Carson-Roberts, Inc

Text (OCR of the scan at left; may contain errors)

IN THE MATTER OF MATTEL, INC.

and CARSON-ROBERTS, INC.

Dockets C-2071 2072. Consent Orders, Nov. , 1971- Modifying Order, Sept. , 1984 In response to ajoint petition from a toy manufacturer and its advertising agency, this Order reopens the proceedings and modifies Paragraph One of two 1971 consent orders issued against Mattei, Inc. , 79 F.T.C. 667, and Carson-Roberts, Inc. (succeeded by Ogilvy & Mather U. , a Division ofOgilvy & Mather International Inc.), 79 C. 674. Noting that Paragraph One of the Orders prohibited the use of certain tim or camera techniques in children s advertising only if they misrepresented advertised product's performance; operation or use, the Commission held that modification clarifying the circumstances under which these techniques may be used would be in the public interest. The Commission therefore added a proviso permitting the non-deceptive use of the techniques if respondents possessed and relied on competent and reliable tests establishing this lack of deception; and a requirement that the supporting data be maintained for a period of two years, However, the Commission declined to modify Paragraph Four on the ground that petitioners had not demonstrated that the requested modification to remove the broadcast advertisement disclosure requirement regarding the incompatibility of some " Hot Wheels" toys with others is supported by a change in law, change in fact or the public in terest.

ORDER REOPENING THE PROCEEDINGS AND MODIFYING CEASE AND DESIST ORDERS On January 19, 1984, Mattei, Inc., and Ogilvy & Mather U. , a Division ofOgilvy & Mather International Inc., (successor corporation to Carson Roberts, Ine., respondents in the above captioned matter fied a joint petition pursuant to Rule 2.51 of the Commission s Rules of Practice to reopen the proceedings and modify the consent orders entered therein. (79 F. C. 667, 674 (1971)) By letters dated June 7 1984, June 21, 1984, July 19, 1984, and August 13, 1984, Petitioners modified their original joint proposal.

The Orders relate to alleged unfair and deceptive practices by Mattei and Carson-Roberts in connection with the advertising oftoy racing sets and dancing or walking dolls. The first three paragraphs of the Orders against the two petitioners are substantively identical. These paragraphs: (1) prohibit the use of certain fim or camera techniques in advertising addressed to children that misrepresent the product's performance, operation or use, (2) limit the circumstances under which endorsements may be used in advertising addressed to 556 FEDERAL TRADE COMIVISSION DECISIONS Modifying Order 104 F.

children, and, (3) require that, whenever two or more "Hot Wheels products which are sold separately are advertised together, that separate status be disclosed. Paragraph Four of the Order against Ogilvy & Mather requires a disclosure in broadcast advertising regarding the incompatibility of some Hot Wheels products with other Hot Wheels products. Paragraph Four of the Order against Mattei requires this disclosure in print advertising and on packages as well as in broadcast advertising. The Order against Mattei also prohibits the misrepresentation of the velocity at which a toy car travels (Paragraph Five) and requires a disclosure regarding the ability of dancing or walking dolls to stand by themselves (Paragraph Six).

Petitioners seek modification only of Paragraphs One and Four of the Order-the paragraphs restricting the use of certain camera techniques 1 and requiring compatibility disclosures. Petitioners seek to add a proviso to Paragraph One ofthe Order. This proviso would state that respondents' use of the camera techniques restricted by Paragraph One would not constitute a violation of that Paragraph as long as they possessed and relied upon results of competent and reliable tests showing that such techniques, in the context of the advertisement as a whole, do not misrepresent the advertised product's performance to the age group of children to whom the advertisement is addressed. Petitioners also propose a provision which requires that proper records pertaining to these tests be maintained. Petitioners argue that the addition ofthis proviso to the Orders is in tbe public interest. They point out that the current Order does not flatly prohibit the use of the camera techniques described in Paragraph One. They are proscribed only if the result of their use "in the context of the advertisement as a whole is to misrepresent the product' s performance, operation or use to the age groups of children to whom the advertisement is addressed, taking into consideration the level of knowledge, sophistication, maturity, and experience of such age group or age groups. " Paragraph 1 , Clauses (a)(b)(c) and (d). Petitioners claim that even though the Order is thus not designed to totally ban the use of these camera techniques, such is its practical effect. They argue that the order language fails to give them clear guidance by which they can tailor their actions for compliance, and that whether any given use of a particular camera technique in a , Thc\;e t.t:chniques are: (1) tlw use of "any vis",,1 perspective which purports to be but i not. one which iI child can experience .. J "laJny sequence ufdifferentvisual perspectives which purportstu depict perspectives which a child can experiencEC but which changes laster than il child can change his visual perspective . " (3) "(aJny visual per3pectivp. whicr. purports t.o dcpict the "dual pcrf()rmllDCe ofa particular function of the product and which differs 5ubstanlia1ly from ,he length of time required to perfurm that fuvction. ; " and (4) "klamera over-cranking ur under-cranking to rJpict a performance ch,-ractcristic of s\lch prorJud which does not ,"exist or cannot b perceived under ordirmry conditions ofthc product s use. " 179 F, C. 672, 679 (19711J , Ogilvy &. Mather, successor to Carson-Rolwrts, Inc., a180 proposes that. t.he name, Ogjjvy & Mather U.s., a division ofOgiJvy & Mather International, Inc., be ;mbstituted for t.he name C"rson-Robert., inc, wherever t.h latter name "pp ars in t.he Order In Dockel C .2072 iVJAll.Csl.J, )...v, 555 Modifying Order particular advertisement "misrepresents" the product can only- be decided in a post hoc Commission civil penalty action. Therefore, to avoid such an action, they must totally abstain from the use of such techniques.

Petitioners also propose that Paragraph Four ofthe original Orders be modified to remove the broadcast advertisement disclosures currently required. In place of these disclosures, Paragraph Four of the Matte! Order would be modified to require all package disclosures to appear on the front ofthe packages. Also, the clear and conspicuous standard would be applied to both the package and print advertising disclosure requirements. Finally, Paragraph Four of the Ogilvy & Mather Order would be modified to include the same package front and print advertising requirements as in Paragraph Four of the Mattei Order.

The Commission believes that the proposed proviso to Paragraph One is in the public interest. The theory bebind the Orders clearly was not that the specified camera techniques are always deceptive. Had it been so, they would simply have been prohibited in all circumstances. The Orders do not proscribe those techniques, but only prohibit their use when the effect of such is to misrepresent the product's performance. The practical effect ofthe Orders, however, appears to be to raise substantial questions with regard to what constitutes proper compliance with Paragraph One. The proviso would allow the non-deceptive use ofthese camera techniques. Petitioners would have to possess and rely upon competent and reliable tests establishing this lack of deception. The requested modification would thus provide Petitioners with additional guidance as to the circumstances under which these camera techniques may be used while continuing to prevent that advertising from misrepresenting the performance of the advertised products.

However, the Commission has determined that Petitioners have not demonstrated that the requested modifications in Paragraph Four of the Orders are supported by a change in law, change in fact or the public interest. Petitioners argue that consumers are clearly aware that some Hot Wheels vehicles are not compatible with some Hot Wheels playsets, and can determine from pre-purchase visual inspection whicb vehicles are compatible with which sets. However Petitioners present no evidence in support ofthis assertion. Furthermore, Commission stafl' attempted to test Petitioners' assertion by examining one Hot Wheels set and a variety of Hot Wheels vehicles. Staff was unable to reliably predict from visual inspection which vehicles would be compatible with the set. Petitioners also argue that the broadcast advertisement disclosures currently required by the Order provide no consumer benefit. They claim that an order which Modifying Order 104 F.

requires package-front disclosures, rather than broadcast disclosures and written disclosures whicb may be placed anywhere on the package, would provide better protection to consumers. However, Petitioners have presented no evidence to support this claim. The Commission concludes that Petitioners' arguments for the modification of Paragraph Four of the Orders do not meet the requirements of Section 51(b) of the Commission s Rules of Practice. By letter of August 13, 1984, Petitioners requested that the Commission reopen the proceedings to modify both Paragraphs One and Four of the Orders, or, in the alternative, if the Commission should decide not to reopen the proceedings to modify both of those Order paragraphs, that the Commission reopen the proceedings to modify just Paragraph One as they have requested or reopen the proceedings to modify just Paragraph Four as they have requested. The Commission accepts only Petitioners' alternative request to reopen the proceedings solely to modify Paragraph One of the Orders. The Commission denies Petitioners' other requests to reopen the proceedings for the purpose of modifying both Paragraphs One and Four of the Orders and to reopen the proceedings for the purpose of modifying only Paragraph Four of the Orders.

The modified Orders wil ensure that Petitioners' use ofthe camera techniques specified therein is not deceptive. At the same time, Petitioners' duties under the Orders have been clarified and their burdens of compliance lessened.

It is therefore ordered That the proceeding in Docket Number C- 2071 is hereby reopened and the Order issued November 1, 1971, is hereby modified as follows:

1. The fonowing language shah be added to the end of Paragraph 1 of the Order.

Provided, however That respondent shall not be in violation of the provisions of this paragraph if, at the time an advertisement incorporating any visual perspective or fim or camera technique described in this paragraph is broadcast, respondent possesses and relies upon a test(s) which compares the impressions created by two advertisements which are identical except for the presence or absence of the visual perspective(s) or film or camera technique(s) being tested, and which is suffciently sensitive to measure differences between the impressions conveyed by the advertisements with an accuracy of plus or minus less than five (5) percentage points, and which is otherwise competent and reliable, or any other competent and reliable test(s), which demonstrate(s) that such visual perspective(s) or fim or camera technique(s), in the context of the advertisement as a whole, does not 1\11.11 J I' 1I'L-. , rd Au. iJiJ;J 555 Modifying Order misrepresent the product's performance, operation, or use to the age group or age groups of children to whom the advertisement is addressed, taking into consideration the level of knowledge, sophistication, maturity, and experience of such age group or age groups. Competent and reliable" shall mean for purposes of this Order a test conducted and evaluated in an objective manner by persons qualified to do so, using procedures generally accepted by others in that profession to yield accurate and reliable results. 2. The following paragraph shall be added following Paragraph 6. It is further ordered That respondent Mattei, Inc., shall maintain for at least two years from the date of the last dissemination of the representation, and upon request make available to the Federal Trade Commission for inspection and copying, all test results, data and other documents or information relied upon for any representation containing any visual perspective(s) or fim or camera technique(s) described in Paragraph 1 of this Order, and any information possessed or known of which contradicts, qualifies or calls into serious question that representation.

It is further ordered, That the proceeding in Docket Number C-2072 is hereby reopened and the Order issued November 1 , 1971, is hereby modified as follows:

1. The corporate name, Ogilvy & Mather U. , a division ofOgilvy & Mather International, Inc., shall be substituted for the corporate name Carson-Roberts, Inc., wherever the name Carson-Roberts, Inc. appears in this Order.

2. The following language shall be added to tbe end of Paragraph 1 of the Order.

Provided, however That respondent shall not be in violation of the provisions of this paragraph if, at the time an advertisement incorporating any visual perspective or fim or camera technique described in this paragraph is broadcast, respondent possesses and relies upon a test(s) which compares the impressions created by two advertisements which are identical except for the presence or absence of the visual perspective(s) or fim or camera technique(s) being tested, and which is suffciently sensitive to measure differences between the impressions conveyed by the advertisements with an accuracy of plus or minus less than five (5) percentage points, and which is otherwise competent and reliable, or any other competent and reliable test(s), which demonstrate(s) that such visual perspective(s) or fim or camera technique(s), in the context ofthe advertisement as a whole, does not misrepresent the product's performance, operation, or use to the age .. , ., ;,,,, , Dissenting Statement 104 F. group or age groups of children to whom the advertisement is addressed, taking into consideration the level of knowledge, sophistication, maturity, and experience of such age group or age groups. Competent and reliable" shall mean for purposes of this Order a test conducted and evaluated in an objective manner by persons qualified to do so, using procedures generally accepted by others in that profession to yield accurate and reliable results. 3. The following paragraph shall be added following Paragraph 4. ft is further ordered That respondent Ogilvy & Mather U. , a division ofOgilvy & Mather International, Inc., shall maintain for at least two years from the date ofthe last dissemination ofthe representation, and upon request make available to the Federal Trade Commission for inspection and copying, all test results, data, and other documents or information relied upon for any representation containing any visual perspective(s) or fim or camera technique(s) described in Paragraph 1 ofthis Order, and any information possessed or known of which contradicts, qualifies or calls into serious question that representation.

Commissioner Pertschuk dissented.

DISSENTING STATEMENT OF COMMISSIONER MICHAEL PERTSCHUK The Commission today takes yet another giant step backward from protecting children against commercial exploitation. By agreeing to modify these 1971 orders, the Commission continues to ease the restraints against the sophisticated psychological manipulation of six seven and eight year olds.

The 1971 orders against Mattei and its ad agency were entered to sette charges that ads for MatteI's "Hot Wheels" toys made the toys look bigger and move faster than they would appear to children. Although the orders did not ban Mattei' s use ofthe questionable video techniques, for thirteen years Mattei apparently felt constrained from using a distorted sales pitch to sell its toys. But armed with the knowledge that the current Commission refuses to take a hard look at advertising directed to children ' last January Mattei pleaded hardship and asked the Commission to lift the burden of the orders. The solution agreed to by the Commission permits Mattei to use the techniques listed in the 1971 orders as long as Mattel's comparative I For a more complete critiqlH' of I. he Commission s recent approach lo children' s adverti;;ing, see my Report to the Subcommittee on Oversight ..nd IrlVe t.igiltjons, House Committee on E:energy and Commerce The I'erform- ,mec of the Federa) Trade Commission, 1977-- 1981 " at IV-61- IV- 'In its petition, Mattelnotcd that the staff has recently closed inve tigalions of two ofiL competitor fur u"ing video techniques similar to those prohibitetl by the 1971 orders, The fact that those cases were rJotpursued is indeed ..." nl" n, . , ;A a that, "nd r,I" t1,n rnm.,, ;oo;cm ' nnt;"" hrl;,, t;nn "f ,I ..t "I.,t,, v rnlp .._ .. . .....

555 Dissenting Statement testing ofthe camera techniques on two groups of children shows that the ads do not misrepresent the toys. But comparative testing of ads directed toward children is not adequate. I am skeptical that techniques for testing children s perceptions are suffciently reliable to demonstrate that an ad is nondeceptive. As the FTC staff noted in its 1983 recommendation to close the Children s Advertising Rulemaking, "(YJoung children do not possess the cognitive ability to evaluate adequately child-oriented television advertising." To the best of my knowledge, the Commission has previously respected this fact and has never before entered an order that incorporates a requirement for testing directed at children. Mattei has not presented any reason why the Commission should alter this long-standing policy and chance sanctioning deception.

g., Initial Decision 104 r' to determine from their forms where the patients had applied the cream and which area(s) of pain had been relieved (Golden, Tr. 2875- 76; Marlin, Tr. 3378). Most ofthe Golden study subjects had at least two affected areas of pain (Golden, Tr. 2969; CX 213F- 057). 265. According to Dr. Marlin, the reason that patients were not asked to identify the specific site(s) of pain reJiefwas that at the time the Golden Study protocol was designed, the focus of the study was on pain relief generally (Marlin, Tr. 3204). The implication of nonspecificity in these forms, however, is that since the patient reporting forms used in the Golden study were similar to those subsequently employed in the Golden and Altschuler study (RX 50/CX 214), the inability to determine the location of pain relief from the forms detracts from the result claimed from the latter study that Aspercreme achieved statistically significant pain relief in non-weightbearing areas of the body.

266. A further problem that compounded the failure to pinpoint the location of pain relief on the Golden study patient forms is the fact that rheumatic disease, and particularly arthritis, is cyclical in nature. Thus, a study subject could come in with pain in the shoulder but three days Jater could experience pain in the back or hip (Golden Tr. 2880-81). The actual site of cream application thus could be different from the location of pain recorded on the diagnostic portion ofthe patient forms (see, e. CX 213Z-030). 267. Also, as regards to completion ofthe patient forms, Dr. Golden acknowledged that due to the way the Clinical Data form and Patient Reporting form were designed, if a subject had six affected body areas of pain, the subject would record (81) "partial relief" rating regardless of whether he or she experienced pain relief in only one area or five. And there is no way on the face of the form that the person reviewing it could tell exactly in which affected area or areas the pain relief occurred (Golden, Tr. 2887).

268. Another problem with the Golden study is that several test subjects (seven in number, comprising 18% ofthe sample) were found to have taken one or more concurrent analgesic, antirheumatic, or mood-altering drugs (CX 45Z-022, Admission No. 4 (patients 1, 2, 6 , 10, 24, 31); CX 213S (patient 8)). The drugs included Tylenol, Medrol, Valium and Librium. Use of such concomitant medications is unacceptable in a non-crossover study and may have seriously affected the study results (Adriani, Tr. 1192; Roth, Tr. 1577). Moreover since the subjects were not given a washout period from preexisting aspirin usage, it is unclear what was being measured as far as subjects in the aspirin/placebo cream group were concerned (Adriani, Tr. 1190 91). Thompson s expert, Dr. O'Brien, agreed that subjects using a significant amount of another analgesic or anti-inflammatory drul! J IIV1V.r \.l Ul.cVJ'-,H. I-, v\..

648 Initial Decision should have been dropped from the data base (O'Brien, Tr. 3759-60). However, if you delete the data for these patients in order to eliminate the problem of concurrent medications, you further reduce the sample size of each subset, with the result that the value ofthe reported results is further diminished (Marlin, Tr. 3415-16). 269. The record also shows instances of significant protocol breach. For example, Dr. Golden included patients with particular (i. joint) pain, patients who were outside the age parameters specified in the study protocol, and patients with mild to moderate pain as study subjects. These were not trivial deviations. 270. Departures from a study protocol should be minimized, and any major change or amendment to the protocol should be in writing. Dr. Golden did not make any written changes to the study protocol set forth at ex 213A-e (Golden, Tr. 2785). Although Dr. Golden testified that the protocol was subsequently amended orally, he was unable at his deposition to recall any amendments, either written or oral to the study protocol indicated at ex 213A-e (CX 45Z-005 (Admission No. 257)). The claimed oral modification of the study protocol is also contrary to respondent' s admission that Dr. Golden had "inadvertent- " included patients with arthritic conditions as subjects in the ex 213 study (Golden, Tr. 2800; ex 45N (Admission No. 268)). 271. As regards the age range for subjects specified in the study protocol, Dr. Golden acknowledged that at least thirteen of the forty subjects (about 30%) were outside the age parameters specified (Gold- , Tr. 2804). And despite the (82) protocol requirement that study subjects have moderate or severe pain, at least six ofthe forty subjects had only mild to moderate pain symptoms (Golden, Tr. 2831- , 2834- 35) 272. Another problem with the Golden study is the lack of consistency in Dr. Golden s "global evaluations" of his patient's condition following their participation in the study, which reflected his own subjective global opinion and was not based on any numerical scoring system (Golden, Tr. 2850, 2853, 2857-58). As a result, Dr. Golden was less than convincing in explaining some of his "global evaluations" at trial.

273. There are also some questions about the validity ofthe study blinding process, based on comments by subjects about headache from the test agent's odor, and bitter taste (Roth, Tr. 1578-79). Finally, there are acknowledged errors in two tables and on three out of the six pages oftextual material in the published report (Golden, Tr. 2923; Marlin, Tr. 3239- , 3246; ex 200 , Tables I and and pages B, e and D). Dr. Marlin s explanation of the errors in the two tables was that they stemmed from the fact that data from different sources were Initial Decision 104 F. inappropriately combined and patient ratings of degree of pain were confused with the doctor s ratings (Marlin, Tr. 3243, 3246). 274. For all of the foregoing reasons, the Golden study (RX 48/CX 200) is not an adequate and well-controlled study, and its results cannot be relied on to demonstrate the effcacy of Aspercreme for pain and anti-inflammatory relief in this proceeding. B. The Golden And Altschuler Studies (RX 50/CX 214) 275. In 1977, Dr. Steinberg asked Dr. Marlin to design and develop a second study for Aspercreme, a study that would compare Aspercreme to a. placebo cream (Marlin, Tr. 3254). Dr. Marlin decided to use two doctors and two sites so that a large patient population could be tested. Dr. Stanley Altschuler, an internist, was chosen as the second investigator and Dr. Golden again served as the rheumatologist (Steinberg, Tr. 5157-58). Dr. Marlin again drafted the protocol and reviewed it with the two investigators. The patients were to be divided into two groups, one which would apply the placebo cream and one which would use Aspercreme (RX 50). Dr. Marlin periodically visited each ofthe two investigators to monitor the study and to ensure that the protocol was being followed by both investigators (Altschuler, Tr. 3010 3065; Golden, Tr. 2710; Marlin, Tr. 3257, 3259-61) (83) 276. Drs. Golden and Altschuler each instructed their patients to apply as much of the topical medication to the affected area as was possible and to record the pain relief experienced over a four hour period-at the one-half hour, one hour, two hour, and four hour marks (Altschuler, Tr. 3015-16; Golden, Tr. 2719-21) 277. Prior to the commencement of the studies, it was decided that the protocol should be changed to raise the maximum age allowable in the patient population because Dr. Golden was having diffculty finding patients that fit within the original protocol. The exclusion of patients with rheumatoid arthritis or osteoarthritis was also modified to allow for the use ofthese patients provided they had non-particular features which could be treated by the cream (Altschuler, Tr. 3085-86; Golden, Tr. 2714-17). These changes in the written protocol are not serious since the purpose of the study was to test the pain relieving properties ofthe test drug on non-particular involvement (O'Brien, Tr. 3755).

278. Dr. Freudenthal followed the same basic procedures that she had utilized in the Golden study. She prepared the randomization and the code, packaged the medication in unmarked boxes, then opened the code and analyzed the data after the study was completed (Freudenthal, Tr. 4915-16).

279. The Golden and Altschuler studies was thus a double-blind tll(h7 lH.lnd" n inphr-.:iv n!'tjpnfr; t turn rJiff'prpnt ..iip.. urhi,.h tp-.tt::.,j ).

.l.l.lVH.l.l ,-v.. .Iu.u.I.l'-.n.u '-'-0'H''-. 648 Initial Decision Aspercreme against placebo cream for the relief of various types of musculoskeletal pain. The results of the studies, including Dr. Marlin s 1981 analysis (RX 368F-I), fail to show any statistically significant differences overall between Aspercreme and placebo for pain relief (F. 281 infra. This study was rejected by the FDA in the Tentative Final Monograph on OTe Analgesic Products, dated February 8, 1983 (eX 443D). The reason given by the FDA was: . . . Ofthe six results reported, only one was statistically significant. Furthermore, the selective reporting of these six results renders this report uninformative, and no conclusion can be made concerning the effectiveness of trolamine salicylate LTEA/Sj. 280. According to Thompson s expert witnesses, the results of the study indicated a clear tendency toward greater pain relief with Aspercreme than with the placebo creme at the one-(84)half hour and one hour marks. Twelve patients using Aspercreme, as against ten using the placebo, experienced complete relief at the end of at least two ofthe four time periods. And 58% ofthe Aspercreme subjects, as against 30% of the placebo subjects, reported relief for eight or more hours (Freudenthal, Tr. 4924-25; RX 82, 84). In patients who suffered from cervical (neck) pain, significantly more patients obtained relief with Aspercreme than with the placebo. This difference was statistically significant at the 95% confidence level (Freudenthal, Tr. 4925- 26; Marlin, Tr. 3272; RX 82, 84). Groups of patients with pain in other areas were not large enough in size to compare, but analysis of the data in terms of weight bearing and non-weightbearing areas revealed that there is statistical significance for the superiority of Aspercreme at the one-half hour mark for non-weightbearing areas and joints, and in the cervical area, there is statistical superiority at one-half hour one hour, and two hour marks (Freudenthal, Tr. 4925; Marlin, Tr. 3277).

281. The record shows that the overaJi results of the Golden and Altschuler studies fail to show any statistically significant differences between Aspercreme and placebo for pain relief(Adriani, Tr. 1195-96; Freudenthal, Tr. 4988; O'Brien, Tr. 3834; Roth, Tr. 1585; ex 45Z-21 (Admission No. 1); ex 214L). Three separate analyses of the data including the initial analysis by Dr. Freudenthal (Freudenthal, Tr. 4988), and the subsequent reanalyses of the data first by Dr. Winick (Marlin, Tr. 3425-26) and later by Dr. Marlin (eX 45Z-021, (Admission No. 1); ex 214L), reached the same conclusion in this respect. Even using the Marlin analysis of 1981, it would be fair to say that none of the overaJi figures approached statistical significance at the 95% confidence level (Ehrlich, Tr. 4195; ex 214L). Rather, the likelihood that the differences measured were due to chance ranged from Initial Decision 104 F. slightly under 20% to over 90%, depending upon which time period is considered (See ex 214L).

282. Thompson relies heavily on the Marlin analysis, particularly on the subgroup involving non-weightbearing areas and joints, the only subgroup where statistically significant differences favoring Aspercreme could be shown on some ofthe time parameters (See Marlin Tr. 3276, 3282; O'Brien, Tr. 3834). However, the Golden and Altschuler studies cannot fairly be interpreted as providing any reliable evidence of Aspercreme s superiority to placebo, even in nonweightbearing areas.

283. The original analysis ofthe Golden and Altschuler data by Dr. Freudenthal did not conclude that Aspercreme was superior to placebo in the non-weightbearing areas (Steinberg, Tr. 5246-7). In fact placebo was equal or superior to (85) Aspercreme in three out of the five non-weightbearing areas that Dr. Freudenthal analyzed (i. hands, shoulders, and arms) (Jd. See RX 85J-L). Of the two areas where Aspercreme was superior to placebo (i. cervical and back), only one (pain relief in the cervical area) was statistically significant. However, there were not enough subjects with cervical arthritis in the study for Dr. Freudenthal to reach a conclusion about statistical significance with respect to arthritis in the cervical area (Freudenthal Tr. 4996). The fact that Dr. Freudenthal did not find that Aspercreme was superior to placebo in non-weightbearing areas is entitled to considerable weight. Not only was Dr. Freudenthal' s the first and contemporaneous analysis, but also her analysis employing a nonparametric statistical test (Freudenthal, Tr. 4958-59) was of greater value than Dr. Marlin s later reanalysis employing a parametric test (O' Brien, Tr. 3838-39).

284. It is significant to note that the distinction between weightbearing and non-weightbearing joints and areas came not from the study protocol or Dr. Freudenthal' s biostatistical report, but from a computer analysis done by Dr. Marlin in 1981, three years after the study was conducted (Marlin, Tr. 3423; Steinberg, Tr. 5245). Such post hoc analysis of clinical data calls into question the integrity of the result because of the potential bias present in any rearranging or manipulation of data (Roth, Tr. 1591-92).

285. The rating scale to be used in a study should be developed and set forth in the protocol before the study is begun so as to avoid data manipulation (Adriani, Tr. 1199-200; O'Brien, Tr. 3757, 3759; Roth Tr. 1591-92). The record shows that the scaling system used in the Golden and Altschuler studies (RX 50/CX 214) was developed by Dr. Marlin long after he had broken the code, learned what the raw data showed, and read Dr. Freudenthal's original biostatistical report lin rrr .1..1').'1 L1?7-,)A) 'Th-j !'n lrnnrronpr nrf,,"p. l1rp. !:nl rcm- THOMPSON MEDICAL CO., ING l;jl 648 Initial Decision ders the results questionable (Adriani, Tr. 1199-200; Roth, Tr. 1591- 92). Dr. Marlin acknowledged that the results of the Golden and Altschuler study can be significantly affected by the type of scaling system used (Marlin, Tr. 3429). The record demonstrates that Dr. Marlin s scaling system have affected the analysis in favor of Aspercreme (See ePF 354).

286. Another problem with the Marlin analysis of non-weightbearing joints and areas is the nature of the underlying data. The patient report forms used in both the Golden and the Golden and Altschuler studies did not require the patients to distinguish between weightbearing and non-weightbearing parts of the body in recording pain relief (Marlin, Tr. 3365). Moreover, for patients having two or more areas of pain, it was impossible in the years following the study to determine from the patients' (86) forms exactly the area(s) ofthe body patients applied the cream to and the areas ofthe body the pain relief ratings came from (Altschuler, Tr. 3062, 3065, 3066-69; Golden, Tr. 2954-56; Marlin, Tr. 3378). There were also many instances where the data summary sheets (eX 366Z-114-19) and the entries in individual patient forms were not in agreement (See Marlin, Tr. 3375-91). 287. It is well-recognized that if a body of data is divided into a large number of small cells, some statistically significant differences wiu eventually appear among some of them (See Marlin, Tr. 3433- 3475; O'Brien, Tr. 3841-42; Roth, Tr. 1587; (CX 417 received at 3847 3848-49; CX 435F -G received at 3853, 3856)). This is called random statistical significance (Marlin, Tr. 3475; Roth, Tr. 1587-88). In the Marlin analysis, the summary table shows no statistically significant differences in the total sample (Roth, Tr. 1586; See ex 214L). The fact that after the data is divided into a large number of subsets, a few differences in the non-weightbearing areas can be shown to favor Aspercreme at statistically significant levels is consistent with the concept of random statistical significance (Roth, Tr. 1586-88). 288. Apart from the major problems related to data analysis and interpretation discussed hereinabove, the Golden and Altschuler studies suffer from several significant flaws in its design and execution. As with the earlier Golden study, the patient population included an unacceptably wide array of conditions and diseases (Roth, Tr. 1589-90).

289. There is also some question regarding the propriety of pooling the Golden and Altschuler study data in this case. 290. In the Golden and Altschuler study, the decision to involve a second investigator occurred after Dr. Golden s portion of the test had begun (eX 45Z-003 (Admission No. 213)). The study protocol did not provide for a multi-site clinical trial (eX 214Z-022-D24). The study was to consist oflOO subjects to be studied by Dr. Golden. Only when Initial Decision 104 F. Dr. Golden expressed his doubts about finding the requisite number of subjects after his study got under way, the study was changed to a multi-site study (Marlin, Tr. 3255-57).

291. It is a requirement of a multi-site study that not only the same protocol be adhered to by all investigators, but also the patient groups be homogeneous in order that the data obtained from the different groups may be combined (Marlin, Tr. 3259-60; O'Brien, Tr. 3831) If the patients in the different groups are dissimilar or if they are being treated for different conditions, pooling the data is inappropriate (O' Brien, Tr. 3831). In the Golden and Altschuler study, (87) Dr.Gold- , a rheumatologist, provided mostly rheumatology patients while Dr. Altschuler, an internist, provided general medical patients (eX 45N (Admission No. 253)). There were other significant difierences between the patients in the two groups to render pooling of the data questionable (Roth, Tr. 1592-93). For example, in Dr. Altschuler group, thirty-two of the fifty patients (or 64%) self-rated their baseline pain as severe while only four of Dr. Golden s forty-five patients (or 11 %) did so (Altschuler, Tr. 3086-87). Dr. Altschuler attributed this difference to the differences in the underlying conditions of the two patient groups rather than to the differences in their pain perception (Altschuler, Tr. 3087-90).

292. It is also important to determine that the results obtained by the investigators are similar before pooling the data in a multi-site study (O'Brien, Tr. 3815). Thus, the criticism of the FDA' s Bureau of Drug directed to the Golden and Altschuler Study and its underlying data submitted by Thompson, included a comment that the study reported the diagnosis and location of the pain on one table for both groups and the statistical analyses treated all ninety-six patients without distinguishing investigators (eX 342B). 293. It is also important in a multi-site study that the different investigators adhere to the same protocol (O'Brien, Tr. 3815). In the Golden and Altschuler study, however, both investigators stated that they knew of no checks to insure that the two physicians conducted the test in the same manner (eX 451 (Admission No. 140); CX 45N (Admission No. 256)). Thus, there is no assurance that the two physicians applied the same criteria in gathering the clinical data and background information on the patients, or that they identified the primary diagnosis or the primary site in the same manner (ld. Thompson has admitted that Drs. Golden and Altschuler never discussed any aspect ofthe test with each other (eX 45N (Admission No. 255)). As ofthe time of Dr. Golden s deposition in this case (December 1981) the two physicians had never even spoken to one another (eX 45M (Admission No. 241)). In fact, Dr. Altschuler testified that during hi!' ron(hH't. of t.hp r;t.nnv. hp w:; nn:;w::rp t.n::t. :;nnt,npr invp.o:t.ip':;t, l.tlUlVlY;)Ul," Nlr.VICJ\L CU., ll,"C. 648 Initial Decision was conducting a trial with a protocol that was identical to his (Altschuler, Tr. 3059).

294. Another problem with the Golden and Altschuler study is the fact that nine subjects (or about 10% ofthe total sample) at the time of their participation were using anti-inflammatory or mood-altering drugs which may have distorted the study results (Adriani, Tr. 1198; Roth, Tr. 1588-89; ex 45Z-016-17 (Admission No. 5) (patients, 101 105, 109, 130, 131 , 132, 135, 139, 140)). The drugs included tranquilizers such as Valium and Librium, and anti-inflammatory drugs such as (88) Prednisone and Motrin. In view of the fact that the study was not a crossover study, the use of such concomitant medications should have been discontinued (Adriani, Tr. 1198; Roth, Tr. 1588 89). 295. Although the protocol of the Golden and Altschuler studies prohibits concomitant use of "other analgesic medication" during the four-hour test period, neither of the two physicians was instructed about medication usage in the period prior to the four-hour test period, nor was either instructed as to specific types of non-analgesic medications which could affect study results if used during the test period (Marlin, Tr. 3396-98). Dr. Altschuler admitted that he did not question study participants after the test about any concomitant pain medications that they might have used either before or during the four-hour test period and that some subjects, especially those who applied the test cream after a lapse oftime following their initial visit with him, could have taken analgesic medication prior to commencing the four-hour study period (Altschuler, Tr. 3082-83). Such analgesic medication by subjects, Dr. Altschuler agreed, would not have violated his instructions yet could have affected the test results reported by the patient during the four-hour test period. Thus, because ofthe lack of a washout period for analgesic medications prior to the subjects participation, and because a number of study subjects took concurrent anti-inflammatory or mood-altering drugs which could have affected study results, concomitant medication usage is a significant problem in the Golden and Altschuler study.

296. In addition, there were several patients in Dr. Altschuler Aspercreme group who are known to have breached the protocol by applying the cream two times rather than once (patients 11, 30, 32 and 52 (CX 214Z-079 Z-96, Z-98 Z-112)), aU of whom recorded that they had experienced pain relief (eX 366Z-114-19). This raises a question of potential bias in favor of the Aspercreme group. 297. Dr. Golden testified at trial that the written protocol, requiring exclusion of patients with diagnosed rheumatoid arthritis or osteoarthritis as weU as patients older than fifty-five years, was subsequently modified orany to allow inclusion of arthritis with non-particular pain and persons of older age (Golden, Tr. 2714-17). This testimony is ,..

Initial Decision 104 F. contrary to his deposition testimony where he had recalled no written or oral amendments to the protocol (Golden, Tr. 2937). In any event the patient's forms for some twenty-five subjects indicated arthritic pain and some others only name the affected area or areas ofthe body. With respect to those subjects, it is not possible to determine from their patient forms whether the pain was of non-particular nature and whether their inclusion (89) was proper under the orally-amended protocol (eX 214Z-060, Z-070- , Z-077, Z-078, Z-081- , Z-084- Z-094- , Z-097- , Z-100, Z-102, Z-104, Z-106, Z-108- , Z-113- Z-118-1l9).

298. For all of the foregoing reasons, the Golden and Altschuler study (RX 50/eX 214) is not an adequate and well-controlled trial and does not constitute a reasonable basis for Aspercreme effcacy claims. This determination is in accord with that ofthe FDA Final Tentative Monograph for OTC Analgesic Products, published on February 8 1983 (eX 443D).

e. The French Studies 299. The French studies relied on by Thompson as substantiation for Aspercreme effcacy claims include the two clinical studies conducted in France for the L'Oreal corporation by Dr. Alain Patel (RX 34/eX 209) and Dr. Pierre Andre Chappelle (RX 35/eX 208), respectively, and a follow-up study conducted by Dr. Patel (RX 36/CX 120), aJl during the period 1976 and 1977. A June 4, 1981 statement authored by Dr. Patel (RX 37/eX 253) and a July 6, 1981 letter of Dr. Patel to Dr. Steinberg of Thompson (RX 38/eX 266) also pertain to the Patel studies (RXs 34, 36). Thompson acquired RXs 34 and 35 in early 1977 (Steinberg, Tr. 3139-40) and RXs 36-38 in the summer of 1981. Thompson s Aspercreme advertisement began on a nationwide basis in 1977. These French studies were conducted to meet the French regulatory requirements by government designated investigators. For our purposes, they were uncontrolled and do not permit a proper assessment of their results. Seethe FDA Tentative Final Monograph on OTe External Analgesic Products, dated February 8, 1983 (eX 443D). In any event, they fall far short of adequate and well-controlled clinical trials and cannot be relied on as providing a reasonable basis for making any effcacy claim for Aspercreme. 300. At the time when Thompson purchased the product from the Sperti Drug Company, Inc. in 1976, two studies on the effcacy of Aspercreme were in progress in France. These studies were conducted for the L'Oreal company (" Oreal") which sought to obtain a license to sell Aspercreme in France. Under the French regulatory scheme a drug must undergo toxicology studies and clinical trials designed to hr..., ;h, f'",hr rl ffiro"''' rp",tal 'T.. 1.R17_ 1.R l.R?F;'I 'Th", ('lin -i(' ITIUlVt':)Ull 1Vl Ull.A.L LU. ll'll;. ''ll 648 Initial Decision trials must be conducted by two French physicians whose names appear on the French Ministry of Health list of clinicians approved for such trials (Patel, Tr. 1818, 1826-27). The two studies for L'Oreal were conducted by French physicians, Drs. Alain Patel and Pierre Andre ehappelle. However, L'Oreal subsequently decided (90) that because ofa change in the French Social Security regulations it would not be profitable to market the product in France (Patel, Tr. 1920). Thompson received the results of these studies in the early part of 1977 (Steinberg, Tr. 5137-40).

301. The patient population for these two studies was drawn from the Rehabilitation center ofthe hospital in DeauviIJe, France and the Raymond Poincare Hospital in Paris, France. Dr. ehappelle, head of the Rehabilitation center and a well-known expert in the rehabilitation of rheumatoid and traumatic injuries, supervised the study at the Deauville site. Dr. Patel supervised the Raymond Poincare site. The patients who participated in the study were suffering from pain and swelling either as a result of a rheumatologic disease or a traumatic injury. All of the patients at Dr. Chappelle s site were hospital inpatients, all of those at Dr. Patel's site were out-patients (Patel, Tr. 1868- 1875). According to Dr. Patel, in initial meetings with representatives from L'Oreal, Drs. Patel and ehappelle decided that they would observe the effects of the drug over a two-week period and would confine the study to pain caused by trauma and pain in and around the joints. It was decided that fifty patients would be evaluated, twenty-five from Dr. Patel' s hospital and twenty-five from Dr. ehappelle s Rehabilitation center. Drs. ehappelle and Patel worked closely together and conferred on the instructions to be given to the patients and the reporting form which would be used to record the results. According to Dr. Patel, the test was structured in a standard fashion that had been used many times before to test anti-inflammatory and analgesic preparations (Patel, Tr. 1872-76). Also, patients were told to apply the cream to the painful area twice a day, or three times a day if needed. The patients were seen by the doctors three times during the course of the study at the beginning, after one week, and after two weeks. The total patient population numbered fifty-two (Dr. Patel included two additional subjects), a number considered adequate for drug studies in France (Patel, Tr. 1875- , 1881- , 1976).

302. Reports from clinical examinations were compiled on relief of pain, swelling, inflammation, and on improvement in ease of movement (RXs 34, 35). Results of the study were recorded on the patient' medical record and data collection forms. According to Dr. Patel neither the doctors nor the patients knew the contents of the unmarked tubes of test drug received from the sponsor and sealed en- ;.. .. .. ..,, ;.... + ; + , Initial Decision 104 F. velopes contained the identity of the contents of each bottle. According to Dr. Patel, this is standard practice for drug testing in France (Patel, Tr. 1876-77, 1879, 1881, 1962, 1996). Dr. Patel also testified that, in forming their conclusions, the physicians also consulted the medical charts they maintained for each patient as well as the clinical data reported on the data (91) collection forms. When the study was completed, they studied the charts and the clinical case forms before writing their recommendations to the Ministry of Health (Patel, Tr. 1962, 1996).

303. Nineteen of the twenty-seven patients at Dr. Patel's site were reported to have shown "good results " and all but one of the twentythree patients at Dr. ehappelle s site were reported to have "noticed a very clear improvement" (RXs 34A, 35D). According to the investigators, these findings indicate that test medication was very effective in relieving mq.scale aches, pain from tendonitis, and pain from inflammatory disease. No results were observed where the patient was suffering from serious arthritis affecting large, weightbearing joints (Patel, Tr. 1906-07, 1918- , 1931; Steinberg, Tr. 5142; RXs 34 35).

304. Dr. Patel testified that in February or March of 1977, he conducted an informal follow-up study of Aspercreme in order to satisfy himself that the product did work as well as his earlier findings indicated. This time he knew that the cream he was administering was Aspercreme, and he used the product on approximately forty patients and kept records on twenty-five. Dr. Patel testified that the second study results confirmed the findings of the first. Dr. Patel concluded that Aspercreme is effective in providing relief of pain and swellng and in improving facility of movement in cases where there is rheumatic involvement around the joint. He also concluded that Aspercreme is effective in cases that have rheumatic participation inside the joint, that is, in the synovial fluid. Furthermore, patients with arthritis ofthe small, non-weightbearing joints get good relief of pain when using Aspercreme (Patel, Tr. 1920- , 1931-32). Dr. Patel testified that he was so impressed with Aspercreme that he has used it since for his hands and has continued to recommend it to patients who are able to purchase Aspercreme in this country (Patel, Tr. 2042 2045).

305. However, the record shows clearly that neither of Dr. Patel's studies (RXs 34, 36) constitutes a well-controlled, double-blind clinical study. Dr. Patel's studies consist mainly of clinical observations expressing a global evaluation ofthe product (Roth, Tr. 1604-05). There was no written protocol for either RX 34 or RX 36 (Patel, Tr. 1980), no scoring or scaling system for pain relief, no record ofthe subjective hf.. f;"".. f-1... 'nO '''' (D.. l '1.. 1aQ s;A'\ ",nrl..". ).

I I IVlnrDVl'l lH.LdJl,"n.l. '"V. , H',". 648 Initial Decision information on concomitant medications (Roth, Tr. 1604-05; Adriani Tr. 1214). Both investigations also involved self-limiting conditions ' design Ud. Thethat were not taken into account in the studies studies are notable in their lack of adequate data entry and analysis customarily found in (92) other reports of clinical trials. Finally, the Patel studies are seriously flawed by the absence of a placebo or any other control and the unblinded conditions under which his observations were made (eX 342e).

306. The ehappeJJe study (RX 35/CX 208) shares ajj the flaws discussed hereinabove with respect to the Patel studies (See Adriani, Tr. 1215).

D. The Batterman And Sanders Myoflex Study (CX 254) 307. Thompson also relies on a Myoflex study conducted by Drs. Batterman and Sanders (eX 254/CX 344Z-148-56). 308. This was a double-blinded, placeblo-controlled, multi-center study to which Dr. Batterman contributed twenty-eight patients and Dr. Sanders thirty-five. The study employed a cross-over design and compared the effcacy of Myoflex cream (a 10% topically-applied TEAlS ointment like Aspercreme) with that of placebo cream in patients with arthritic involvement ofthe hands. The nature of a crossover design is that each subject in the study uses the test and control agents sequentially, and the subject as weJJ as the investigator are blinded (Roth, Tr. 1534).

309. The Batterman and Sanders study employed a total of six measures, or three measures for each of the two groups. There were two objective measures (hand grip strength and finger joint circumference) as weJJ as one subjective measure (global improvement) for each of the two groups. The results showed no difference in patient response between Myoflex and placebo cream in five out of the six measurements taken between the two groups. Neither investigator found any difference between TEAlS and placebo cream in terms of either of the two objective measures of improvement. Also, Dr. Sanders found no difference between TEAlS and placebo in terms of patients' subjective impressions of improvement. But Dr. Batterman reported a significant difference between TEAlS and placebo in terms of patients' subjective impression.

310. For several reasons, the Batterman and Sanders study cannot be regarded as adequate support for Aspercreme effcacy claims. First, as was emphasized in the FDA' s Tentative Final Monograph on OTe External Analgesic Products, on five out of the six parameters used to measure drug effcacy, TEAlS was no more effective than placebo (CXs 343B, 443D). In light ofthese results, the FDA concluded that the study does not indicate any clear superiority of TEAlS over Initial Decision 104 F. placebo (eX 443D). Secondly, in (93) terms of the subjective impress group, thesions of improvement reported by subjects in Batterman report lacks any information as to what subjective measure(s) of im- Itprovement was employed (Adriani, Tr. 1201-03; Roth, Tr. 1595-96). may have measured the patient's impression of reduced pain or improved function, and the improvement mayor may not have been c1inicaJ1y significant (Adriani, Tr. 1202-03; Roth, Tr. 1594). Without knowing what questions the patients were asked by the doctors, the subjective improvement" parameter is too vague to be relied on (Roth, Tr. 1597). For this reason, the FDA External Analgesic Panel considered the Batterman and Sanders study to be not adequate and well-controlled. It felt that the report relied heavily on the subjective improvement reported by the Batterman subjects but did not indicate what the subjective improvement consisted of (Adriani, Tr. 1459). 311. There are other important information gaps in the Myoflex study. It lacks information on concomitant medication usage by the on the typesubjects, on the frequency and duration of applications, ofbJinding techniques used, and on how study dropouts were treated (Adriani, Tr. 1204; Roth, Tr. 1598). The inclusion often subjects listed as being in a "quiescent phase" also means that those patients had no active disease (Roth, Tr. 1596).

312. In view of the foregoing omissions and problems, the Batterman and Sanders study is not an adequate and weJ1-controlled clinical trial. In any event, this study was not seen by Thompson before it began Aspercreme advertisements, for Thompson acquired it some- P). time between 1979 and 1981 (Admissions, ex 45D, E. The Bioavailability Studies 313. A drug is said to be "bioavailable" when it has been absorbed into the body and is present in the blood, urine, or other body tissue or fluid. A drug is said to be "bioactive" when it also shows a significant therapeutic effect in the human body. For example, if a person with a bacterial infection takes an antibiotic that is generally recognized as being effective, but the strain of bacteria causing the infection is resistant to that antibiotic, the antibiotic wil in fact get into the person s system and thus be nbioavailable " butnot " bioactive" in that its presence in the body wiJ1 not show a therapeutic effect (Adriani, Tr. 1178-79).

314. According to the FDA, demonstrating that a drug i bioavailable and demonstrating its effcacy are not the same thing: (94) It is not. . . the intent of a bioavailability study to demonstrate effectiveness. The purpm:;e of a bioavailability study is to determine the rate and extent of absorption. If a drug product is not bioavailable, it cannot be regarded as effective. However, a Lt1\J1Vlr,:vn lH.lUHJ'1.J' VV. .u.

648 Initial Decision determination that a drug product is bioavailable is not in itself a determination of effectiveness. The requirement of evidence ofbioavailability is intended to supplement no(tJ replace, clinical evidence of ctiveness. 42 FR 1640 (1977).

The record in this proceeding also clearly shows that bioavailability studies are not a substitute for well-controlled clinical trials for the purpose of showing the effectiveness of a drug (Adriani, Tr. 1178; Ehrlich, Tr. 4087; O'Brien, Tr. 3964-66; Rabinowitz, Tr. 3519; Roth Tr. 1566).

315. As part of its reasonable basis materials, Thompson heavily relies on a radioisotope experiment on dogs and humans conducted by Dr. Joseph L. Rabinowitz and others (RX 70lex 374). The canine portion of the study involved ten dogs, five of which were given oral aspirin tagged with radioactive carbon-14 and five which were given radioactive TEAlS topically applied at the knee. The human portion involved six subjects who were first given radioactive aspirin orally, and two to six weeks later were given radioactive TEAlS topically applied at the knee. In each case, tissue and fluid samples were analyzed for radioactive material at specific intervals after the drug had been administered. The presence of radioactive materials in tissue samples will show drug penetration of skin and absorption into subcutaneous tissues. However, only clinical studies can demonstrate analgesic effcacy (O'Brien, Tr. 3868-69; Roth, Tr. 1601 , 1728). Dr. Rabinowitz agrees that there is nothing in his study to show that TEAlS is an effective analgesic agent (Rabinowitz, Tr. 3518; accord Roth, Tr. 1599; Adriani, Tr. 1206-08).

316. In 1978, Thompson issued a grant to the University ofPennsylvania for a study that Dr. Joseph Rabinowitz wanted to conduct. From time to time after the study was underway, Dr. Steinberg of Thompson received from Dr. Rabinowitz preliminary written reports relative to the amount of salicylate available in tissues after the application of TEAlS as compared to the ingestion of aspirin. There were no real differences between the results reported in each of the preliminary reports (95) (Steinberg, Tr. 5163-66). Prior to the initiation of the study, three scientific committees at the Veterans Hospital reviewed and approved the proposed investigation. Approval by the committees required a finding by them that the study had scientific merit (Rabinowitz, Tr. 3499). Dr. Rabinowitz maintained full independence in conducting the study. The other participants in the study were physicians ofthe University of Pennsylvania specializing in arthritis and rheumatology, who had no contact whatever with Thompson (Rabinowitz, Tr. 3495-97).

317. The Rabinowitz study involved "tagging" the TEAlS molecule Initial Decision 104 F. with a radioactive isotope. The TEAlS was made radioactive at the New England Nuclear company by using radioactive carbon dioxide (e02) and bubbling it with phenol to yield radioactive salicylic acid. The radioactive salicylic acid was then added to TEA to yield radioactive TEAlS (Rabinowitz, Tr. 3500-01, 3521). The salicylate molecule in the TEAlS compound was tagged with radioactive earbon 14 (14e). According to Dr. Rabinowitz, it is certain that the salicylate molecule retained the radioactive tag (Rabinowitz, Tr. 3503). There is no difference other than the radioactive carbon between the radioactive TEAlS used in the study and commercially available TEAlS (Rabinowitz, Tr. 3503-04). This fact was proven by use ofa nuclear magnetic resonance (Rabinowitz, Tr. 3504).

318. The first stage of the Rabinowitz study was a canine study conducted on beagles. Beagles are considered to be good models to test the absorption of salicylate (O'Brien, Tr. 3890; Rabinowitz, Tr. 3505- 06; Silverman, Tr. 2209). The study was conducted by rubbing Aspercreme into the shaved right knee of the dog until all of it was absorbed. After one hour, the area was wiped off with alcohol. The animal was sacrificed and the knee was cut off and sections were taken from the skin. Each section was weighed carefully and the tissue was extracted. Each section was then treated with either and sulfuric acid to extract the radioactive salicylate. The extract was chromatographed for further purification, and the amount of salicylate present was measured by assessing the amount of radioactivity with a radioactivity counter (Rabinowitz, Tr. 3505, 3507-09). The amount of radioactivity measured represented the amount ofconcentration of salicylate in the tissue (Rabinowitz, Tr. 3511-12). 319. The results of the canine portion ofthe study showed that the skin, the muscle, and the fascia absorbed significant quantities of salicylate in all five dogs in the group. The data also revealed that the salicylate level in the blood in the Aspercreme group was 10 to 100 times lower than that of the aspirin group. (96) 320. In the canine study conducted by Dr. Rabinowitz, the following salicylate levels were found in body tissues following oral and topical administration:

Oral Cream Muscle 38.20 1.6 Fascia 16.40 04 Fat Pad 00 5. Tendon 20 3. Cartilage 43 1. Synovium 62 .

These data reveal that TEAlS topical application resulted in higher .l.lH.nnr":R.Jlj luc.U.lvfil, VV. , .!'jv.

648 Initial Dccision local salicylate concentrations than did ingestion of oral aspirin in dogs and indicate that topical TEAlS was primarily absorbed locally by direct penetration (Rabinowitz, Tr. 3513-14; RX 70C-D). 321. The second stage of the study studied human patients with rheumatoid arthritis. All patients met American Rheumatism Association s criteria for classical or definite rheumatoid arthritis. Dr. Ralph Shumacher, Professor of Medicine at the University of Pennsylvania and a highly regarded rheumatologist, screened the patients for study eligibility. The same patients participated in both the oral aspirin and topical TEAlS parts ofthe study. Patients abstained from salicylates for six hours prior to each study period. Each patient first received orally 500 mjJigrams of e aspirin. Two to six weeks later ten gms. of the triethanolamine C- salicylate cream was massaged into the skin over one knee. Blood and urine samples were obtained before the administration of either the oral or the topical medication and again at 60 to 120 minutes, at which time a synovial fluid aspiration was performed. The fluid samples were extracted, chromatographed, and measured for radioactivity (Rabinowitz, Tr. 3509-10; RX 70).

322. In the human study conducted by Dr. Rabinowitz, salicylate concentrations in the synovial fluid after the application of TEAlS were found to be approximmately 60% of the concentration found after the oral ingestion of aspirin. However, the concentration in the blood from orally ingested aspirin was four to eight times higher than that resulting from topical application of TEAlS. These data indicate that the TEAlS was absorbed by direct penetration into the joint through the skin since these levels were achieved despite low blood salicylate levels (RX 70). (97) 323. The Rabinowitz study shows that salicylate can be and is absorbed through the skin in measurable amounts and that the salicylate component of the TEAlS molecule is capable of penetrating through the skin, muscles, and tendons right down to the joint connecting the bones. Thus, Aspercreme can deliver salicylate to joints and tissues (Adriani, Tr. 1298; Ehrlich, Tr. 4030; O'Brien, Tr. 3675; Rabinowitz, Tr. 3513-14; Roth, Tr. 1728/7-12; Silverman, Tr. 2208; RX 70).

324. Dr. Howard Maibach, professor of dermatology at the University of ealifornia Medical School and a recognized expert in the field of percutaneous absorption of drugs, observed that while topical drug absorption was generally believed to be dependent upon being transported through the blood or the general circulatory system (the "microcapillary network") (RX 289), recent studies by Dr. Maibach, Dr. Jean Paul Marty, and others have demonstrated that a drug may be capable of penetrating into the body without being carried by the Initial Decision 104 F. blood (RXs 1000-05). These studies also show that subcutaneous drug levels can be achieved following the penetration of a topically applied drug into the layers of the body (RXs 289B- , 1020). In Dr. Maibach' opinion, topically applied TEAlS is one of those drugs which diffuses into muscles and tissues beneath the skin (RX 289D). Dr. Maibach' paper, accepted for publication in the Journal of Pharmaceutical Sciences reviewed some of the scientific literature in the area of subcutaneous delivery of chemical substances, including the Rabinowitz study (RX 70), the early St. Thomas Institute study (RX 45), and the Golden study (RX 49) and concluded:

Better (or at least equivalent) therapy is possible, therefore, without systematic distribution of the drug, significant blood levels, and possible side-elects (RX 102M). 325. Dr. Rabinowitz' bioavailability study is interesting in that it shows that topically applied TEAlS may be capable of delivering salicylate to subcutaneous body tissues by directly penetrating the skin and thus offer an alternative method of administering salicylate for relief of pain. What is needed to verify this potential, however, is an acceptable demonstration of TEAlS' bioactivity. 326. The record also shows some ambiguity as to precisely what chemical was carbon- 14 tagged in the experiment itself TEAlS was obtained by adding radioactive salicylate acid to TEA. Although Dr. Rabinowitz testified that there was no (98) diflerence between the radioactive TEAlS he used and the commercially available TEAlS and that the TEA molecule in the radioactive TEAlS could not have been tagged in the process, the question as to which of the four differ- TEAlS or TEA)ent chemical entities (the salicylate ion, salicylic acid was measured in the radioisotope experiment remains in the record (Adriani, Tr. 1211 (salicylate ion, salicylic acid, or TEAlS); Brien Tr. 3873 (salicylate ion); Rabinowitz, Tr. 3500-01 (salicylic acid); Roth Tr. 1599-600 (TEA or salicylate)). There also is some evidence indicating that the sulfuric acid used to extract the tagged material from the samples could have caused the TEAlS to disassociate into TEA and a salicylate moiety (Adriani, Tr. 1209; Rabinowitz, Tr. 3536). 327. Thompson has also admitted that the human portion of the Rabinowitz experiment did not study whether TEAlS broke down into a salicylate ion in the body and that the canine portion did not study whether TEAlS broke down into salicylate in the body (eX 45Q (Admissions Nos. 354-55)).

328. The Rabinowitz data do not show that topically applied TEAlS in the experiment penetrated below the skin in therapeutically significant quantities (Adriani, Tr. 1263; Roth, Tr. 1599-600, 1724). Rather the studv shows that most ofthe ta!!!!ed material staved on the skin n1U1Vlr:)Ul lVlr.Uli.AL I.U. , U\jL-.

648 Initial Decision (Adriani, Tr. 1263; RX 70e (Table 1)) and that the base was actually better absorbed than the tagged material (Rabinowitz, Tr. 3611-12; RX 70e). Dr. Roth, complaint counsel' s expert witness, testified that the clinical relevance of the presence of the tagged material in any ofthe tissue or fluid samples is dubious (Roth, Tr. 1728). Furthermore the record fails to answer many important questions that would have to be addressed. The Rabinowitz study does not show the rate at which the tagged material penetrated. It has been demonstrated, however that if an analgesic is slowly absorbed, the minimum effective concentration ofthe drug at the site of action may never be reached (O'Brien Tr. 3879-83).

329. Furthermore, there are some significant problems in the experiment' s design. The dosage of radioactive aspirin used in the experiment (500 miligrams) is lower than the recommended single dose of aspirin for analgesia (650 millgrams) and is far lower than the recommended dose of aspirin for inflammation (4200 miligrams a day) (O'Brien, Tr. 3868; Rabinowitz, Tr. 3534; Roth, Tr. 1602). In contrast, Dr. Rabinowitz added 20% more radioactive molecules to the TEAlS in order to compensate for the TEAlS that would remain on the glove of the physicians who applied the cream (Rabinowitz, Tr. 3502). Also, although all six human subjects had been on a stable dose of oral aspirin for six months or more, they were instructed not to take aspirin only for six hours before each test period (99) (RX 70D). More than six hours, however, would be necessary for the nonradioactive aspirin to be totally cleared from the subjects' systems (Adriani, Tr. 1395; O'Brien, Tr. 3765-67; Roth, Tr. 1600). To the extent the nonradioactive aspirin remained in the body, it may have influenced the test results by diminishing the radioactive aspirin that could have been absorbed.

330. As part of its reasonable basis, Thompson relies on submitted documents (RX 42/eX 202; RX 62/eX 216) pertaining to blood and urine level tests on human volunteers. In both studies, topical TEAlS was applied and blood and urine samples were measured for salicylate at fixed intervals. The only scientific value of these studies is to demonstrate that topically applied TEAlS is absorbed into the blood and excreted in the urine (Adriani, Tr. 1178). Almost any drug applied to the skin will show up in minute traces in the blood. But, the serum levels of salicylate achieved did not reach the minimum levels associated with analgesia (Adriani, Tr. 1390; Roth, Tr. 1760-61). Moreover, in one test (RX 62), the site where the drug was applied was covered with saran wrap overnight. covering the site of a topical application increases the rate of absorption. Thus, the resulting blood and urine levels of salicylate were substantia1Jy higher than they Initial Decision 104 F. would have been after normal consumer use (Adriani, Tr. 1229; Silverman, Tr. 2428-29).

331. In addition, Thompson has submitted two rabbit studies (RX 43/CX 215; RX 44/eX 204) measuring salicylate levels in rabbits after application oftopical TEAlS. The scientific value ofthese tests, which involved a total of five rabbits, is questionable. In any event, for the reasons set forth hereinabove, these so-called penetration studies do not provide substantiation for- the effcacy of topical TEAlS in humans for relief of musculoskeletal pain.

332. The Myoflex bioavailability studies Thompson refers to (eX 344G; Z-48-49; Z-063-65) simply show that salicylate was present in the blood following a topical application of Myoflex, a topical rub containing 10% TEAlS. The egg membrane experiment (RPF 202) and beefmuscule experiment (RPF 203) referred to by Thompson are trivial and oflitUe value to this proceeding. 333. Dr. Steinberg, Thompson s vice president who is responsible for substantiation of advertising claims, testified that Thompson relied on a 1955 article by Howell (Steinberg, Tr. 5175-76; RX 366/eX 366 pp. Z-222-23). Since Thompson acquired the Howell article toward the end of 1981, Thompson could not have relied on it for substantiation of any claims for Aspercreme made before that time. In any event, the Howell (100) article, which reports the results of a singleblind English study on diethylamine, cannot be used to substantiate claims for TEAlS which is an entirely different drug that may not penetrate the skin in the same way.

F. The Gaudin Patent 334. Thompson also relies on certain patents as evidence of drug effcacy. A United States patent was issued in 1952 to Dr. Olivier Gaudin for his discovery (No. 2 596 674) oftopical absorption of amine salicylates (RX 450lex 212). It does not refer to TEAlS but to an entirely different compound, diethylamine salicylate (Adriani, Tr. 1288, 1243-44). Other than the patent holder s assertion that dissociation of diethylamine salicylate occurs, there is no other evidence of dissociation in the patent report (eX 212e). In addition, other patents have been issued by the United States Patent Offce for different topical salicylate salts (Silverman, Tr. 2227-28; RX 451). These patents reflect a determination that some salicylate salts as described were patentable within the meaning ofthe patent laws, and the medical scientific community does not use or accept patents as a source of information on drug effcacy (Adriani, Tr. 1227). Patents are merely descriptive claims for compounds, and have no scientific significance for determining drug effcacy for any disease condition (Roth, Tr. 1609).

THOMPSON MEDICAL CO., INC. Iv'! 648 Initial Decision G. Clinical Observations And Opinions Of Physicians 335. It is natural that in the practice of medicine, the determination of what drug is most appropriate for an individual patient rests with the physician s professional judgment based on the patient's history and disease state and the physician s knowledge of and experience with drugs. However, the practice of medicine is not an exact science but an art and a physician s choice of a drug for a particular patient essentially reflects a process of trial and error based on long experience, insight and wisdom. However, it is something else to argue that clinicians' experience with patients with a particular drug should be accepted as scientific proof of effcacy. It is generally recognized by the medical scientific community that physicians' observations and opinions may suggest or lead to controlled clinical trials or be used to augment such trials, but they are not substitutes for well-controlled clinical trials for the purpose of showing drug effcacy (Adriani, Tr. 1436, 1460-61; (101) Roth, Tr. 1570). The contrary view expressed by respondent' s experts, to the effect that the requirement for well-controlled clinical trials should be substantially relaxed or dispensed with in the case of OTe topical analgesic drugs, such as Aspercreme do not reflect the prevailing view ofthe medical scientific community including the FDA. See 21 C.F.R 330.1O(a)(4)(ii); FDA OTC Drug Review Policy Statement, 46 FR 47 729, 47 731 (1979). 336. A physician s observation of his patient's response to an analgesic may be affected by bias (Adriani, Tr. 1181) or may be incorrect due to a variety of other factors, such as placebo effect (Ehrlich Tr. 4155-56). For example, the enthusiasm the physician may consciously or unconsciously communicate to the patient may contribute to a high placebo response rate (Ehrlich, Tr. 4133-36). Moreover there have been numerous instances where drugs used over a period of years with positive consumer response and physicians' observations were later subjected to clinical tests and found to be ineffective (Adriani, Tr. 1180-83; Ehrlich, Tr. 4117-18; O'Brien, Tr. 3775-76; Roth, Tr. 1571).

337. Physicians' observations and opinions based upon case reports random experiences, and other reports lacking details necessary for scientific evaluation do not constitute adequate substantiation for Aspercreme s efficacy claims (F. 335-36 supra). 21 e. 330.10(a)(4)(ii). Thus, for example, letters such as RX 47 ICX 260 which represent isolated, uncontrolled, and undocumented observations of physicians do not constitute scientific evidence (ld. See Adriani, Tr. 1230-32; Golden, Tr. 2763- , 2767; Roth, Tr. 1609-12). On the other hand, where appropriately documented and systematic, physicians observations and opinions may constitute reports of significant Initial Decision 104 YT. human experience during marketing and may thus be viewed as evidence capable of corroborating, but not supplanting, clinical studies (F. 218 supra). 21 e. R. 330. 10(a)(4)ii). 338. In 1982, B.F. Ascher & eo., Inc. ("Ascher ), the marketer of Mobisyl, another 10% TEAlS product, conducted a survey of certain physicians and health practitioners in an effort to determine their opinions of Mobisyl (Borchers, Tr. 5045-56; RX 346A-e). The study was submitted to the FDA in an effort to corroborate the results of clinical tests undertaken by Thompson Medical (Borchers, Tr. 5057- 58; RX 346AJ. The submission to the FDA was made pursuant to 21 R. 330.10(a)(4)(ii), which provides that proof of effectiveness shall consist of controlled clinical investigations, which may be corroborated, inter alia, by reports of significant human experience during marketing (Borchers, Tr. 5058; RX 346A-B). (102) 339. Because the Ascher survey was neither possessed nor relied upon by Thompson prior to its dissemination of any ofthe challenged (See CX 25). advertising, it does not aid Thompson in this proceeding In any event, by virtue of the manner in which this survey was conceived and conducted, it is of little value in this proceeding even as corroborative evidence of TEAlS' effcacy (See ePF 395-99). H. Testimonial Eoidence of Users 340. A significant portion of Thompson s substantiation materials is devoted to testimonial evidence, both by consumers and by persons in health-related occupations. It is well-recognized that such testimonial evidence has no value in determining the issue of drug effcacy (Adriani, Tr. 1180; Roth, Tr. 1567). In FDA regulations and the OTC panel evaluations of analgesic drug effcacy, patient testimonials were not worthy of consideration (Adriani, Tr. 1239; ex 391F, 395B). 21 C. R. 330.10(a)(4)(ii). Since consumers are not incapable of evaluating drug effcacy, testimonial evidence is not a reliable source of evidence of drug effcacy (Adriani, Tr. 1239; Roth, Tr. 1617). 341. Thompson s own witnesses acknowledged the inadequacies of testir.lOnial evidence as a basis for demonstrating drug effcacy. For example, Dr. O'Brien admitted that he had criticized pharmaceutical companies for their reliance on testimonials as evidence that Indomethacin (a prescription anti-inflammatory drug) is efiective (O'Brien Tr. 3786). He had also criticized the companies for their reliance on consumers' reports as to the type of pain reliefthey got from the drug (O' Brien, Tr. 3786-87). Dr. Ehrlich indicated that in dealing with consumers' reports, there is no way to eliminate the possibility that they were due to the placebo eflect (Ehrlich, Tr. 4155-56). Consumer letters to companies also suffer from selectivity in that a company only hears from those who want to he Hn ( lrll 'rhl1 for PYJlmnlp \ .

TtlU1VU' ::Ul-. lVl 1Jlvf\L LV.

648 Initial Decision a nurse, who had returned a consumer response card favorable to Aspercreme and subsequently testified at trial, acknowledged that she probably would not have written a negative letter to a company even if she were "downright disappointed" in a product (Walsh, Tr. 4362).

342. The Arthritis Foundation has devoted significant attention to the area of testimonials because it relates to the problem of unproven remedies (Roth, Tr. 1610). Thompson s expert, Dr. O'Brien, acknowledged that the Arthritis Foundation is a reliable source of information about the treatment of arthritis and the scientific issues surrounding the disease, and that the Foundation takes the position that testimonials and (103) case histories cannot be relied on to show that a remedy works for the diseases of rheumatoid arthritis and osteoarthritis (O'Brien, Tr. 3919- , 3925). The reason the Foundation takes this position as regards case histories is because arthritis has peaks and valleys and may go away by itself just when a person tries a new remedy (O'Brien, Tr. 3926). For this reason, the Foundation takes the position that controlled trials of drugs and remedies are employed to determine safety and effcacy while discounting the placebo effect and other sources of bias. Such controlled clinical trials are acceptable as scientific proof (ldJ 343. The unreliability of Thompson s testimonial evidence is evident from a sampling ofthe testimonial letters it received. For example, one letter is from a chiropractor who used Aspercreme on himself and thought it was good. He also indicated that he was using it on his patients in conjunction with ultrasound treatment. Reports such as these do not constitute valid scientific evidence (Adriani, Tr. 1232; Roth, Tr. 1611-12). Another letter stated that Aspercreme was effective in a case of stroke paralysis. There is no topically-applied medication that would be effcacious for stroke victims, and this type of testimonial may be aptly compared to faith healing (Adriani, Tr. 1233; Roth, Tr. 1612). Still another letter submitted by Thompson as substantiation for its claims is a "Dear Doctor" promotional letter sent out by another pharmaceutical company, B.F. Ascher & eo., to physicians introducing a new oral product for arthritis treatment (Adriani Tr. 1234-35). The letter discusses a topically-applied TEAlS cream as an adjunct to oral therapy for arthritis. This document does not constitute proof of Aspercreme s effcacy (Adriani, Tr. 1235; Roth, Tr. 1612-13).

I. Drug Compendia And General Scientific Literature As Proof Of Drug Efficacy 344. Thompson has also relied on the inclusion of TEAlS in certain compendia of drug products a" an indication of Aspercreme s drug Initial Decision 104 F. effcacy. However TEAlS is not included in any of the authoritative reference works on drugs. The publications in which TEAlS is listed are listings of marketed drugs. Importantly, there are only a few published reports on TEAlS in the medical or scientific literature. 345. For example, there are no listings for Aspercreme or TEAlS in four of the most authoritative drug compendia: the US. Pharmacopoeia (O' Brien, Tr. 3716; Silverman, Tr. 2384; Steinberg, Tr. 5229); the National Formulary (O' Brien, Tr. 3716; (104) Silverman Tr. 2386); Remington s; and Goodman and Gilman s book The Pharmacological Basis of Therapeutics (O' Brien, Tr. 3718; Steinberg, Tr. 5227-29). The omission of TEAlS products from these standard reference works is significant. A drug has to be recognized as effcacious in order to be listed in the US. Pharmacopoeia an offcial standard reference work (Silverman, Tr. 2378- , 2387). Remington is also considered to be an authoritative treatise on drugs, while Goodman and Gilman s book is widely regarded as a major reference work on drug effcacy and drug action (Steinberg, Tr. 5227). Dr. Steinberg of Thompson was well aware that a number of authoritative United States treatises did not include TEAlS (Steinberg, Tr. 5229). And Dr. Silverman, Thompson s pharmaceutical expert, agreed that in ex 393, the FDA Panel on OTC Skin Protectant Products for Human Use only drugs that were listed in standard texts were placed in eategory 1 absent clinical trials (Silverman, Tr. 2391-94). 346. Aspercreme is listed in the Handbook of Nonprescription Drugs. Dr. O'Brien, an expert witness for Thompson, acknowledged that he would not rely on the Handbook to determine whether or not an analgesic product is effective (O'Brien, Tr. 3903). Dr. Silverman conceded that the sixth edition of the Handbook of Nonprescription Drugs repeats the findings of the FDA Panel on OTe External Analgesic Products about TEAlS (Silverman, Tr. 242G-21). 347. Aspercreme is also listed in the Physicians ' Desk Reference on Nonprescription Drugs. Dr. O'Brien agreed that the Physicians' Desk Reference ("PDR") is not an authoritative source of drug information. Typically, much of the book consists of excerpts from package inserts and no one would look to it as an academic source of information (O' Brien, Tr. 3905).

348. Three other books in which TEAlS products are listed, namely the American Drug Index, Facts and Comparisons and Perry Pre. scription and Nonprescription Drugs, are all-inclusive indexes or lists which purport to include all drugs marketed in the United States (Silverman, Tr. 2379 , 2384; Steinberg, Tr. 5226). Neither the American Drug Index nor Facts and Comparisons are compendia in the p t, 1- fhp TT, Phormnr' nnno;n r.cihrprm!'n 'rr 7R-7Q ? Inv1Vll",Vl lVl.rUlLi\L LV., H'\L. limit 648 Initial Decision 82). Unlike the Us. Pharmacopoeia they do not constitute offcial standard reference works on drugs (Silverman, Tr. 2378-79). 349. With respect to published reports about TEAlS in the medical! scientific literature, Thompson s expert witness Dr. Silverman testified that although he conducts a literature search specifically on TEAlS twice a year, he has never seen any articles recommending its use as an analgesic other than those (105) written by Drs. Golden (RX 48/eX 200) and Rabinowitz (RX 70lex 374) (Silverman, Tr. 2347). Dr. Adriani testified that he was not aware of published reports in the literature about TEAlS (Adriani, Tr. 1434). And Dr. Roth indicated that prior to his participation in the instant case, he had never seen anything on TEAlS in the medical literature. Dr. Roth also testified that he had never heard of TEAlS being the subject of a paper at a professional meeting (Roth, Tr. 1761).

J. The Pharmacology of Triethanolamine Salicylate (TEAlS) And Its Mechanism Of Action 350. Much ofthe record information pertaining to the pharmacology and mechanism of action of TEAlS is based on the testimony of Dr. Silverman, Thompson s pharmaceutical expert witness. Respondent' theory, as further elaborated in this case, is essentially that Aspercreme delivers salicylate molecules to the subcutaneous tissues by direct penetration of the skin, and provides pain relief in the site of pain by inhibiting prostaglandin synthesis in the cell (eX 45N (Admission No. 274)). Although Dr. Silverman s testimony in this regard was not directly contradicted or rebutted by other expert testimony, the record is clear that Dr. Silverman s hypothesis is a novel one and was expounded publicly for the first time in this proceeding. The Silverman hypothesis remains to be accepted by the medical scientific community. It also leaves too many important questions unanswered and is inconsistent in some important respects. In any event, theories regarding a drug s mechanism of action are important and useful but they are not substitutes for well-controlled clinicals for the purposes of showing drug effcacy.

351. The active ingredient of Aspercreme is triethanolamine salicylate (TEAlS). TEAl IS is manufactured by combining equal amounts of triethanolamine and salicylic acid. The resulting compound has a relatively low molecular weight of 280 (Silverman, Tr. 2113). According to Thompson s pharmaceutical expert, this low molecular weight helps the TEAlS molecule to be absorbed through the skin (Silverman, Tr. 2114). Radioisotope testing and bioavailability studies in this record have suggested that some topically applied drugs can penetrate the skin.

352. The human skin is constructed of several layers of cells. There _ : _ .... ; ,.. Initial Decision 101 F. are five layers on the stratum corneum and two on the diadermis. Even though the cellular layers are differentiated in function, they have one common characteristic: (106) they consist of cells which have a cell membrane on the outside and protoplasm on the inside. The cell membrane consists of both lipid (fat) and water the cell is phospholipid (Silverman, Tr. 2140-1). Thus, for a drug to pass through the cell membrane, it needs to have both water solubility and lipid solubility (Silverman, Tr. 2117, 2141). According to Dr. Silverman TEAlS has very good water solubility and some lipid solubility (Silverman Tr. 2115). Molecules which have a solubilty ratio of one to one 50% in water and 50% in lipid, have the best ability to penetrate the biological cell membrane. According to Dr. Silverman TEAlS has 60/40 solubility ratio (60% in water and 40% in lipid) and penetrates the cell membrane well (Silverman, Tr. 2118). 353. The outer layer of the skin consists ofa dead layer of dry cells. According to Dr. Silverman, when Aspercreme is applied, these dry cells are hydrated by the oil in the product, and penetration is facilitated. According to respondent's experts, the TEAlS molecule dissociates into a triethanolamine ("TEA") molecule and a salicylate ("SA" molecule in the presence of water (O'Brien, Tr. 3938; Silverman, Tr. 2125; RX 1013). This fact can be demonstrated by either of two tests: a Beckman thermometer test (which measures temperature differences between a freezing point of the molecule and a freezing point of the dissociated mix) or an osmometer test (which measures the rapidity and depth of electrical signals to determine how many particles are in solution (Silverman, Tr. 2125-26). Dr. Silverman hypothesized as follows: when Aspercreme is placed on the skin, the waxes in the product's vehicle (triethanolamine) soften the dry outer layers of the skin creating an occlusive efiect which hampers the evaporation of water from the skin (Silverman, Tr. 2146), the TEAlS molecule starts to dissociate or ionize into its components, a TEA ion and a SA ion, and TEA and SA ions are very water soluble and wil pass through the skin slowly (Silverman, Tr. 2127- , 2146-7). According to the Silverman hypothesis, in addition to ionization, another chemical process, hydrolysis, is taking place. As the TEA ion comes in contact with water in the skin, the TEA ion reacts with the hydroxyl ion of water (i. - OH) and reverts back to the TEA molecule. The TEA molecule, which has biphasic solubility (solubility in both water and lipid), penetrates the skin (Silverman, Tr. 2128-29). As the SA ion comes in contact with the water in skin, the SA ion reacts with the hydrogen ion of water (i. H +) to form salicylic acid. According to Dr. Silverman, the salicylic acid molecule has good lipid solubility with some degree of water solubility and penetrates through the skin. 1 - 1. - rnt:,l lco 1- ..1-- 'jHUMI' UN Mr;UICAL CU., INC.

648 Initial Decision reactions called ionization and hydrolysis occur so that the TEAlS molecule dissociates and recombines as needed to pass through the membranes. These (107) reactions establish a constant topical reservoir in equilibrium that provides TEA molecules and SA molecules which penetrate the skin and gradually migrate through the epidermis and diadermis to the underlying tissues of the body (Silverman Tr. 2132- , 2136-37, 2146-7). Migration is possible because the outer membrane of the cell wall is phospholipid (i. both lipid and water in nature), while the interior ofthe cell is largely protoplasm with inorganic salts (and hence largely water in nature). Penetration of a cell by the molecule involves movement through the lipid barrier of the cell membrane and movement through the water barrier inside the cell (Silverman, Tr. 2150). According to Dr. Silverman, because the TEAlS molecule has both water and lipid solubility, it will "percolate" its way through the various layers of the skin, connective tissues, and muscles to the bone (Silverman, Tr. 2152- , 2165). 354. Penetration of a drug through the skin is enhanced if a drug is soluble in both water and lipid and if it has a molecular weight of less than 1 000 (Adriani, Tr. 1294; ex 269, p. 69 774). Penetration through the skin is also enhanced if the skin is damaged (Adriani, Tr. 1291-92; ex 269, p. 69 774). Damaged skin is skin in which the stratum corneum remains intact, but there is edema (fluid retention), inflammation, or other pathological processes present in the lower layers of the skin as a result of an injury or a disease (Adriani, Tr. 1293; ex 269, p. 69 773). According to Dr. Silverman, where there is inflammation present, as there is in arthritis or rheumatism, penetration is increased (Silverman, Tr. 2167, 2169). And drug absorption is further facilitated if the substance is rubbed or massaged into the affected area (Adriani, Tr. 1295; Silverman, Tr. 2169, 2176). 355. Inflammation is characterized by heat, redness, swelling, and tenderness in the affected tissues (eX 269, pp. 69 777-78). The salicylate ion exerts an anti-inflammatory effect (Roth, Tr. 1658; Silverman, Tr. 2486-87; CX 269, p. 69 778). It has generally been hypothesized ofJate that the salicylate ion achieves this anti-inflammatory effect by interfering with the biosynthesis of prostaglandins (PGs) at the cellular level. Prostaglandins are complex, hormone-like molecules which are synthesized from arachidonic acid which is present in the body. Prostaglandins EI and F2 have been shown to be capable of producing local inflammation. Trauma to the body causes the cells to produce PGs. PGs E, and F2 cause pain and inflammation by intensifying the pain producing properties of certain compounds within the body (Stipulated testimony of Dr. Ehrlich with respect to medical literature, Tr. 4006; RXs 1014-15). Thus, in theory aspirin and other salicylates can be useful in interfering with the develop- Initial Decision 104 F. ment 01'108) prostaglandins from arachidonic acid (Ehrlich, Tr. 4005; ex 269, pp. 69 777-78; RX 1015).

356. Respondent's medical experts testified that Aspercreme through its TEAlS component, achieves its analgesic and anti-inflammatory action by penetrating through the skin to the underlying tissues and working at a cellular level, through its salicylate ion, to inhibit the formation of PGs (Ehrlich, Tr. 4008-9; Heller, Tr. 2612- 13).

357. It is diffcult if not impossible to determine at what blood (serum) level the salicylate ion becomes an effective analgesic agent. The amount of salicylate in the blood does not correlate to clinical analgesia (Adriani, Tr. 1290; CX 268, p. 35 382). 358. Therapeutic serum levels of salicylate differ from the levels of salicylate at the local site (Roth, Tr. 1688). Therefore, the pain-relieving effectiveness of an analgesic agent cannot be measured by analysis of the salicylate serum level (Roth, Tr. 1688). According to Dr. Silverman, because aspirin is ingested orally and must first circulate through the blood stream before reaching the affected site, the action of oral aspirin in relieving pain is slower than the action of topically applied salicylate (Silverman, Tr. 2174; RX 49). 359. Theories regarding a drug s mechanism of action are not a substitute for clinical testing for purposes of demonstrating drug efficacy. Thus, attempts to evaluate a drug s mechanism of action are generally made after the drug s effcacy has been established clinically (Ehrlich, Tr. 4008-10; Roth, Tr. 1613-15). In any event, even on a purely theoretical basis Thompson s theories of Aspercreme s action leave too many questions unanswered. The record also shows many inconsistencies with respect to the assumptions about salicylates TEAlS and prostaglandins upon which the theories essentiaUy rest. 360. As to salicylates, the Thompson s theory apparently is founded on the tenet that topical TEAlS arrives in the muscle or other point of pain as salicylate and that the salicylate in TEAlS is the same as the salicylate in aspirin and wil therefore provide relief in the same way as oral aspirin does (Steinberg, Tr. 5131). Aspirin and TEAlS are not the same drug: aspirin is a salicylate to which an acetyl group has been added; TEAlS is a milder nonacetylated salicylate (O'Brien, Tr. 3729 3877-78; Roth, Tr. 1516-17). The assumption that aU salicylates are the same is untenable given that no one reaIJy knows how analgesics work and that the metabolism of aspirin in the body is highly complex (Ehrlich, Tr. 4047-48; O'Brien, (109) Tr. 3877). Currently, there are at least two schools of thought on aspirin s action. One is that the salicylate molecule itself is anti-inflammatory, and that the acetyl moiety is merely a means of delivering the salicylate. The other C' tho:t t"hp o;("ptuh,:.t-inn il"l"tnrprQihlv ,.ptvl h:,,;:: fhp t.plpt t.n ::rp THOMPSON MEDICAL CO. , INC. IQ" 648 Initial Decision part of the inflammatory process and this is essential for fuJJ relief (O' Brien, Tr. 3879; Roth, Tr. 1658-59). Moreover, there is a body of opinion to the effect that nonacetylated salicylates do not inhibit prostaglandin synthetase (Roth, Tr. 1675). Regardless of how either aspirin or TEAlS may work, there is also support for the proposition that aspirin and salicylates other than methyl salicylate are not effective as topical analgesics (Adriani, Tr. 1480). 361. Disregarding the chemical differences between TEAlS and aspirin, there are stil unanswered questions surrounding the "site of action" component of Thompson s theory. Although theories as to how aspirin works are continuously evolving, the prevailing view is that in producing an analgesic effect, aspirin acts peripheraJJy, or locally, as well as centrally on the central nervous system (Ehrlich Tr. 4057-58; Roth, Tr. 1655-56; CX 268, pp. 35 351, 35 381). When Thompson acquired Aspercreme from Sperti in 1976, there were two theories about Aspercreme s mechanism of action. One was that topical TEAlS exerted an analgesic effect by achieving therapeutic levels in the bloodstream. The other was that TEAlS worked by penetrating directly to the point of pain (Steinberg, Tr. 5261 62; RX 41/eX 251Z- 058). For purposes of this proceeding; Thompson has adopted the position that topically applied TEAlS acts 10caJJy "at the point of pain " (eX 45N (Admission No. 274)). Topical TEAlS unlike aspirin is not a systemic drug and thus would not provide whatever relief is produced by the central nervous system ("ens") effect of aspirin. In addition TEAlS could do litte to modify a systemic process like inflammation, which is a major factor in many arthritic conditions (Roth, Tr. 1536, 1757).

362. In addition TEAlS is a relatively obscure drug. There is little medical literature on TEAlS and it is not included in any authoritative drug treatise (O'Brien, Tr. 3718-18; Roth, Tr. 1761; Silverman, Tr. 2347, 2385-86). An expert advisory panel to the FDA has approved TEAlS in concentrations of 5% to 12%, as an OTe drug for only one use, as a mild sunscreen (Roth, Tr. 1684; ex 394B, H). It is the salicylate, not the TEA , in TEAlS that acts as a sunscreen (Roth, Tr. 1684; ex 394B (triethanolamine salicylate listed as an active sunscreen ingredient, triethanolamine listed as inactive)). This would suggest that most of the salicylate in TEAlS remains in the dermis, rather than penetrating into the deeper tissues, (110) since all sunscreens work in this fashion (Adriani, Tr. 1458; Roth, Tr. 1684). 363. Another problem in Thompson s theory of Aspercreme s effcacy is the assumption that TEAlS works by inhibiting prostaglandin synthetase. Prostaglandins are enzymes found throughout the body, some of which are now thought to be implicated in the inflammatory process (Adriani, Tr. 1288-87; Roth, Tr. 1651-52). However, there is Initial Decision 104 F. no evidence in this proceeding showing that TEAlS blocks prostaglandin synthetase. Moreover, blocking prostaglandin formation is not the only way of curbing the inflammatory process, since several mechanisms are known to be involved in inflammation (Ehrlich, Tr. 4049- 51; Roth, Tr. 1745-46). Furthermore, prostaglandin theory is an evolving theory and not alj prostaglandins have been discovered. Some known prostaglandins, which were initially thought to playa central role in the inflammatory process, are now considered to be less important as new prostaglandins are discovered (Ehrlich, Tr. 4049-50; Roth Tr. 1755-57). In addition, no one knows which prostaglandins are involved in arthritis (Adriani, Tr. 1445; Ehrlich, Tr. 3996-97). Finally, prostaglandins are only one of a series of inflammatory mechanisms studied over the years (Roth, Tr. 1613-15). In any event, the requirement for well-controlled clinical trials remain unaffected by any theory of the inflammatory process (Adriani, Tr. 1446; Ehrlich, Tr. 4008-10; Roth, Tr. 1624).

364. From all of the foregoing, it is found that Thompson s clinical trials, bioavailability studies, and theories about TEAlS' mechanism of action fail to provide an acceptable level of scientific support for the claim that Aspercreme acts by penetrating through the skin to the site of arthritic disorders. On the contrary, the failure to show significant bioactivity in controlled clinicals suggests that TEAlS in Aspercreme is not absorbed in amounts necessary to demonstrate its analgesic effcacy (Roth, Tr. 1574).

K. Marketing Data Related To Aspercreme 365. In support of its claims of Aspercreme s effcacy, Thompson relies on various marketing-related data, including Aspercreme package insert cards mailed in by purchasers of the product, a survey of pharmacists, information on repeat purchases, and unsolicited consumer letters received over the years. According to the FDA Panel Report on External Analgesic Products, as well as expert opinion in this case, marketing experience related to an OTe analgesic product is at best a corroborative or confirmatory type of evidence, and under the (111) prevailing, scientifically accepted principles, it does not constitute the direct evidence or primary evidence needed to prove drug effcacy in the first instance (Adriani, Tr. 1433-34; Ehrlich, Tr. 4155-56; Roth, Tr. 1764; ex 269, p. 69 780). The FDA regulations governing the advisory panel OTe drug review process specifically provides that reports of significant marketing experience are appropriate only as a source of corroboration for proof of eflectiveness, and that isolated case reports, random experiences, and reports of product effcacy lacking the details which permit scientific evaluation are not n hp rplipi! nn in rlptprmininD' nrn rll1"t pftprtivpnp.;,; ')1 It1Ular;:Ul IVl.tUlLAL lA). Ith 648 Initial Decision 330.10(a)(4)(ii). Therefore, marketing experience by itself cannot be regarded as constituting adequate proof of drug efficacy (eX 395BJ. 366. Marketing data includes information on the number of units of the product sold to consumers and the number of consumer complaints received by the manufacturer (Adriani, Tr. 1346-7). Marketing data can indicate wide consumer acceptance of a drug (Adriani Tr. 1345). And while not a primary source of information on a drug effcacy (Roth, Tr. 1704), postmarketing data can provide a level of important information related to product safety-primarily longterm toxicity and idiosyncratic reactions resulting from product usage (Roth, Tr. 1709-10).

1. Consumer Response eards And consumer Letters 367. Evidence relied on by Thompson includes various data showing consumer satisfaction with the product, such as the following: (i) consumer response cards (package insert cards): Beginning in 1978, Thompson Medical included response cards in Aspercreme packages which were to be filled out and returned by the user. While there have been various forms of response cards which asked somewhat different questions, consumers were generally asked their general opinion of Aspercreme, particularly whether it worked better than aspirin. By 1982, some 30 000 response cards had been returned to Thompson. Two different tabulations of various groups of response cards were prepared. The results generally indicated that the consumers who reiurned the response cards had a favorable opinion towards Aspercreme (RXs 292, 521, 711). (112) (ii) Consumer letters: When Thompson acquired Aspercreme from the Sperti Drug company in 1976, all consumer letters which had been received by Sperti were turned over to Thompson (Siegal, Tr. 4599). Since then, Thompson has continued to receive consumer letters regarding Aspercreme (Siegal, Tr. 4598-99). eurrently, Thompson s fie of consumer letters contains almost 800 letters, most of which comment favorably on Aspercreme s efficacy (Siegal, Tr. 4603- 04).

(iii) Thompson Medical has received approximately 3 400 requests for refunds since 1978 (RX 94).

368. The consumer response cards and letters represent only small fraction (about 2%) of the Aspercreme purchases (Siegal, Tr. 4666; Silver, Tr. 5884, 5886). Such a low response rate is considered unacceptable for a survey because the respondents could not be considered representative in any meaningful sense (See Silver, Tr. 5883- 84). This is particularly true with respect to the response cards and letters because, unlike a survey, these consumers were completely ;.. , , g., Initial Decision 104 F. self-selected. It is also clear that those who sent in the response cards and letters differ substantially from those who did not in that the former feel much more strongly about Aspercreme than the latter (Ross, Tr. 6449-50).

2. Pharmacy Times Survey (RX 143) 369. RX 143 OTe Products The Pharmacist Recommends And Why, " is a 1981 survey of pharmacists by The Pharmacy Times offered to support Aspercreme s effcacy claims. The Pharmacy Times is a marketing publication for the pharmacy trade and is distributed free of charge to pharmacists and pharmaceutical houses. Its costs are borne by its advertisers. The Pharmacy Times purports to provide marketing information in the field of pharmacy (Reis, Tr. 5488). The Pharmacy Times survey of OTe drug products is conducted on an alternate year basis (Reis, Tr. 5492). RX 143, the report ofthe results ofthe survey, was first published in The Pharmacy Times in approximately late spring of 1981 (Reis, Tr. 5493). The survey was conducted by mailing questionnaires to a selected group from among the retail trade portion of The Pharmacy Times mailing list (Reis, Tr. 5493-94). The survey questionnaires were mailed concurrently with a request for verification of address, as part (113) of The Pharmacy Times audit of circulation conducted once every three years (Reis, Tr. 5495, 5498- 99). A total of 3 000 questionnaires were mailed out, and there were some 807 completed returns at the time the report was prepared (RX 143e). The survey questionnaire first asked respondents to estimate the number of recommendations they or their employees make for all brand name products in each category of OTe products each month. Secondly, the questionnaire asked which single product within each category the respondent would recommend by brand name (Reis, Tr. 5511-12; RX 143).

370. The Pharmacy Times survey is thus essentially a marketing survey. The results show that only 11.5% of the respondents recommended Aspercreme, and about 70% of the respondents made no recommendations for topical analgesics or did not recommend a TEAl based product (RX 143). In any event, pharmacists' recommendations are substantially influenced by advertising and other promotional activities (Ross, Tr. 5541, 6404-05). Studies in the literature have demonstrated the influence of advertising on physicians and dentists (Ross, Tr. 6405). The effect of advertising on consumer experience has been noted by FDA expert advisory panels (See, e. 396B). In the present case, Thompson s trade ads, such as RX 166E (a trade ad to druggists), made strong express claims that Aspercreme effectiveness was clinically proven, and proven better than aspirin in 1;.. dll r1;t:,, (Roc;c; '"r hdOfL(7) nr Rot.h t.p lfip,rl tnat a survev of 648 Initial Decision pharmacist recommendations of pain relief products is not reliable evidence that the product in fact relieves pain (Roth, Tr. 1706-07). Dr. Brien testified that he does not rely on retail pharmacists as a source of information on the effcacy of OTe drug products (O'Brien Tr. 3906-07). Since pharmacists' recommendations are based on a number of factors other than product effcacy, they do not provide reliable evidence ofthe effcacy of Aspercreme. The survey (RX 143) also suffers from a number of methodological problems (SeeePF 435- 38). It is of litte value in this proceeding. 371. In any event, the 1981 Pharmacy Times survey (RX 143) was not published until late spring of1981 (Reis, Tr. 5493). Thompson did not possess and rely on RX 143 when those advertising claims were made (See CX 25), nor did it cite or rely on this survey as part of its reasonable basis materials (See CX 44A). L. The Significance of Other Clinical Trials Showing Negative Results For 10% TEAlS Topical Products 372. Just as the clinical trials which failed to show significant dillerences between TEAlS and aspirin do not prove (114) that the two are equally effective (F. 237 supra), the five other clinical studies in evidence which compared TEAlS and placebo with negative results do not prove that there is no difference between the two. These negative studies simply show that the analgesic effcacy of TEAlS remains to be established. The five studies are the Roth study (CX 344Z-195); Ehrlich study (CX 344Z-157); charles study (eX 344Z-168); Brown study (eX 344Z-182); and Algozzine study (eX 255). All of the five tested a 10% TEAlS creme as an adjunctive drug to be used in conjunction with other therapy (Adriani, Tr. 1454). 373. The first four studies (CX 344 series) had been submitted by Warren-Teed Pharmaceuticals Corporation, the marketer of Myoflex (a 10% TEAlS rub), to the FDA in connection with the FDA's monograph proceeding involving OTe external analgesic drug products (eX 344A). In its Tentative Final Monograph on OTe External Analgesic Products, published February 8, 1983, the FDA referred to tb.e four studies and stated that none ofthem reported any significant differences between TEAlS and placebo for any of the measurements recorded (eX 443D).

374. While some ofthese negative studies have been criticized as not being well-controlled in several respects, these studies, together with the other studies discussed earlier (i. the Golden study, the Golden and Altschuler study, the French studies and the Batterman and Sanders study) show the many opportunities that TEAlS has hadwith different methodologies, by difierent investigators, and under Initial Decision 104 F. various conditions-to show that it is significantly better than placebo (Roth, Tr. 1622-23).

375. The five negative studies discussed here which tested TEAlS as an adjunctive drug also show that Aspercreme s effcacy when used to relieve the so-called breakthrough pain (See RPFs 138 and 140) remains to be demonstrated (Adriani, Tr. 1454-55). In any event, the evidence regarding the five so-called negative clinicals is summarized below.

1. The Roth Study (CX 344Z-195) 376. The first study, entitled Myoflex Arthritis Study (CX 344Z- 195), by Dr. Sanford Roth, was a double-blind, two-way crossover study which compared the eflects of Myoflex cream and a placebo cream in arthritis patients. One hundred and two patients were enrolled in the study, which took from July 1976 to November 1978 to complete (Roth, Tr. 1521). Effcacy measures in the study included the average daily duration of morning stiffness; average joint size in both hands; grip strength in both hands; an particular index with categories for (115) pain, swellng, and limitation of movement; a nine-point scale showing the patient's overall assessment of pain; and a composite particular index. Observations on each of these variables were taken at baseline and at the end of the first and fourth weeks of each treatment sequence. In addition, a treatment preference rating was obtained from the subjects at the end of the second treatment sequence. The study results shows that in comparing the two creams on ten objective effcacy parameters and on the subjective parameter of patient treatment preference, there were no statistically significant differences between Myoflex cream and placebo cream (Roth, Tr. 1518 , 1524-25; ex 344Z-200-04).

377. The subjects in the Roth study were patients with diagnosed chronic rheumatoid arthritis. The study design provided that the subjects could continue to use, as concomitant medications on an as needed basis, those oral anti-inflammatory and lor analgesic agents that they had been using prior to their participation in the study (CX 344Z-197-98). The reason for this was that rheumatoid arthritis is a systemic disease and since there was no evidence that the topically applied agent had any systemic effect, the existing forms of systemic therapy were continued to provide a baseline. The topically applied creams were being added on to the patients' regimens as a layer of further treatment to see if the test ingredient made an extra difference in controlling the pain symptom (Roth, Tr. 1752). By crossover of the test and control substances in a double-blind manner, each of the patients acted as his own control (Roth, Tr. 1552). There were no major changes in the subjects' regimens during the study period lDUlnrOV1., JVU' l.tvt"J, VV. , 11.,v.

648 Initial Dccision (Roth, Tr. 1523). An additional reason for the subjects' continuing their systemic medications during the study period was that the TEAlS product tested in the Roth study, Myoflex, was intended for use as an adjunct to other treatment and was in fact marketed as an adjunctive product. Where the tested product is so marketed, it would be inappropriate to take study subjects off their concomitant medications (Roth, Tr. 1524).

378. In the Roth study, many subjects suffering from moderate to severe osteoarthritis did not get any relief from aspirin despite the fact that therapeutic doses of aspirin have been proven effective in treating moderate osteoarthritis (Roth, Tr. 1714-15). Dr. Adriani also noted that the use of oral anti-inflammatory or analgesic drugs and the use of physical therapy during the test period made the results unreliable as evidence of Myoflex s analgesic eflcacy. 379. The Roth study finding that TEAlS is indistinguishable from placebo was one of those cited in the Tentative Final Monograph on OTe External Analgesic Drug Products as a basis for the FDA's conclusion that the evidence does not support Category I status for TEAlS as an OTe external analgesic (116) (eX 443D). In a letter to the manufacturer of Myoflex (eX 343e), the FDA's Bureau of Drugs likewise referred to the Roth study in reaching a similar conclusion. 2. The Ehrlich Study (eX 344Z-157) 380. The second clinical study, entitled Myoflex Creme in Putients With Chronic Musculoskeletal Complaints (eX 344Z-157), by Dr. George Ehrlich (one of Thompson s expert witnesses in this case), was a double-blind, placebo-controlled, crossover evaluation of a TEAlS cream and a placebo cream. The fifteen study subjects applied the TEAlS cream or placebo cream to the affected area three times daily for two two-week test periods. The objective ofthe study was to evaluate the analgesic effectiveness of Myoflex cream in the treatment of prolonged or chronic disability of musculoskeletal origin. Patients with arthritic symptoms, with sporadic musculoskeletal complaints, or with histories of spontaneous remission were excluded from participation in the study (eX 344Z-158).

381. The Ehrlich study was originally scheduled to have thirty subjects, but the investigator was unable to secure suffcient patients with the specified qualifications within a reasonable period of time and a decision was made to terminate the study at fifteen patients. This number is obviously inadequate (F. 232 supra). Of the fifteen patients, two did not complete both study periods and thus their results were excluded from the final evaluations. Physical therapy programs and oral anti-inflammatory drugs used by the subjects prior to entering the study were normally maintained. Ten ofthe subjects Initial Decision 104 F. were using a concomitant medication during the study period. This was in keeping with the usage and marketing ofMyoflex creme as an adjunct analgesic product. No oral or other typical salicylates were permitted during the study (eX 344-158-59). 382. The study results showed that Myoflex provided some pain reliefto six of the thirteen patients completing the study, while four of the thirteen patients reported some relief from use of the placebo cream. The degree of pain relief provided by Myoflex did not appear to be significantly greater than the degree of relief provided by the placebo. For the parameters of onset of pain relief and duration of relief, the differences between Myoflex and placebo were small and probably not significant. In terms of increased strength in the affected area, the improvement with Myoflex was not significantly different from that of placebo. In alj, there were no statistically significant differences between active drug and placebo for any ofthe measurements recorded. (117) 3. The charles Study (eX 344Z-168) 383. The third study by Dr. Alix A. charles, entitled Myoflex Creme in the Treatment of Chronic Musculoskeletal Complaints, had the same basic design as the Ehrlich study. Like the Ehrlich study, it was criticized for inadequate sample size, use of concomitant physical therapy and the maintenance of drug therapy in many of the subjects (Adriani, Tr. 1411-14; Roth, Tr. 1621).

384. Out of a total of thirty subjects planned for inclusion in the charles study, twenty-six patients entered the study, but final data is available for only twenty who completed the full study regimen. Of these, twelve continued their prior use of physical therapy andlor concomitant medication during the study period, again reflecting the intended use of Myoflex cream as an adjunctive product. The study showed that Myoflex provided some pain relief in fifteen ofthe twenty patients who completed the study. However, an equal number of patients reported relief from the placebo agent. Relief, as measured by improvement from the initial pain, was statistically significant for both treatments, but there were no significant differences between Myoflex and placebo for any of the parameters studied. The study write-up noted that the high placebo response rate observed in the study was not unusual for analgesic type studies. Both treatments required the topical cream to be rubbed into the affected area, it noted, and this massage action itself may be beneficial to patients wit,r, ml1 r1Jl()"k-plpt,Rl r.nmnlRlnt, .. .

It1U1Vl1 lVl 1.H ,,v.

648 Initial Decision 4. The Brown Study (eX 344Z-182) 385. The fourth clinical trial is entitled MyofiexlChronic Musculoskeletal Complaints(CX 344Z-182). It was conducted by Dr. Burnell Brown for Adria Laboratories, manufacturer of Myoflex. The objective of the Brown study was to evaluate the effcacy of Myoflex in the adjunctive relief of pain associated with chronic musculoskeletal complaints. The study included fifty-two patients (of whom fortytwo completed the full study regimen) in a double-blind, matched placebo controlled, crossover evaluat;on. The subjects were assigned to the test groups randomly, with half the subjects starting with Myoflex and half with placebo the first week, before being switched to the crossover medication for an additional one week. After the two one-week periods with TEAlS and placebo (without washout between periods), all patients were treated for a final one-week period with methyl salicylate. Statistical analysis of the data for the forty-two evaluable patients in the Brown study revealed (118) no statistically significant differences between the Myoflex, placebo, or methyl salicylate treatments based on any of the four study parameters analyzed: reduction of pain; increase in mobility; reduction in muscle tenderness; and overall musculoskeletal complaint improvement (CX 344Z-187). The Brown results are questionable because ofthe use of oral anti-inflammatory agents and failure to distinguish different clinical entities and diseases (Adriani, Tr. 1414-15; CX 344Z-182). Dr. Roth testjfied that the Brown study s methodology was flawed and that it did not measure up to the criteria for a valid study (Roth, Tr. 1719).

5. The Algozzine Study (eX 255) 386. The most recent ofthe five cl;nical studies in which TEAlS was reported to be no more effective than placebo is the study entitled Trolamine Salicylate Cream in Osteoarthritis of the Knee (eX 255), by Dr. Gary Algozz;ne et al. reported in the March 5, 1982 issue of the Journal of the American Medical Association. The study involved twenty-five patients, drawn from a Veteran s Administration hospital in Florida, who had symptomatic osteoarthritis of the knee. Study subjects were asked to designate their most painful knee and to apply either 10% TEAlS or placebo cream four times daily for two one-week periods in a randomized double-blind crossover study. Drug effcacy was measured along both subjective and objective parameters. 387. Patients were prohibited from using any concomitant oral or topical salicylates or any other analgesic drug during the study period and for the two days preceding the test period. However, because TEAlS was considered adjunctive treatment, patients currently re- Initial Decision 104 F. ceiving other forms of drug treatment for osteoarthritis (e. nonsteroidal anti-inflammatory agents) were eligible for inclusion in the study provided that their condition had been stabilized on a stated dose of the drug for the preceding one-month period. There were no changes in drug or dosage permitted during the study period. No patient received an intra-particular injection of a corticosteroid within the preceding six weeks, and no other form of treatment, such as external heat, exercise, or massage was used during the study period. Thirteen patients received concomitant non-steroidal anti-inflammatory medication during the study period.

388. The study results showed no statistically significant difference either in subjective or objective measures of relief between the treat- " testment and control groups. Eight patients preferred the "active cream, while six preferred placebo, and eleven had no preference (eX 255B). The investigator (119) concluded that the clinical data show the total effect of 10% TEAlS cream to be no better than that of placebo (CX 255C). In any event, this study suffers from a few defects. More than halfofthe test subjects received concomitant non-steroidal anti-inflammatory agents (Adriani, Tr. 1419), and they were virtually ajj bedridden males in a VA hospital, a group not weJJ suited for the trial of a topical drug intended for minor pain of arthritis. 389. The presence of conflicting evidence with respect to a drug effcacy leaves the scientific community in doubt about the drug effcacy, as respondent's own expert acknowledged (O'Brien, Tr. 3738). Dr. Steinberg, an offcer and employee of Thompson, testified TEAlS studiesthat in and around 1980 , he became aware ofthe four conducted by Drs. Roth, Ehrlich, charles and Brown (Steinberg, Tr. 5255). He became aware ofthe Algozzine study when it was published in the Journal of the American Medical Association in 1982 (ld. Despite this information, the company continued to make its unqualified effcacy claims for Aspercreme.

M. The Argument That The Evidence Of Perceived Benefits Constitutes Adequate Substantiation For Aspercreme Efficacy Claims 390. Respondent vigorously contends that Aspercreme has been shown to be a safe topical drug capable of providing perceptible pain relief to a significant segment of the consuming public and that this evidence alone is suffcient to satisfy the eommission s reasonable basis requirement. In this connection, respondent points to the fact that a number of FDA OTe drug panels, including the OTe External Analgesic Products Panel, have determined a number of ingredients to be effective on the basis of uncontrolled studies, long-term clinical use and marketing experience. Respondent urges that Aspercreme THOMPSON MEDICAL CO" INC. 769 648 Initial Decision should be found effective on the same basis (See RPF 123, 206-28; RB 49-59). These arguments, however, cannot be reconciled with the consistent refusal of both the FDA and its External Analgesic Panel to find TEAlS to be eflective for labelinglmarketing purposes (F. 393- infra).

391. The FDA OTe drug panels which based a finding of drug effcacy on evidence short oftwo well-controlled clinical trials include the OTe Internal Analgesics Panel:

(a) The Panel found, without well-controlled studies, the ingredient choline salicylate to be an effective analgesic. The Panel based its findings (120) solely on a survey of physicians who had given the ingredient to their patients. In some cases, the physician compared the eflects of choline salicylate in some patients with the known effects of aspirin in his patient population at large (eX 268, p. 35 418). (b) The Panel found the ingredient magnesium salicylate to be an effective analgesic. The Panel relied upon a clinical study with only twenty-two patients which compared magnesium salicylate with aspirin and found no statistically significant differences in the levels of analgesia. Because magnesium salicylate produced less gastrointestinal irritation than aspirin, the authors of the study concluded, and the Panel agreed, that magnesium salicylate was not only an effective analgesic, but preferable to aspirin for conditions requiring long-term therapy (eX 268, p. 35,419).

(c) The Panel found calcium carbaspirin to be an effective antipyret- , not based on controlled clinical studies, but on the fact that the absorbed moiety is aspirin and its established adequate bioavailability demonstrated an effect similar to aspirin (eX 268 , p. 35,448). 392. In reaching its conclusion of effcacy of a wide variety of externally applied ingredients, the OTe External Analgesic Panel considered data from both controlled and uncontrolled subjective studies (eX 269, p. 69 778). The Panel also gave consideration to reports of long-term, widespread satisfactory clinical use and marketing experience in evaluation of ingredients. For example, the Panel based its determination that:

(a) the ingredient Stronger Ammonia Water was an effective external analgesic on the ingredient's wide use, its clinical acceptance, and on published reports in the literature (eX 269, p. 69 793); (b) the ingredient Juniper Tar was an effective external analgesic based on the ingredient' s wide use, clinical acceptance, (121) and on published reports in the literature (eX 269, p. 69 824); (c) the ingredient Turpentine Oil was an effective external analgesic based on the ingredient's wide use, clinical acceptance, and on published reports in the literature. No scientifically controlled Initial Decision 104 F. studies concerning the use of turpentine oil alone for the treatment of rheumatism, arthritis, or muscular aches and pains were submitted (eX 269, p. 69 840).

VIII. FDA DETERMINATIONS ABOUT TEAls ANALGESIC EFFICACY 393. The most authoritative record evidence that topical TEAlS' analgesic effcacy remains to be demonstrated is the consistent refusal of both the FDA and its External Analgesics Panel to find TEAlS to be effective for labeling purposes under the Food, Drug and Cosmetics Act. The first decision on TEAlS' lack of eflcacy was that of the FDA' s Advisory Review Panel in December of 1979 (eX 269). The Panel was unimpressed by the evidence of TEAlS' effcacy. The Panel does not give serious consideration to the claim that the drug penetrates the skin and passes directly into the affected deeper structures in suffcient concentrations to be cfIcctive because there is no data available to substantiate this claim (eX 269. p. 69 8561.

394. In response to the Panel's rejection of TEAlS as a topical analgesic, both Thompson and Warren-Teed Pharmaceuticals, Inc. the maker of another topical TEAlS product, submitted additional material in 1980 to the FDA's Bureau of Drugs in an effort to reverse the Panel's decision (exs 342-43). The material submitted by the companies included the Golden study (RX 49, discussed in F. 246-74 supra), the Golden and Altschuler study (RX 50, discussed in F. 275supra), the Batterman and Sanders study (eX 254, discussed in F. 307- , supra), the Patel and ehappelle materials (RX 35- , discussed in F. 299-306 supra), the canine portion of the Rabinowitz study (RX 70, discussed in F. 315- supra), the Roth study (eX 344Z- 195 discussed in F. 376-79, supra), the Ehrlich study (CX 344Z- 157 discussed in F. 380- supra), the charles study (CX 344Z-168 discussed in F. 383- supra), and (122) the Brown study (eX 344Z- 182 discussed in F. 385 supra). As a result, the FDA's Bureau of Drugs had before it most of the clinical studies in evidence on this proceeding. The Bureau of Drugs reiterated the Advisory Panel's decision with respect to TEAlS that there were not suffcient data to support the drug s effcacy as a topical analgesic (exs 342e, 343DJ. 395. The current position of the FDA with respect to TEAlS' analgesic effcacy was formally announced upon publication of the tenative final monograph for external analgesics (eX 443) in which the FDA adopted the Panel' s determination and affrmed the Bureau s position. The pertinent portion of the FDA's tentative final monograph (proposed rule) on external OTe analgesic drug products nnnl1n(,pr- hrll!'rU R 1 QR 1.. 0:"" tnllnuH;:. THOMPSON MEDICAL CO., INC. 771 648 Initial Decision Two comments submitted data on the effectiveness of trolamine salicylate (TEA/SJ as a topical analgesic. Based on these data, one of the comments suggested that the monograph include a class of external analgesics that "act upon painful structures below the skin by absorption of the active ingredient directly into subcutaneous structures" and that trolamine salicylate (TEAlS) be placed in this class. The comment also suggested the following indications for this class: "For the temporary relief of minor aches and pains of muscles and joints. Also as a topical adjunct for pain due to arthritis and rheumatism." Both comments suggested that trolamine salicylate (TEA/Sl be placed in Category I based on the data submitted. Because the submitted information fails to demonstrate that this ingredient would be effective for application at the site of pain or for any use as an external analgesic the agency does not agree with the comments that trolamjne salicylate (TEA/SJ should be placed in a new class of external analgesic drug product. Trolamine salicylate (TEA/SJ remains in Category III as an anesthetic, analgesic, and antipuritic in this tentative final monograph (CX 41: , p. 58551. (123) The FDA's proposed rule also specifically rejected as inadequate ajj of the material referred to in F. 394 supra (ld. 396. Concomitantly, Dr. Wiliam E. Gilbertson, the Director of the Division of OTe Drug Evaluation, advised Thompson by letter dated February 9 , 1983:

All of the data and information in your Citizen petition and subsequent correspondence, as identified above, have been included in the Administrative Record. When the review of the data and information is completed, you will be notified of the Agency findings pursuant to the Agency s feedback procedures (RX 296). 397. Included in the data submitted by Thompson to the FDA but not specifically considered by the FDA in its tentative final monograph (eX 443) are:

(i) eitizen Petition dated November 24, 1981 (RX 366), containing inter alia the published report of Dr. Rabinowitz (RX 70 discussed in F. 315- supra); the pilot investigation of TEAlS' ability to influence prostaglandins synthesis (RPF 146 discussed in F. 355pra); backup data from the Golden and Altschuler study (RX 50 discussed in F. 275- supra); written report and patient report forms on the French elinical Trials conducted by Drs. Alain Patel and Pierre Andre ehappelle in 1976 and 1977 (RXs 35-37 discussed in F. 299-306 supra); as well as considerable marketing and consumergenerated information.

(ii) B.F. Ascher company, Inc. survey of physicians (RX 346 discussed in F. 338- supra).

(iij) "Drug Delivery to Local Subcutaneous Structures Following Topical Administration" a review article by Drs. Guy and Maibach ..., .. _ . ( Initial Decision 104 F. accepted (124) for publication in The Journal of Pharmaceutical Sciences (RX 1020).

(iv) Double-blind clinical study conducted at the Baltimore center for elinical Studies entitled "Double-Blind Clinical Evaluation of Aspercreme Plus Two Placebo Tablets Versus Aspirin Plus a Placebo ereme For Relief of Pain Associated with Arthritis " purporting to show that Aspercreme is as effective as aspirin. SeeRB 17.

398. Ofthe pending FDA submissions by Thompson, referred to as Your Citizen Petition and subsequent correspondence" in Dr. Gilbertson s letter of February 9, 1983 (RX 296). the only biomedical studies not specifically commented on and rejected as insulIcient by the FDA to date are: the Rabinowitz study (RX 70 dicussed in F. 315- supra); the Guy and Maibach article (RX 1020 discussed in F. 324 supra and a clinical study conducted for Thompson by the Baltimore center for elinical Studies ("Baltimore study ). The last two evidently postdate Thompson s November 1981 FDA petition and thus could not have been possessed or relied on by Thompson as substantiation for the various advertising claims challenged in this proceeding.

399. In any event, the Rabinowitz study and the Guy-Maibach article are penetration studies and do not purport to show TEAlS' bioactivity. The Baltimore study compared Aspercreme plus placebo tablets and aspirin plus placebo creme, as did the Golden study (eX 200), and thus lacked placebo-control. The lack of placebo control is a basic methodological flaw in a pain study (SeeF. 228 supra). Furthermore, its failure to show significant difference between the two treatment groups ("Aspercreme was as effective as, if not better than aspirin ) does not demonstrate that Aspercreme and aspirin are equally effective (See F. 237 supra). 400. Therefore, it is highly unlikely that the FDA, as a result of its review and feedback procedures referred to in the Gilbcrtson letter (F. 396 supra), will reverse its position with respect to TEAlS' topical analgesic eflcacy. Against this background, for the FTe to hold, on the basis of essentially the same evidence considered by the FDA, that Thompson s eflcacy claims for Aspercreme are based on adequate medical/scientific substantiation for advertising purposes would not only be contrary to the prevailing view of the medical scientific community but also be tantamount to establishing a different and (125) lower standard of effcacy for OTC drug advertising than that applicable to OTe drug marketing.

401. In sum, the clinical trials on which Thompson relies are inade- "-- 4-- .J .l-._ ,,t. i' rfp IQ t-""-..,, THOMPSON MEDICAL CO. , INC. '10 648 initial Decision analgesic. The bioavailability studies are not a substitute for wellcontrolled clinical trials, but also on the whole they show that TEAlS is poorly absorbed and thus probably not bioactive. The physicians observation and consumer testimonials upon which Thompson relies are corroborative evidence at best. And the general scientific literature and the various theories of TEAlS' mechanism of action are not a substitute for clinical demonstration of TEAlS efficacy. 402. From all the foregoing, it is found that at the time respondent made the advertising claims alleged in Paragraphs 12(a) and 14 ofthe complaint, it did not possess and rely on a reasonable basis for its claims that Aspercreme is an effective topical pain reliever, including arthritic pain.

IX. THOMPSON DID NOT HAVE A REASONABLE BASIS FOR COMPARATIVE EFFICACY CLAIMS 403. The record shows, and Thompson admits, that in order to show the comparative efficacy oftwo drugs, clinical trials directly comparing the two drugs are required (Adriani, Tr. 1177; Byers, Tr. 4384, CX 45Q (Admission No. 339); see also, American Home Products Corp., 98 C. 136 804- 15 (1981), modified 396 F.2d 681 (3rd eir. 1982) 1101 e. 698 (1983)j. The only clinical trial in evidence comparing the results of using oral aspirin and topical TEAlS is the Golden study (RX 48/eX 200) (Altschuler, Tr. 3060; ex 451 (Admission No. 142). The Golden study is not a well-controlled clinical study for the purpose of showing topical TEAlS' analgesic eflcacy (See F. 246supra). Accordingly, there is no reasonable basis for any claim about either TEAlS' parity with or superiority to oral aspirin. Moreover since Thompson s substantiation materials as a whole fail to provide the required level of scientific proof of Aspercreme s basic analagesic effcacy, there is clearly no substantiation for claims regarding the comparative effcacy of TEAlS. Thus at the time of the representations alleged in Paragraphs 12(b), 12(c) and 14 of the complaint Thompson did not have a reasonable basis for the claims that Aspercreme is either as effective as or more efiective than oral aspirin for the relief of minor pain associated with arthritis and rheumatism, as alleged in Paragraphs 12(b), 12(c) and J 4 of the complaint. 404. Accordingly, the claim that valid studies have scientifically proven that Aspercreme is more effective than (126) orally-ingested aspirin, as alleged in Paragraph 10(c) of the complaint, was, and is false.

Initial Decision 104 F. X. THOMPSON DID NOT HAVE A REASONABLE BASIS FOR EFFICACY CLAIMS ABOUT RHEUMATIC PAIN 405. Thompson has advertised Aspercreme as being effective for the relief of pain associated with rheumatic as weU as arthritic conditions. With the exception of only one document (RX 63/eX 265), Thompson relies on the same materials to justify its claims for rheumatic conditions that it relied on for claims about arthritic conditions (eX 44A-B). For the reasons discussed in Section VII, these materials fail to demonstrate that topical TEAlS is an effective drug for pain relief. Since rheumatism is musculoskeletal pain by definition, the materials Thompson used for arthritic conditions do not constitute a reasonable basis for any claims about Aspercreme s effcacy for relief of rheumatic conditions and symptoms.

406. RX 63/eX 265 is a letter from an employee of Warren-Teed Pharmaceuticals which markets Myoflex (a 10% TEAlS rub), to Dr. Saul Heber, a psychiatrist who specializes in acupuncture and has been a consultant to Thompson for more than twenty years (Heber Tr. 2566, 2604). Without disclosing his affliation to Thompson, Dr. Heller wrote to Warren- Teed for information about its topical TEAlS product (Heber, Tr. 2629). Warren-Teed' s reply (RX 63) briefly describes some bioavailability and clinical studies conducted by Warren- Teed on Myoflex. Dr.' Heller never saw any of the studies mentioned in the letter, and the letter only suggests that Warren-Teed' s product may be used as "an adjunctive therapeutic agent" (Heller, Tr. 2625- 28; RX 63B). This letter adds nothing as a practical matter to the reasonable basis materials discussed in Section VII above. Accordingly, when Thompson made the representation alleged in Paragraphs 12(d) and 14 of the eommission s complaint there was no acceptable scientific support and no reasonable basis for the claim that Aspercreme is effective for the relief of rheumatic pain (Adriani, Tr. 1185; Roth, Tr. 1573-74).

Xl. THOMPSON DID NO'!' HAVE A REASONABLE BASIS FOR CLAIMS ABOUT DRUG ACTION BY DIRECT PENETRATION 407. For the same reasons underlying the foregoing findings that there is no reasonable basis for Aspercreme s effcacy claims, it is found that there is no reasonable basis for the claim, alleged in Paragraphs 12(e) and 14 of the complaint, that Aspercreme provides pain relief by direct penetration to the site of arthritic pain. Moreover, the mechanism of pain or its (127) relief has not been definitively understood, and the bioavailability or penetration studies in evidence fau short of showing TEAlS' bioactivity or analgesic effcacy. 408. From all of t.hp fort-'jJoinD' it -ic; frmnrl t'h !1U1Vll'::Ul.. lVlJ Ull1'1J LV., ll'llJ. , 'u 648 Initial Decision (a) the advertising claims alleged in complaint paragraphs lO(a), 10(b) and lO(c) were false, misleading or deceptive; (b) the advertising claims alleged in complaint paragraphs 12(a), 12(b), 12(c) and 12(d) lacked a reasonable basis and they were false as alleged in complaint paragraph 14; and (c) the use of the brand name "Aspercreme" in advertising, labels and promotional materials is false, misleading or deceptive. XII. RELIEF 409. With respect to certain Aspercreme advertising claims which were false or for which Thompson did not have a reasonable basis, the customary remedy is an order to cease and desist. It is necessary and appropriate in this case to require Thompson to refrain from making false and unsubstantiated claims in the future. Such false or unsubstantiated claims include:

(a) False claims:

(1) Aspercreme contains aspirin (eomp. n 10(a)). (2) Aspercreme is a newly discovered drug product (eomp. 111O(b)). (3) Valid studies have scientifically proven that Aspercreme is more effective than orally-ingested aspirin for the relief of minor pain (128) associated with arthritis and other rheumatic conditions (eomp. n lO(c)).

(b) Unsubstantiated claims:

(1) Aspercreme is an eflective drug for the relief of minor pain associated with arthritis or rheumatic conditions (Comp. n 12(a) and (d)).

(2) Aspercreme is as effective as, or more efiective than, orallyingested aspirin for the relief of minor pain associated with arthritis (eomp. n 12(b) and (c)).

(3) Aspercreme acts by directly penetrating through the skin to the site of arthritic pain (Comp. n 12(e)).

410. With respect to an advertising claim for Aspercreme s effcacy, either in simple or comparative terms, the required reasonable basis comprises two or more adequate and well-controlled clinical trials which demonstrate Aspercreme s simple or comparative effcacy as an analgesic drug for the relief of minor pain associated with arthritis or other rheumatic conditions. This is what the prevailing view ofthe medicallscientific community requires, as does the FDA for OTe topical analgesic products for labeling purposes. Acceptance by the Federal Trade eommission of a lower level of scientific evidence of drug effcacy would not only be contrary to the prevailing and established view of the medical scientific community but also be tantamount to Initial Decision 104 F. establishing a different and lower standard of effcacy for OTe drug advertising under Section 5 of the FTe Act than that applicable to OTe drug marketing under the Food, Drug and Cosmetic Act. 411. The need for adequate substantiation is greater where, as here a relatively obscure topical drug is being touted as proven effective pain reliever to the long-suffering arthritics, a group known to be singularly disposed to grasp at new promises of relief. 412. Thompson has been aware for some time that its advertising and the product's brand name may imply that (129) Aspercreme contains aspirin. For example, the networks and the NAB eode Authority challenged Aspercreme commercials with respect to the aspirin content suggestion, as well as challenging Thompson s substantiation for its effcacy claims (See, e. ex 88D, 92). Thompson also was aware ofthe results of the Mapes & Ross copy test (which showed that an aspirin content message was being communicated) at least as early as June 14, 1979, at which time the test was discussed in a meeting between Ogilvy & Mather and Thompson (See, ex 99A: F. 95 supra). In a document setting out the strategy for Aspercreme advertising, Thompson s advertising agency stressed the importance of the aspirin communication (eX 54Z ("the 'aspirin' component of Aspercreme is uniquely relevant for arthritis ), Z-002 (discussion of "aspirin in a rub" as the ad strategy), Z-005 (creative strategy: "Aspercreme contains the pain relieving ingredient of aspirin ), Z-007 (" it is the local aspirin relief that is important"), Z-012 (with respect to an Aspercreme print ad: "the client has grown to believe it doesn t communicate arthritis or aspirin very well, and we are in the process of developing a new one )). Since representatives of the ad agency communicated regularly with Thompson in the creation of ad strategy and had to get direct approval from Thompson for the strategy (Jasper, Tr. 4815-16), it is reasonable to infer that the ad agency s expressions in ex 54 were known to Thompson. ex 66 is a conference report summarizing a meeting between Thompson and Ogilvy & Mather, at which Thompson gave approval for certain Aspercreme ads to be run (Jasper, Tr. 4794-97). Among those ads were ex 3 and ex 4, television commercials including the network-mandated ncontains no aspirin disclosures. The report states, "the client fThompson) agreed and wil pursue approval of the aspirin equivalency claim. In the meantime, the agency wil pursue using the ' strong relief of aspirin ' claim offset the contains no aspirin super (eX 66B, emphasis added). ex 66 was sent to Thompson s president, as well as to other Thompson representatives (CX 66A).

413. The record also clearly shows that a significant segment of consumers are likely to get the impression that Aspercreme contains aspirin. either from the brand name !!Aspercreme" alone or from THOMPSON MEDICAL CO., ING 777 648 Initial Decision other language in the ads or both, and that ingredient disclosure statements (such as "Aspercreme contains salicyn, a strong non-aspirin pain reliever ) or aspirin disclaimers in the ads (such as does not contain aspirin" super) are ineffective in preventing the consumer perception that Aspercreme is an aspirin product (cohen, Tr. 287- 549- , 558-60). Therefore, stronger measures are required in Aspercreme advertising in the future in order to insure that the viewer of such advertisements are not misled thereby (See F. 84, 101 , 116). (130) 414. It is the determination of the administrative law judge that based on the record as a whole an eflective aspirin disclaimer should be explicit and unequivocal and should include both audio and video messages which are prominent enough and conspicuous enough to assure that the intended objective wil be achieved. In television commercials, a prominent and conspicuous video aspirin disclaimer (such as "Aspirin Free" super) should be displayed throughout the commercial as well as a vocal aspirin disclaimer statement (such as "Aspercreme does not contain aspirin ) made at the end of each commercial. In print or radio advertising, the printed or vocal aspirin disclaimers should be explicit and unequivocal and be prominent and conspicuous in relation to each advertisement as a whole. It is the opinion of the adminstrative Jaw judge that without an eflective disclaimer measure as described above, a continued use of the brand name Aspercreme in advertising will be clearly misleading and deceptive. 415. The brand name Aspercreme was registered in the United States Patent Offce under Registration No. 933, 107 issued in 1972 (RX 921). It also has been registered in many foreign countries (RX 601A). Furthermore, the perception that Aspercreme contains aspirin is not a result of the natural and literal meaning ofthe trade name itself but a result of the confusion created by its proximity to "aspirin." The trade name excision complaint counsel advocate is notjustified under the circumstances of this case. On the basis of this record I am unable to conclude that the brand name excision is the only adequate remedy or that a less restrictive remedy (such as effective disclaimers) cannot be devised that will prevent the impression on the part of the viewer that Aspercreme ads represent that the product contains aspirin. See Beneficial Corp. u. FTC, 542 F.2d 611, 619 (3rd eir. 1976).

416. It is found that the cease and desist order provisions should be limited to OTe analgesic products. The record evidence does notjustify an "alj drug products" provision sought by complaint counsel. (131) Initial Decision 104 F. DISCUSSION The record shows that the capacity of the challenged Aspercreme advertisements to mislead a significant segment of reasonable consumers is palpably real, that the deception is material and substantial in health, economic and societal terms, that the information regarding the availability of a topical OTC drug alternative to oral medication can easily and effectively be communicated in advertisements without implying false, deceptive or misleading claims, and that the benefits of requiring adequate substantiation of Aspercreme effcacy claims clearly outweigh potential costs of such requirement. The record also shows that the brand name "Aspercreme" is misleading and that the kind of fleeting aspirin disclaimers (such as "Does not contain aspirin " super displayed for a few seconds) or equivocal ingredient statements (such as contains salicyn, a strong non-aspirin pain reliever ) in Aspercreme ads are not suffcient and more effective, straight-forward aspirin disclaimers are needed. The basic findings with respect to the various disputed issues of law and fact and the reasons for the conclusions reached are set forth in the preceding sections. However, a brief discussion of some key issues may be in order.

1. The Meaning Of Aspercreme Adoertisements And The Materiulity Of Aspirin Content Thompson argues that its Aspercreme ads sought merely to inform arthritis suflerers that there is an alternative topical OTC remedy which provides as much pain relief as do aspirin tablets without stomach upsets often caused by aspirin tablets. A careful examination ofthe individual advertisements in evidence, however, has convinced me that, whatever Thompson s intent, these ads also made, sometimes directly but mostly by implication, the various representations alleged in the complaint, except for paragraph 12(D. Empirical data including the four copy tests conducted in connection with this litigation (exs 26 and 27; RXs 500 and 520), generally confirm what common sense and experience would tell us, namely that the Aspercreme ads, including the use ofthe brand name "Aspercreme " can reasonably be expected to mislead a significant number of consumers in the manner alleged in the complaint.

The record also amply demonstrates that the presence or absence of aspirin in an OTe analgesic product clearly is a material fact within the meaning of Section 12 of the FTe Act (See F. 162supra). (132) , .

lilV1Vlr.:VJ'I l Ulv.tl IvV. , U'lV.

648 Initial Decision 2. The Requirement For Well-Controlled Clinical Studies As A Reasonable Basis For Aspercreme Efficacy Claims The thrust of Thompson s argument regarding the reasonable basis issue is essentially that, on the basis of the totality of evidence of TEAlS including clinical trials, drug penetration studies, physician opinions and clinjcal experience, marketing experience, user testimonials and consumer reactions regarding Aspercreme over the years, it was reasonable for Thompson to make those advertising claims challenged in this proceeding (RB 59-72). Thompson argues that Aspercreme is a mild and harmless topical product which demonstrably provides tangible benefits to a significant segment of the target population and that, therefore, it should be permitted to make effcacy claims in advertising without the kind and level of medical scientific substantiation required for more potent, potentially harmful drugs (RB 61-62). However, the eommission has determined that with respect to the issue of efficacy of an OTC analgesic drug, two or more well-controlled clinical trials are required to prove simple or comparative effcacy. American Home Products Corp. 98 F. e. at 376. Therefore, an argument that a level of substantiation less than two well-controlled clinicals constitutes a reasonable basis for the simple and comparative efficacy claims made for Aspercreme, a topical OTC analgesic drug, is not acceptable.

Thompson also argues that the FDA does not require two wellcontrolled studies in the case of the so-called "old" OTC drugs (133) (RPF 64-66). Thompson further argues that, in any event, the FDA relies on uncontrolled clinical trials, clinical observations of physicians and marketing experience related to the product in determining the issue of effcacy of OTe drugs (See RPF 123-25)4 Thompson also urges that " under the principles of res judicata and collateral estoppel" the FTC should apply to drug effcacy issues in FTe proceedings the same standard the FDA employs in determining drug effcacy. Indeed Thompson points out that "the eommission has itself used the finding of the FDA to reach its own finding that advertising was J In this conn ction, the record show that TEAlS is not an effective sysl mic Or internal analgesic(SeeCX 268) hut is an effective topical sunscreen agent (F, 362 supra;CX 270, PI'. : 2Gl-53). An important property of5- 12% TEAlS a a sunscreen product is its ability to n main On the skin long enough to form an ef1ectiv urface barrier against ultraviolet ray of the sun, Yet, Thompson hypotl",sjs regarding Aspcrcrerne s 00% TEAlS) anajg sic r.ion is that it penetrat t.he skjn and delivers a therapeutic level ofsalicy!ate to th site of pain in the underlying deep tissues of t.he musculoskeletal system in humans. The records also shows that WI"the of an unproven UTe drug for self-medication t.o treat rheumatic pain poses a real danger to th consumer and may result in significant losses of individual and societal resources (F. 207- supra). In these circumstance, common sense dictates that Aspercreme s analgesic elIcacy be subjected to as rigorous a test as ifnot more rigorous t.han, that required of ate interned analgesic product..A/so see,F, 242 "pm 'These argumf'nts ignore Thompson s own statement that. "the FDA OTe monographJ procedures now provides a period until April 9, 1984 for the development and review of evidence that will permit final classification ufthe cffectiveness of TEAlS" and that "during this period the marketing of TEAlS products is, ofeourse, permitt.ed" (RB !8). Thr. "evidence that will permit final clas. ification of the efJectivp.ness of TEAlS" in this context, of course includes two or more adequate OInd well-contro1Jed clinical trials Initial Decision 104 F.T.C. deceptive " citing Simeon Mgt. Corp. u. FTC, 579 F. 2d 1137, 1145 n. 10 (9th eir. 1978) (SeeRB 99-100). According to this argument, the FTe should accept, as the FDA allegedly would, evidence less than two or more well-controlled clinicals as proof of analgesic effcacy of Aspercreme. In my view, these arguments are unpersuasive. As the CaID mission has explicitly recognized in American Home Products the FDA requires under the Food, Drug and Cosmetic Act two or more well-controlled clinical studies as proof of effcacy with respect to all drugs, including OTe analgesic drug products. 98 F. e. at 378-81. The most authoritative record evidence that topical TEAlS' analgesic effcacy remains to be demonstrated is the consistent refusal of both the FDA and its OTe External Analgesic Panel to find topical TEAlS to be effective under the Food, Drug and Cosmetic Act (See F. 393- supra). Having reviewed substantially aji of the material and information presented in this proceeding, the FDA has announced, in the February 8, 1983 Tentative Final Monograph (proposed rule) on OTe External Analgesic Drug Products, its determination that there was insuffcient evidence to conclude that TEAlS is an efiective topical analgesic agent. Against this background, the need for regulatory harmony and uniform standards among federal agencies with respect to the issue of drug effcacy dictates that the FTe require in this proceeding the same level of medical scientific (134) evidence the FDA requires under the Food, Drug and cosmetic Act. Thompson has urged as much (RB 99-100). In light ofthe foregoing discussion, Thompson s argument that the requirement for well-controlled clinicals in this case is somehow excessive" and that it is likely to result in limiting the amount of useful information " made available to consumers is inapposite (See RB 59-61). The record clearly shows the substantial harm that misinf L:nder the FDA's monograph procedures for OTC ext.ern,!! analgesic dmg pJ'oducb, the rnarkeling of TEAlS products, incJlJding Aspcrcreme, will be permitted fur an imerim period until April 9 , 1984 , pending development and revi"w of"eviden"e that wjJj permit tinal classification oftl", !,fr' "livene55 of TEAlS " presumably includinf! two Uf more wejJ.contruHed clinical trials (F. 9 'npm In this connection, Thompson obliquc1y sllggcsL t.hat the FDA might change iL position regardingTF:A/S when the FDA completes n review ofThompson\; " pending " submissions (111: 16 18), lIowev(,r, on the basis of the record theoutcorne Thompson S\1ggHsts is highly unlikely (SeeF, 396-1\00 supra). In any event. the Order makes provision for t,he cootingency thilllhe FDA migbt, on the basis of new cvi(hmrc, determine TJ:A/S to be an eITeNive external aoalgesic agent. Should that contingency occur, till' Order would ,,!luw respondent to rely on IJ", Sij"''' ",vidence the FDA relied on in the monograph pro"""ding Th;tt pruvision will also cover other OTC anillgesic drug pruduct. that may be markeled by respondent in t1,. future- In this connection, it should he nOLed that the FDA, in the OTC monograph prol:eerJing, dd"rmines the conditiulls ullder which OTC anillges;c drug products are "gencnlily recognized OIS safe and eITedive OInd not misbr,-ltded" and does 001. dt al with the issue of """,parativceffcacy as Ruch (SeeCX 11\3, pp- 51:5 53) In sum, should the FTC Order herein become eft,dive be furl' Apl-il!:, 1!:84 , it would have t.he practical etrect of barring Thompson from advertising Aspen:renw as a topical analgesic while Thompson is being permitted by the FDA to continup. marketing Aspercreme during the interim period, It may well be that the time reqlJi(f d foi thefinalddr,rrninOltionofthisproceedingmaymoottheissue- Juwt' ver int.heinterestofcomityandr"'ilJlctory harmony, the Order should provide for a grace p"riod similar to the FDA' s (namp.ly a period lIntil April 9, 1984) during which simple, non-comparative effcacy cbjms for Asperueme may be f\!lowed in advertising It is the law judge s vip.w, however, that,dl ad daims ofcomparatiucor superior dIicf\cy for Aspercreme should cease forthwith aIJd not. I", allowed unt,il such claims ,,,m hI' "rJ"o""t."lv ."lrh"t."nt.i;,tp 648 Initial Decision formation regarding OTe (135) analgesic "rugs may cause to the long suffering arthritics as well as to the society as a whole (F. 207supra). The record also shows that the message Thompson claims to have sought to convey to the consumer, namely that Aspercreme is an effective topical alternative to oral aspirin, can easily and clearly be conveyed without implying unproven claims of comparative effcacy. In these circumstances, a ban against unproven claims of simple and comparative effcacy for Aspercreme is imperative. Thus, the determination that a reasonable basis for Aspercreme effcacy claims should include well-controlled clinical studies reflects a careful weighing ofthe possible benefits to consumers if Thompson s effcacy claims are true and the possible costs to consumers if the claims are false. SeeThe FTe Statement of Policy on the Scope ofthe consumer Unfairness Jurisdiction, 43 Trade Reg. Rep. , 570 (1982). In any event, as the eommission has recognized, the Supreme court has indicated that the First Amendment does not shield deceptive advertising against appropriate prior restraints needed to prevent its recurrence. See American Home Products Corp. 98 F. e. at 403and the cases cited therein.

In this connection, a proposed restriction on commercial speech (including the brand name) designed to prevent recurrence of deception should be viewed in the information cost perspective so as not to inhibit dissemination of economically effcient information.6 Howev- , it should also be borne in mind that information in this context means accurate, truthful information.' False, misleading or deceptive advertising or brand name is not economically effcient. On the contrary, it increases the costs of information about the qualities of products and is economically inefficient. This is such a case. 3. The Advertising Claims That Aspercreme Is A Newly Developed Drug And That It Has No Side Effects It is true that certain early Aspercreme advertisements contain an express or implied claim that Aspercreme is a new drug, as alleged in paragraph 10(b) of the complaint. However, (136) the history of Aspercreme advertising supports the view that Thompson was attempting to bring Aspercreme to the consumer s attention during the period when Aspercreme was being nationally advertised for the first time. In my view, Section 5 gives an advertiser some leeway in making a "new drug" claim during the introductory phase even though the same or similar products may have been available on the market. In 6 SeeCmJse, JUI. The Problem of 5"':1,,/ COST 3 J of' Law & EcotJ., 1 , 15 (1960); Alchian, A. & Allen, W, EXCHANGE A:-D PHODCCTION: COMPETITION, COORDINATION AXD CONTROL, 110- , 124 294-95 (2d ed1977) "I Alchian & Allen supra, at 121 , , Initial Decision 104 F. any event, Thompson should not make "new drug" claims for Aspercreme in the future.

It is also true that some of Thompson s advertisements contain an express "no side effects" claim, as alleged in paragraph 12(1) (exs 7 , 10, 11). The record also shows that Aspercreme in fact can irritate the skin in some users (Golden, Tr. 2700-1; RX 35Z002). However when each advertisement (CXs 7, 8, 10, 11) is viewed separately, I am unable to conclude that any of these advertisements can be reasonably understood by viewers as saying that Aspercreme has literally no side effects. It is common knowledge that any topical drug may irritate the skin in some persons. In may view, these ads were saying either that Aspercreme does not cause stomach discomforts as aspirin tablets often do or that Aspercreme has no serious side effects to worry about. To impute any more to the "no side effects" claim in these ads would be unreasonable. Therefore, complaint allegations in paragraph 12(1) wil be dismissed.

4. The Trade Name Excision Issue Complaint counsel vigorously argue that the record evidence demonstrating the misleading eflect ofthe brand name "Aspercreme clearly justifies its excision (eB 109-16). They urge essentially that the brand name "Aspercreme " is so close to aspirin cream" that any qualification of the name will necessarily involve a contradiction in terms, adding to confusion rather than removing it. complaint counsel further assert that, when the names of drugs or other medical products are at issue, courts readily have found confusion from less proof of confusing similarity because of the risk of physical harm to the consumer (eB 111). However, trade name excision is an extreme and harsh remedy and should not be employed except in cases where there is no meaningful alternative.

Although the evidence does not show that the brand name "Aspercreme" has acquired a secondary meaning associating it with some standard or price Aspercreme" has been in use as a registered brand name since 1972 and extensively advertised on a national scale as an OTe topical analgesic since 1979. As such Aspercreme" is a valuable business asset to respondent. C(, Friedman u. Rogers 440 U.s. 1, n. 11 at 12 (1978). Furthermore, in my view, there is a reasonable, common sense distinction to be made between " aspirin cream " or "Jay (137) aspirincreme" on the one hand and "Aspercreme" on the other. In the former, the natural and literal meaning ofthe names would a creme made of aspirin." These names do not involve any ambiguity. In the latter, the construction that the product contains aspirin is due to the name s ambiguity, or proximity to Haspirin " and the confusion is not attributable to the natural and literal meaning ofthe name , ......, . , .& . . nu.. UV"" HU-".I''''' U' u,,.

648 Initial Decision itself. In these circumstances, applicable precedents require that the eommission fully examine less restrictive measures including the feasibility of requiring rewriting of advertising copy in lieu of total excision. Beneficial Corp. v. FTC, 542 F.2d 611 , 619 (3rd eir. 1976); Continental Wax Corp. v. FTC, 330 F.2d 475, 479-80 (2d Cir. 1964). The evidence shows that the direct or indirect aspirin disclaimers used heretofore in Aspercreme ads are ineffective and that, therefore, stronger, more conspicuous aspirin disclaimers are required (F. 84 101 116 413-14, supra). Thus, simple, unequivocal aspirin disclaimer statements must be prominently and conspicuously made both in audio and video form in all Aspercreme ads in order to insure that the ambiguity and confusion resulting from the brand name "Aspercreme" are effectively removed. Therefore, the order wil include appropriate provisions which will require, in television advertisements Aspirin- free" video super throughout the entire commercial in addition to a vocal disclaimer ttAspercreme does not contain aspirin" at the end of each Aspercreme commercial. In radio ads, one clear aspirin disclaimer statement at the end of the commercial wil be suffcient. In print ads, a simple and unequivocal aspirin disclaimer statement, such as "ASPEReREME DOES NOT contain ASPI- RIN " should be prominently and conspicuously placed in relation to each advertising copy as a whole (F. 415 supra). CONCLUSIONS OF LAW 1. The Federal Trade eommission has jurisdiction over the advertising of Aspercreme under Section 5 of the Federal Trade Commission Act.

2. Respondent's use of false, misleading and deceptive advertising representations (including the brand name "Aspercreme ) as herein found has had and now has the capacity and tendency to mislead members of the purchasing public into the erroneous and mistaken beliefthat said statements and representations were and are true and into the purchase of substantial quantities of Aspercreme by reason of said erroneous and mistaken belief. In the absence of an appropriate order, such members of the purchasing public are likely to continue to purchase substantial quantities of Aspercreme in the mistaken (138) beliefthat respondent's past advertising representations regarding the aspirin content of Aspercreme, Aspercreme s novelty, and the effcacy and comparative effcacy of Aspercreme (including the existence of valid studies proving Aspercreme to be more effective than aspirin), were true or were supported by a reasonable basis. 3. The acts and practices of respondent as herein found were and are alj to the prejudice and injury of the public and of respondent' ....,.

Initial Decision 104 F. competitors and constituted and now constitute unfair methods of competition and unfair and deceptive acts and practices in commerce in violation of Sections 5 and 12 ofthe Federal Trade Commission Act. 4. The accompanying order is necessary and appropriate for the purpose of prohibiting the continuation of the proscribed acts. ORDER It is ordered That respondent, Thompson Medical Company, Inc. a corporation, its successors and assigns, and respondent's offcers representatives, agents and employees, directly or through any corporation, subsidiary, division or other device, in connection with the labeling, advertising, offering for sale, sale or distribution of any over-the-counter analgesic "drug" as that term is defined in the Federal Trade Commission Act, in or affecting commerce, as "commerce is defined in the Federal Trade eommission Act, do forthwith cease and desist from:

A. Employing the brand name "Aspercreme" for such products or otherwise representing directly or by implication that an active ingredient of such product is aspirin, unless such product contains aspirin in therapeutically significant quantities or unless the advertising and labeling for such product clearly and prominently discloses that the product does not contain aspirin.

(1) In television advertisements, an explicit and simple aspirin disclaimer statement (such as (139) "ASPIRIN-FREE") shall be superimposed and prominently displayed throughout the length of each advertisement as well as a vocal aspirin disclaimer statement (such as "Aspercreme does not contain aspirin ) at the end of each advertisement.

(2) In radio advertisements, an explicit aspirin disclaimer statement (such as Aspercreme does not contain aspirin ) shall be made at the end of each advertisement.

(3) In print advertisements, an explicit aspirin disclaimer statement (such as "ASPERCREME DOES NOT contain ASPIRIN" shall be displayed prominently and conspicuously in relation to each such advertisement as a whole.

(4) In labeling, an explicit aspirin disclaimer statement (such as DOES NOT contain ASPIRIN") shall be prominently and conspicuously printed in the front package panel (or in the front of the container if no package is used).

"Q "R,.-n..,.C'",.. t;..,. ..nr.tlh .. hu ;1'T' l;,... t;.... t"h..d- n",."h ..nn 648 Initial Decision new, or involves a new mechanical or scientific principle except during the product's introductory period, when such product or one involving such principle has been generally available for purchase in the United States for more than one year.

e. Misrepresenting the contents, validity, results, conclusions, or interpretations of any test or study.

It is further ordered That Thompson Medical company, Inc. , a corporation. its successors and assigns, and respondent' s offcers, representatives, agents and employees, directly or (140) through any corporation, subsidiary, division or other device, in connection with the labeling, advertising, offering for sale, sale or distribution of any OTe analgesic "drug," as that term is defined in the Federal Trade Commssion Act, in or affecting commerce, as commerce is defined in the Federal Trade Commission Act, do forthwith cease and desit from:

A. Representing, after April 9, 1984, that such product is effective for the reJiefofminor pain and other symptoms of any musculoskeletal disorder (such as arthritis, tendinitis, bursitis, or rheumatic disorders);

B. Representing that such product is as fast as or faster than, or is as effective as, or more effective than any other drug or device in the relief of minor pain and other symptoms of any musculoskeletal disorder (such as arthritis, tendinitis, bursitis, or rheumatic disorders); unless at the time of the dissemination of any such representation respondent possesses and relies upon a reasonable basis for such representation consisting of competent and reliable scientific or medical evidence. For analgesic drug products competent and reliable scientific or medical evidence shall include at least two adequate and wellcontrolled, double-blinded clinical studies which conform to acceptable designs and protocols and are conducted by different persons independently of each other. Such persons shall be qualified by training and experience to conduct such studies. Provided however with respect to any representation covered by this part other than claims of superior or comparative effectiveness or safety, if the Food and Drug Administration promulgates any final standard which establishes conditions under which such product is safe and efiective under the Food, Drug and cosmetic Act, then in lieu ofthe above, respondent may rely upon scientific evidence which fully conforms to such final standards as a reasonable basis for said representation. Opinion 104 F.

It is further ordered, That so much of the complaint as relates t 0 Paragraph 12(1 be, and the same hereby is, dismissed. OPINION OF THE eommission By DOUGLAS, Commissioner:

I. INTRODUCTION Thompson Medical company, Inc. ("Thompson ) is a pharmaceutical company that sells several difierent over-the-counter ("OTC" drugs. One such drug is Aspercreme, a topical cream rub or lotion whose active ingredient is 10% triethanolamine salicylate ("TEAlS" Thompson has marketed Aspercreme as a remedy for relief of minor arthritis pain since 1976, after purchasing it from the Sperti Drug company. At first, Thompson followed Sperti' s practice of marketing the product locally in Ohio. Starting in 1978, however, Thompson began a successful national advertising campaign for Aspercreme which saw sales of the product increase from $589 000 in 1978 to $5. million in 1981 (IDF 74.)1 (2) On February 5, 1981, the Commission issued a complaint against Thompson and its advertising agency, Ogilvy and Mather, Inc. The complaint charges that Thompson s marketing campaign for Aspercreme involved the following deceptive representations: Complaint Alleged Paragraph Number Representation 10(a), 16 Aspercreme contains aspirin. 10(b) Aspercreme is a recently discovered or developed drug product. , The following abb!'eviatioi", an ed in thi opinion: initial d ion page number !DF initial decision finding numbnr - tran criptoftestimony page number complaintcounscl' s exhibit number CI'F complaint. counsel' proposed nndings of fact and conclusions of law finding number CMF - complaint coll" ,,I' memorandum supporting propos d findings of fact and conclusions of Jaw page number CAP complaint counsel's appeal brief page number CAB - complaint counsel's an wering brief page number - respondent s exhihit number HPF respondent's proposed finding of fact and conclu ions of law finding number RMF respondent's memorandum supporting proposed findings of fact and conclusions of law page RAP numb",'pondent s appeal brief page number RRB - respondent' s reply brief page number ,. Complaint Paragraph 1O(a) charges Thompson generally with representing that Aspercreme ';contain aspirin Complaint Paragraph 16 specifically charges that use of the trademark "Aspercreme" constitutes such a representation.

648 Opinion 10(c) Valid studies have scientifically proven that Aspercreme is more effective than orally-ingested aspirin for the relief of arthritis, rheumatic conditions, and their symptoms.

12(a) Aspercreme is an effective drug for the relief of minor arthritis and its symptoms.

12(b) Aspercreme is as effective a drug as orally-ingested aspirin for the relief of minor arthritis and its symptoms. 12(c) Aspercreme is a more effective drug than orally-ingested aspirin for the relief of minor arthritis and its symptoms. (3) 12(d) Aspercreme is an effective drug for the relief of rheumatic conditions and their symptoms.

12(e) Aspercreme acts by directly penetrating through the skin to the site of the arthritic disorder.

12(t) The use of Aspercreme.will result in no side effects. Thompson had a reasonable basis for representations listed in Complaint Paragraphs 12(a)-(f) at the time they were made. The complaint alleges that representations listed in complaint Paragraphs 10, 14, and 16 are false, misleading, and deceptive. It further alleges that Thompson lacked a reasonable basis for the representations listed in complaint Paragraph 12.

The case was assigned to Administrative Law Judge Montgomery K. Hyun for hearing. On July 7 1983, Judge Hyun entered his initial decision, which found against Thompson on aji charges except those relating to claims that Aspercreme had no side eflects (eomp. 12(D). The order he adopted requires Thompson to have a reasonable basis for claims that any OTe analgesic drug is effective for the relief of symptoms of musculoskeletal disorder, and for comparative effcacy claims made for such a (4) drug. The order further prohibits Thompson from: (1) using the brand name t!Aspercreme" unless it is accompanied by a disclosure that the product does not contain aspirin; (2) representing that a product is new if it has been generally available for purchase in the United States for more than one year; and (3) misrepresenting any test or study.

Thompson appeals from the ALJ' s findings as to liability. Thompson s principal arguments on appeal are: (1) the ALJ erred in finding that Thompson lacked reliable and credible information constituting a reasonable basis for the effcacy claims it made for Aspercreme; (2) the ALJ erred in finding Aspercreme not to be an effective drug for the relief of minor arthritis or rheumatism pains; and (3) the ALJ erred in finding that either the trademark "Aspercreme" or the product' s advertising deceptively represent that the product contains aspirin. complaint counsel appeal from the ALJ's decision not to prohibit Thompson from using the "Aspercreme" brand name. Com- J Prior to Judge Hyun s decision, the Commis ion adopted a Decisioll and Order ettJjng thc chClrgcR agaiJlst Ogilvy & Mather. In. the MatterofOgliuy& Mather International, Inc.. Dockct No. 9149, Decision and Order issued January 4, 1983. l101 F. C. 1 (1983)) Opinion 104 F.T.C.

plaint counsel also appeal the ALJ's decision to limit Part I of the order to OTe analgesic drugs, rather than an OTC drugs, and not to have Part I prohibit false ingredient representations involving any brand name and any active ingredient.

We generally agree with the ALJ' s findings and conclusion and except as noted in this opinion, adopt them as our own. Our analysis of the issues presented here is in four parts. First, we determine whether or not Thompson made the representations alleged in the complaint. Second, we evaluate whether or not the representations were material. Third, we examine whether or not (5) the claims were likely to mislead consumers acting reasonably under the circumstances.' Finally, in light of our determinations as to the extent and severity of Thompson s deceptive advertising, we consider the appropriate scope of an order prohibiting such conduct in the future. II. DID THOMPSON MAKE THE REPRESENTATIONS ALLEGED IN THE COMPLAINT? 1. Legal Framework In this part of the opinion, we examine whether Thompson s advertising expressly or impliedly conveyed to consumers the representations listed in the complaint. As our discussion below will show, we conclude that it generally did. Before analyzing the advertising itself however, we win set out the standards by which we do so. (6) As we noted in Chffdale Associates we will deem an advertisement to convey a claim if consumers acting reasonably under the circumstances would interpret the advertisement to contain that message. The purpose of such a requirement is to ensure that the flow of useful accurate information to consumers win not be deterred by advertisers' fears that they could be held responsible for claims that they could not reasonably have known consumers were going to receive from the ads in question.

In evaluating what message an ad could reasonably be interpreted as containing, the Commission has traditionally distinguished between express and implied claims. Express claims are ones that directly state the representation at issue. Because the message is stated unequivocally, it is reasonable to interpret the ads as intending to make the claim.6 Implied claims are any claims that are not express. 'As Commissioner Bailey noted in her Concurring and DissentingStlitement inCliffd(lleAsso 'i(ltes. l"c, Docket No 9156 (March 23, 1984), she believes the elements of deceptiun are th"t an ad or practice holVe the tendency or (..pacity to mislead a suhstantial number of consumers in a material way. She did not endorse the Commi ion Cliffdoledescription (slip op. at 7) of the element of deception nor does she endorse it in this opinion. Huwever Commissioner, Bailey agrees that respondent s practices in this case were dec ptiy used itl"'r description of rm. (103 F. C 110 (1984)) ; Id Ii This is rhe only is ue we look at, We do not additionally con ider wheth r the adyerli er inllmded to deceiye consumers with lh claim or whether an objective product claim (i. one not involving puffery or subjectiy value 648 Opinion They range from claims that would be virtually synonymous with an express claim (7) through language that literally says one thing but strongly suggests another to language which relatively few consumers would interpret as making a particular representation. Because of this wide range, the eommission employs two different techniques in evaluating whether an advertisement contains implied claims. One is to look at evidence from the advertisement itself. We often conclude that an advertisement contains an implied claim by evaluating the contents of the advertisement and the circumstances surrounding it. This technique is primarily useful in evaluating advertisements whose language or depictions are clear enough, though not express, for us to conclude with confidence after examining the interaction of aU the different elements in them that they contain a particular implied claim.

If our initial review of evidence from the advertisement itself does not allow us to conclude with confidence that it is reasonable to read an advertisement as containing a particular impEed message, we will not find the ad to make the implied claim unless extrinsic evidence aUows us to conclude that such a (8) reading ofthe ad is reasonable8 For example in this case the conflicting messages in some elements of Aspercreme s ads caused us to examine extrinsic evidence to determine what net impression(s) the entire ad could reasonably be interpreted as giving to consumers.

The extrinsic evidence we prefer to use and to which we give great weight is direct evidence of what consumers actually thought upon reading the advertisement in question. Such evidence wil be in the form of consumer survey research for widely distributed ads, such as those involved in this proceeding. Ads ofthat sort are directed at so large an audience that it is too costly to obtain the statements of enough individual consumers in another manner (e. by way of afjudgment1) is so far-fetched that reasonable consumers would not believe it. Thus, if an ad expressly daims that a shampoo will cure baldness, we presume that reasonable consumers would be deceived by the ad. We presume this even though we might think few people arc unaware of the fact that there is no cure for common baldness. See g, Keele Hair Scalp Specialists, Inc. 55 F. C. 1810 (1959),af(d 275 F.2d 18 (5th Cir. 1960). 7 Adverr.isern('nt do not neceS'arily convey one mcssOige to ajj persons, One subset of consumers reading an ad may interp!' t it to contain one message, while another subset may interpret it to contain a different message. Each interprdation is reasonable as Jong as the subset making it is representill.ive of the group of per SODS to whom the ad is OIddressed.See, e,g, Heinz W. Kirchner 62 F. C. 1282, 1290 (1963) B This longstanding principle of Commission case law was most recently reiterated by us Bristol-Myersin Co. 102 F. C. 21 , 319 (1983), ofrd, 738 F.2d 551 (2d Cir 1981)peli/ionforcert, filed(t'o.84- 650) (Ocl 23, 1984), a case involving deceptive advertisements for DTC internal OImdgesic products: "There also may be instances where daims cannot be inferred from a facial examination of the advertisements and resort to extrinsjc evidence is necessary. See, e.g, Leoflurd F. Porter, lnt".. 88 F. C .54n, 626 (1976)" 9 Similarly, pijstCommissjon cases recognize that affrmative disdosureR can he an etTective method for preventing consumer misunderstanding of ads. However, Commission cases also recognize that such disclosures need to be made clearly and pruminently tu be effective. Whether or not a particular disclosure was clear and prominent is a question fur whose answer we allen would seek information in ijddjtion to that from the ad itself. However where a simple inspection of the ad kayes us confident that the disclosure was ineffective for ordinary consumers we wiJI not require extrinsic information.See, e.g. Litton lrulw;tries, lnc"97 FTC. 1 , 71 , n.6 (1981), afrd modified 676 F.2d 364 (9th Cir. 1982) j g., Opinion 104 F.T.C.

fidavits) to be reasonably confident that the consumers' views on the record of the proceeding were representative of the entire (9) group to which the ad was addressed. Where we use surveys in lieu of individual testimony, the surveys are methodologically sound; they draw valid samples from the appropriate population, ask appropriate questions in ways that minimize bias, and analyze results correctly.10 Another type of evidence we wi1look at is evidence not specifically showing how consumers understood the advertisements at issue before us, but showing how consumers might ordinarily be expected to perceive or understand representations like those contained in the ads we are reviewing. For example, we might look at the dictionary definition of a word to identify the word's common usages. Or we might look at principles derived from market research, as expressed by marketing experts, which show that consumers generally respond in a certain manner to ads that are presented in a particular way, and presume that consumer reactions to a particular ad before us would be consistent with the general response pattern. Where we apply such marketing principles, we will derive them from research presented in references generally accepted as reliable in the field of marketing. Such references may be cited by marketing experts called to testify in the proceeding. Alternatively, we may take offcial notice of the references and cite to them directly in our opinion. (10) A third type of evidence we wi1 consider if ofiered is the opinion of expert witnesses in the proceeding as to how an advertisement might reasonably be interpreted. For example, we might consider the opinion ofa marketing expert who stated his or her view that consumers would interpret an ad in a particular manner. However, where the opinions voiced by experts are not adequately supported we ordinarily give them little weightll Whether we are looking at evidence from the ad itself, extrinsic evidence, or both, we look at the overall, net impression made by an ad to determine what messages it reasonably can be interpreted as conveying to consumersl2 For example, we would look here at whether or not consumers thought, after viewing Thompson s television ads, that Aspercreme contains aspirin. We would not look at how consumers reacted to a particular element from the ad in a different context than the ad itself In this case, Thompson has acknowledged. making several of LOSee, e. our dismssiun uf survey methodolugy in Bristol-Myers Cu. 85 F.T.C. 688, 744 n. 14 (1975). II We consider to be adequately support"," opinions th"t describe empirical research or analyses based on generally recugnizIod marketing principles or other objective manifestations ofprofessionaJ expertise. Opinions not so supported may ""sily be contradicted by the contrary opinions of opposing experts and thus may br of little value in resolving the issue.

See, e. , Bristol.Myers, supra nore 8, at 320; Hor;2rm Corp. 97 T.C 164 , 810 (1981) _., _.. ..._ ._....

J --,-- 648 Opinion the represenations listed in the complaint (RMF 142-43). These are: (11) Complaint Alleged Paragraph Number Representation 12(a) Aspercreme is an effective drug for the relief of minor arthritis and its symptoms. 1:l 12(b) Aspercreme is as effective a drug as orally-ingested aspirin for the relief of minor arthritis and its symptoms. 12(d) Aspercreme is an effective drug for the relief of rheumatic conditions and their symptoms 12(e) Aspercreme acts by directly penetrating through the skin to the site of the arthritic disorder.l Because Thompson has acknowledged making the above-listed representations, we may state without reviewing the ads that Thompson intended to make these claims. Therefore, it is reasonable to interpret the ads as making them. (12) Thompson has not acknowledged making any ofthe other representations listed in the complaint. We therefore examine each of these alleged representations in turn to determine whether or not consumers acting reasonably under the circumstances would interpret Thompson s advertisements as making them.

2. Aspirin Content-Paragraphs 10(a), 16.

Of aji the disputed representations, the one that received the most attention at trial was the issue of whether or not Thompson s advertising, and specifically the trade name "Aspercreme" represented that the product contains aspirin. complaint counsel argued that it did that it represented Aspercreme to contain a cream form of acety- Jated salicylate, as aspirin is defined in the United States Pharmacopoeia. Thompson argued that the ads merely suggest Aspercreme is a drug which works like the more familiar aspirin, not that it actually contains aspirin (RAB 41). Thompson also argued that consumers understand the word "aspirin" as a generic reference to the class of drugs that are pain kiJlers, not to the particular salicylate contained in aspirin tablets (RAB 35-36). Thus, the issues presented for our consideration here are, first, whether one net impression it is reasonable to interpret Thompson s ads as conveying to consumers is that Aspercreme contains aspirin and, second, what consumers would LJ In acknowledg-ing that it made t.his dairn for rl!liefof arthritis, Thompson denied representing that Aspercreme will cure the disease. Thompson similarly denied repre enting that Aspercreme wil cure rheurnati!ltn. The Ar found that Thoml' on hold not made such repre entations (IDFs 131-32, 141-42) L1 We note that lhe complaint included another allegedly deceptive represent-tion. Complaint Paragraph 12(1) charged ThomjJ on with representing that Aspercreme does not result in the adver e reaclion associated with aspirin. The ALl concluded (TDF t43) thatin the context of Thompson s advertisements, the statements about " side effects would be taken to mean "no significant' side efrect, for which representation Thompson had a reasonable basi . Accordingly, the ALJ dismissed so much of the complaint as relates to Paragraph 12(0. We agree with the ALJ's disposition of this part of th," complaint Opinion 104 F.T.C.

interpret the word "aspirin" to mean in the context of the Aspercreme ads. (13) We note to begin that none of the Aspercreme ads includes an express representation that Aspercreme contains aspirin. On the contrary, like much advertising we find deceptive, the ads are drafted with an artful choice of words to make what Thompson thought were literally correct statements (for example, that Aspercreme s pain relieving properties are equivalent to aspirin s), However, in considering the net impression of an advertisement, we do not require that aJl consumers reading or viewing it be sophisticated experts in interpreting the nuances of the English language. Absent reason to conclude differently, we presume that advertisements are directed at ordinary members ofthe adult population who, as such, have a range of abilities. 15 We look at how such (14) individuals actually interpret advertisements in real-life situation, not at how they would if they had suffcient time and incentives attentively to review the ads so as to come up with the most semantically correct interpretation of them. We therefore consider here whether one message impliedly conveyed to reasonable consumers by the ads is that Aspercreme contains aspirin. To do this, we need to evaluate the ads' net impression on consumers. The ads are not all alike; rather they fall into several different groups. We describe each group below and consider whether it gives a net impression of' aspirin content.1 (1) The TV Ads With No Disclosures-exs 1 and 2 Thompson s earliest television ad for Aspercreme, ex 1, shows a woman talking. As the ad begins, she is shown holding two aspirin tablets and saying ". . . imagine being able to put the strong relief of aspirin right where you hurt most. " At this point, the two aspirin tablets in her hand disappear and are replaced by a tube of Aspercreme, while she says: "Now, with amazing Aspercreme, you can get ' If advertioem.mls are targeted OItp"cia! audiences who as a group h,wc 11 greater or lesser cOIpahility to recognize deceptive advertising than urdirliry members ufthe adult population or a distinctive readion t.o partieu- \"r advertising claim . a reasonable consumer is an ordinary members of that torge! audience.See. e.g., TTlIJ Kin", 80 F, C. 71.1 (HI75) However, almust. all advertising is t.argeted at some demographic gruup, sllch as farmers, housewives, or residents ofa particular area. This alone does nol mean that we apply a st.andard different from our customary one Previous Commission I:ases hav, recof-'lized that persons with heart.h- relat.ed prublems can he a target audience See, e.!:., Travel King, id. ; jJOI'ler Dielsd 90 F.TC. 770, 864-65 (1977),arrd, 605 F.2d 294 (7th Cir. 1979),ecrt. denied 445 S 950 (1980). In this case, Thompson s ads were directed at arthritics. The ALJ' s opinion (IDF 90) suggests t.hat Iw considered iIft.hritics to be a target audience. Though his finding was stated in conclusory fashion other available evidence and findings (such as the fact t.hat. arthritis is ;J chronic disease (IDF 191) and that con umers spend significant amounts ufmoney on quack remedies (lDF 193) suggest.s t.hat. arthritics may be more susceptible to deceptive analgp.sic claims than ordinary members oft.he adult population and therefore are a target audience Ultimately we find it. unnecessary tu resolve t.he questiun because the evidence in this proceeding shows that. Thompson s ads were deceptive whet.her or nol arthril.ies arc cunsjdered a target audience '" Our discussiun here focuses suleJy on th" advertisements Thompson used in thp. naljonal consumlOr advertising campaign for Aspercreme (CX 1-- , 21- , and 37). We du riot evaluat.e the Aspercr"me ads in the local Ohio tc"t. (r.X ?m \lPC "S" nur conclusiuns abolli. t.he nat.ional ads makes it. unnecessary 1.0 do so. _.._ .. . , . ,., . 648 Opinion the strong relief of aspirin directly at the point of minor arthritis pain." She continues comparing Aspercreme to aspirin, noting (among other things) that Aspercreme contains none of aspirin s side effects. As the ad ends, she leaves the screen and is replaced by a still life (15) showing a tube and bottle of Aspercreme, along with a video super stating: "The strong relief of aspirin right where you hurt." The audio portion ofthe ad concludes with a voice over stating the same phrase as the video super.

The second early Aspercreme ad, CX 2, is similar to CX 1 at its beginning and end. To begin, it shows a woman holding two aspirin tablets. As she holds the tablets, she tells listeners to "imagine putting the strong relief of aspirin right where you hurt." Next an Aspercreme tube replaces the two tablets, whereupon she states: "Aspercreme is an odorless rub which concentrates the relief of aspirin. At this point, CX 2 diverges from CX 1. In its middle portion, ex 2' visual part uses the transparent outline of a man s body and rays streaming outward from both an aspirin tablet in the stomach and a point on the shoulder where Aspercreme was rubbed. There are fewer rays from the stomach than from the shoulder. As the rays are shown visually, the woman states that "regular aspirin. . . goes through your body like this (through the stomach). But, in seconds, Aspercreme starts concentrating all the temporary relief oftwo aspirin directly at the point of minor arthritis pain." The ad concludes with the same still life and audio as does ex 1. (16) The ALJ concluded that a variety of different elements in these ads were representations that Aspercreme contains aspirin.!' Among the " repetition of the elements were use ofthebrand name " Aspercreme , use of the words " Aspercreme" and " aspirin" in the same commercial statement that Aspercreme provides "the strong relief of aspirin" and the visual image in which two aspirin tablets are replaced by a tube of Aspercreme (IDFs 85-89).

Our own examination ofthe net impressions conveyed by these ads also causes us to conclude that they suggest that Aspercreme contains aspirin. We agree with the ALJ' s view that various elements in the ads suggest aspirin content. We further note the absence of any elments giving a contrary impression, such as express disclosures. Therefore, an the evidence from the ads themselves points to one conclusion: that the ads are likely to give consumers interpreting them reasonably a net impression that Aspercreme contains aspirin. Although we do not find it necessary from our perspective to look \, The ALJ did not distinguish between express and implied represenullions. He also did not focuseparately on the net impression given by CX 1 and 2. Rather, he analyz.cd whether or notindividuaJ word, phrases, or visual images used in one or more ofThompsoll s ads implied aspirin content. A we noted ill the texl, above, our method of analysis Jooks at the net. impression created by the inter"ction of different elementsin a given ad, not at t.he elementsbyt.themselves 794 FEDERAL TRAm; COMMISSION DECISIONS Opinion 104 F.T.C.

at extrinsic evidence in evaluating exs 1 and 2, Thompson has offered some in support of its theory that consumers would interpret exs 1 and 2-and all the other Aspercreme ads, as well-to mean that Aspercreme provides relieflike aspirin (17) without actually containing aspirin as an ingredient (RPF 240-60). Because Thompson offered extrinsic evidence, we are obliged to consider it. However, our review of Thompson s evidence does not change our views regarding exs 1 and 2.

One type of extrinsic evidence offered by Thompson consisted of testimony by marketing experts. (See, e. Ross, Tr. 5983-88; Jasper Tr. 4736; Silver, Tr. 5654-55) and the results of a consumer survey, the Video Storyboard Test (RX 165). Having examined the testimony, we find it unpersuasive.1 (18) Thompson also has claimed that the results ofthe Video Storyboard Test (RX 165) show that CXs 1 and 2 do not leave a net impression of aspirin content with consumers who interpret the ads reasonably. The Video Storyboard Test is, of course, consumer survey research and as such is the type of evidence to which we would give considerable weight ifit were methodologically sound. However, we agree with the ALJ's conclusion (IDF's 117-19) that the Video Storyboard Test was improperly designed for the objective of providing data on whether exs 1 and 2 communicate a net impresssion of aspirin content to consumers.

Although the report of the Video Storyboard Test does not include a protocol describing what the study was designed to measure and how the methodology chosen would do so (in and of itself a factor we view negatively when assessing the quality of a study), it appears that the survey was designed to measure the comparative effectiveness of CX 1 and ex 2 in conveying important attributes of Aspercreme to potential users (RX 165A-M). In order to collect this information, the Video Storyboard Test used a shopping mall intercept sample to obtain 100 participants. The respondents were qualified by age, 45 years and older, and by whether they "occasionally have symptoms For "example, Thompson s principal marketing expert, Dr. Ross, spccul"ted on whether the percentage of consumers coming away from ex, ! and 2 with Lhe impression that Aspercreme contains aspirin rose above an irreducible minimum, but he provided little factual or analytical support for his observations- See e.g, Tr. 5985-6.

Q- Is it possible, Dr. Russ, that some consumers might have taken away the impression from eX- I that Aspercreme is aspirin A. Yes, I think it is possible. I think tlmt some may have and I think that that impression would occur simply from the ""alogy between Aspercreme and internal analgesics, That is, some may think ifit relieves the same way, itb"spirin But I think th"t is a kiod of irreducible minimum percentage kind ofthe;ng of a perception, of a mispercepion, in rhilt I think if you are going to compare any external rub product to any internal product, any internal analgesic product, that. some an.., going tu assume thilt you mean aspirin, since that is the word that they use for such internal tilblets or produd , in perceiving that tlw ad is representing that Aspercreme has aspirin, all the consumer is really thinking or . avin!! is that it has an anaJl!esic in it Tt1U1V1Y::Uj lVlI..U1L;1\.L L;U., l1"l".

648 Opinion of arthritis" (eX 165-0). Fifty participants were shown CX1 and the other fifty viewed ex 2. Immediately after viewing either ex 1 or ex , participants completed the survey questionaire. Although the sample was a non-probability sample (i. its results are not projectable to the national population as a whole) and rather small, there is no apparent reason why the (19) persons surveyed would not be reasonably representative of a typical subset (those over 45 who suffer occasional symptoms of arthritis) of those persons who generally might be interested in purchasing analgesics to relieve arthritis pain. However, ex 51 does not provide probative information about the participants' beliefs about aspirin content in Aspercreme for two reasons. First, it does not contain any questions that directly elicit ingredient information. Second, the verbatim responses were not placed on the record and the data from the study were not coded to determine the percentage of responses that included mentions of ingredients (Ross, Tr. 6310). The questionnaire used in the Video Storyboard Test was concerned with the persuasiveness of each commercial (question 1), the impact ofthe messages or copypoints in each commercial (questions 2, 3, 4, 9, 10, 11 and 12), the perceived uniqueness of Aspercreme (questions 5 and 6), the relative effectiveness of Aspercreme compared to aspirin (questions 7 and 8), and demographics (age and sex of respondents). The way in which these questions were phrased makes any information provided about ingredients in general, and aspirin content in particular, incidental to the information gathered about the relative effectiveness of ex 1 and ex 2 as commercials. Of course, some participants nonetheless may have mentioned ingredients in their responses to the questions being asked. If the record contained information showing how many participants gave such responses, we would examine whether enough had done so to make the results reasonably reliable. However, no (20) such information was supplied with ex 51. Therefore, we find that CX 51 is of no probative value in determining the consumers' beliefs about whether Aspercreme contains aspirin.

While we thus find the extrinsic evidence relied on by Thompson not to be probative, we find that complaint counsel' s evidence does not provide extrinsic evidence consistent with the conclusions we reach from analyzing the ads themselves. Complaint counsel cite to the Mapes and Ross Test (eX 50), which was conducted for Thompson in 1979, prior to the start of this litigation. Although Thompson now criticizes the design of this survey (RAB 53-54), we find that its methodological strengths far outweigh its weaknesses. Specifically, its research methodology (or close variants of it) is frequently used in copy tests. For example, the use of prerecruited groups of subjects lotteries to provide incentives for viewing, the attempts to "bury" the Opinion 104 F.

commercials in programmatic material, and day-after unaided recall, aided recall and brand preference measures are typical of advertising copy tests. In addition, although the study did not employ a probability sample and discussed neither survey response rates nor non-response bias, we cannot detect any obvious reason why the sample is not reasonably representative of consumers viewing exs 1 and 2. Such methodology is adequate for the results to be of so me probative value, assuming the survey asked probative questions. Like the Video Storyboard Test, the Mapes and Ross survey contained questions that attempted to elicit participants' recall of the information contained in ex 1 and ex 2 and did not contain direct questions about ingredients. However, unlike the Video (21) Storyboard Test, the Mapes and Ross final report contains the verbatim responses provided by the participants so that the percentage of respondents who mentioned aspirin as an ingredient can be determined. From the resulting responses come two estimates of the percentage of relevant market females who, given one exposure to a televised Aspercreme commercial, perceive that Aspercreme contains aspirin. The first estimate is contained in the Mapes and Ross final report, which calculates the percentage of survey participants who thought Aspercreme "Works Like/Contains Aspirin" after viewing ex 1 or ex 2. The second estimate is contained in a memo from an Ogilvy and Mather researcher who reviewed the verbatim transcripts ofthe data collected in the study and independently calculated the percentage of respondents who believed Aspercreme contained aspirin (eX 116-A). The data from both sources are provided below: Table 1 CX- CX- Visible Stand- Respondents Perceiving: Man % Presenter % Aspercreme "Works Like/Contains Aspirin" (CX50- Aspercreme product contains aspirin-recaUers only; (Oghvy and Mather memo, CX116- 30 (22) Although there are some biases in the Mapes and Ross Test !9 its results support our conclusion that one reasonable interpretation The test r.contains seveml potential sources of bias which), may nave affect.ed the estimate of the percentage of people w\'o believe Aspercrcmc contains asplri!1 On one sld"" the screeniflg pron,ss used to obtain the sHmp!e may have aJerl.cd panicipants about the nature of the study- If so. respondents wuuld have paid more attention to the commercials than they otherwise might have On the uther side, one exposure to the advertisements is less than the three recommeoded for maximum recall and comprehension in a nat.ural viewing sit.nation. See Krugman Why Thr'''' t:xp"" !lre Muy Be EnrHlfih Journa! of Advert.ising Research, Dec. 1972, 11- 11- The omission of aided questions or prohes fJbout ingredients probably caused the estimates of the proport.ion of respondents believing that, Aspercreme contained aspirin to he undorst.ilted ""''-.L'U ..u . U"""""A'-''UJ ",v., u.

648 Opinion consumers would make of these two ads is that Aspercreme contains aspirin.

Thus the Mapes and Ross Test, the only reasonably reliable extrinsic evidence regarding ex 1- , supports our conclusion that it is reasonable for consumers to interpret exs 1 and 2 as making an implied claim of aspirin content.

(2) The TV Ads With Visual Disclosures-exs 3 and 4 The next two Aspercreme television ads, CXs 3 and 4, are essentially similar to the earlier ads, exs 1 and 2 respectively, with one difference. Both ex 3 and ex 4 have video supers that read: "contains no aspirin." Thus, the question presented by exs 3 and 4 is whether the video supers are suffciently clear (23) and conspicuous to overcome the impact of other elements in the ads, elements which caused somewhere between 21 % and 35% of the persons viewing exs 1 and 2 to believe that Aspercreme contains aspirin.

The extrinsic evidence relating to exs 3 and 4 is not of the sort we perfer to rely on in evaluating the possible existence of implied claims, but is suffcient for us to reach the conclusion that the ads' net impression upon reasonable consumers is one of aspirin content. There is no survey research pertaining directly to exs 3 and 4, and we are not persuaded by the expert opinions presented at trial as to what consumers would interpret the ads to mean (See cohen, Tr. 213-18; Ross, Tr. 6016-18, 6185-98) because we find them to be inadequately supported.'o However, copy testing performed on an Aspercreme ad with what appears to be a more prominent disclosure of non-aspirin content, ex 9, found that one net impression given to consumers was that the (24) product contains aspirin." ex 9 was suffciently similar to exs 3 and 4 that we consider it permissible to infer from the ex 9 results that an even higher percentage of consumers would come away with such a net impression after watching exs 3 and 4.

Additional extrinsic evidence in the form of generally recognized marketing principles also leads to the conclusion that the disclosures were insuffciently clear and conspicuous. The disclosures were placed in the middle of the ads and were distracted from by the "The AI.J's initi 1 opinion also cites the views ofCRS and the National Association of RroOldCfmt.ers (NAB) that the disclosures were inadequate as evidence that the disclosures were so The opinion does not clearly alert readers to the ALl's ruling at trial (Tr. 637) that the opinions orCBS add NAB were to he admit.ted into evidence not as primOiry evidence, but only as confirmatory evidence (Tr- 637) (The opinion does describe the CBS OInd NAB views as "confirmatory, but folds to indicat.e that the word haspeciillOIny ignificance, ) We agree with the AlAI's view at tria! that this evidence cannot be reli",d on to establish in the first instance wlwtl,er or not the video supers were adequate. As was argued by Thompson (Tr- 633-36), the record does not show t.he qualifications ofthe individual at CBS and ?\AB who reached the condu ion that. t.he supers were inadequate, t.he facts that were before these individuals, or the stOindards they "ppliE:d to the facts- Howp.ver, we see no logical basis for the AI.J' s decision to use these opinions as "confirmatory" evidence and, accordingly. do not rely upon them ourselves. " Seeoul discussion ofCX 9, below at pp 33- .. .. . .

Opinion 104 F.

conceptually uncorresponding audio message being communicated while they were on the screen. Studies of consumer behavior show that individuals will better remember either information presented to them first (primary efiect) or information presented to them last (recency effect), depending upon the delay between presentation of the messages and evaluation of the recipients' responses. In any event, however, recall typically is a U-shaped function, with the worst recall being of messages presented in the middle. A distracting situa- 22 (25)tion lowers ability to recall at alj points along the curve. (3) The Print Ads With common Headlines-exs 6-7, and 10-11. Six print ads for Aspercreme were introduced into evidence in this proceeding. Four ofthem, exs 6-7 and 10- , have common elements. Principal among these is an identical headline. The ads begin in large print:

At last! A remarkable breakthrough for arthritis pain: Aspercreme.

Below this and in smaller print, the headline continues: Aspercreme is an cfIective arthritis medicine which concentrates all the strong relief of aspirin directly at the point of pain.

Alj four ads also contain a visual representation of an Aspercreme tube and bottle. Both display the brand name "Aspercreme" and the statement external arthritic pain medication. Two of the ads, exs 6 and 7, have additional elements that would contribute to their net impression concerning aspirin content. Below their headlines, each shows a drawing of a man s body, with numbered lines pointing to various body parts. Associated with each line is text in smaller print that the headlines. CX 6' s text compares Aspercreme to aspirin in two ways: (1) "Aspirin tablets go through your body. But Aspercreme concentrates the relief of an effective aspirin-like analgesic (26) directly at the point of arthritis pain ; and (2) "Aspercreme, like aspirin itself, has no liniment smell." CX 7's text has three comparisons between Asprecreme and aspirin: (1) "Aspirin tablets go throughout your body. But Asprecreme concentrates the relief of two aspirin directly at the point of arthritis pain ; (2) Aspercreme works faster than aspirin because you rub it right where you hurt"; and (3) "Aspercreme, like aspirin itself, has no liniment smell."

,1 See B. Slernthal and C. Craig, Consumer Behavior, An Information l'rocessinf: Perspecliue102 03, 282 83 (1982) ami work cited therein 648 Opinion We conclude that it would be reasonable to interpret aU four print ads as conveying a message of aspirin content. For three of the ads, exs 7, 10- , we reach this determination based solely upon our examination of evidence from the ads themselves. In particular, we conclude that the language in the headline ("Aspercreme . . . concentrates all the strong relief of aspirin. . . ) is readily susceptible to the interpretation that Aspercreme contains aspirin and that the language would be so interpreted by consumers. Given this fact, and given that no other elements in these ads suggest other than that Aspercreme contains aspirin, we conclude that at least one of the net impressions communicated by the ads is of Aspercreme s aspirin content.

One of the ads, CX 6, does include textual language amounting to an implied representation that Aspercreme does not contain aspirin. It states: "Aspercreme concentrates the relief of an efiective aspirinlike analgesic. . ." The most direct conclusion to draw from a statement that Aspercreme is "like" aspirin would be that it is similar to but not the same as aspirin. Because this element in the ad conflicts with the (27) element stating that Aspercreme "concentrates all the strong relief of aspirin " we require something in addition to evidence from the ad itselfbefore determining whether or not consumers would interpret ex 6 to state that Aspercreme contains aspirin. The additional element in the case of ex 6 is the general marketing principle, acknowledged by Thompson s marketing expert (Ross, Tr. 6198-99) as weU as complaint counsel' s (cohen, Tr. 223), that persons reading a print ad often will read only the headline, and wjJ take their sole impression of the ad from it. The special significance of headlines has previously beeen recognized in Commission cases which hold that even an express disclosure in the text of an ad may not be enough to change the ad's net impression upon consumers. Applying this principle to CX 6, complaint counsel' s marketing expert testified that a great many people reading ex 6 would believe Aspercreme to contain aspirin (Cohen, Tr. 233).21 We also find that application of the general principle to this specific situation suggests such a (28) result. Upon this basis, we conclude that one net impression consumers reasonably would understand ex 6 to be conveying is that Aspercreme contains aspirin.

23 See, e. , Litton Indu.stries, Inc. 97 F. 70 5 (1981), Giant Food, Inc.61 C. 326, 348-9 (1962),affd 322 F.2d 977 (D.c. Clr. 1963),cat. denied 376 S- 967 (1964). ' Thompson s marketi!,!;; expert argued that the headline itself does not imply aspirin content (Ro . Tr- 6020-21, 6060- 61), frum which it would follow that persons reading the headlines alone would not believe Aspercreme contains aspirin. While we are obliged to consider this opinion beclIuscit was offered as evidence, we find unpersuasivc.

. ,, Opinion 104 F.

(4) The Other Print Ads-exs 8 37.

The other two print ads, exs 8 and 37, are different from the other four and from each other. We first examine ex 8. Beginning with its headline, ex 8 repeatedly contrasts Aspercreme with aspirin. The headline states:

There s always been aspirin. . .

Now there s ASPERCREME25 Works faster, safer than aspirin rclievcs pain in minutes. The first sentences ofthe text, (in considerably smaller print than the headline) continue this comparison approach by stating: "Aspirin has been helping sufferers of minor arthritis pain for years. Now there a different way to get relief ASPERCREME." Other comparisons in the text similarly contrast Aspercreme with aspirin. In addition there is a visual depiction of an Aspercreme tube in its display packaging. Visible on the display packaging, in print slightly smaller than the text ofthe ad, is the statement: "An effective aspirin-like analgesic for temporary relief of occasional minor pain. . . " (29) We are not able to conclude with adequate confidence by looking solely at evidence from the ad itself whether or not one message conveyed to consumers by ex 8 is that Aspercreme contains aspirin. The general tone ofthe ad contrasts Aspercreme with aspirin, emphasizing the supposed diflerence between the products rather than their similarities.26 These contrasts might create among average consumers the impression that Aspercreme is different from aspirin. On the other hand, it might also be that the perceived difference would be means of application (external rather than internal), not identity of active ingredients.

In this situation, we cannot find the ad to convey an implied message without suffciently probative extrinsic evidence upon which to base our conclusions. We do not find any such evidence here. There was no consumer survey research on ex 8 or any other Aspercreme advertisement similar enough to ex 8 to permit the drawing of inferences about ex 8 from it. The testimony of the marketing experts also does not permit us to reach reasonably confident conclusions. eomplaint counsel's principal marketing expert argued that most people looking at a newspaper ad such as ex 8 would only look at the headline, which links Aspercreme with aspirin (cohen, Tr. 226), but he did not show that the link created by such a headline would equate Aspercreme and aspirin; instead, consumers might interpret the link to suggest that Aspercreme is a new product similar to but (30) different 5 This part of th headline i in significantly larger print than other parts :! 'Ih" ntl,pr Aon"" rpmp "rl WP h "" .."vi"w"rl I1n In "OW "mnl, ;.,,,rl Qim;br;t;"c p" " o ..,,1;,, THOMPSON MEDICAL CO., INC. /jul 648 Opinion from aspirin.2? Thompson s principal marketing expert, on the other hand, emphasized the fact that CX 8 contrasts Aspercreme with aspirin, concluding from this that if consumers understand aspirin as a specific ingredient, the headline would communicate that Aspercreme is not aspirin (Ross, Tr. 6026-27). However, he provided no evidence to exclude a possible consumer interpretation of the headline that Aspercreme provides the clinically active ingredient of aspirin tablets in a cream form. While both sides have thus failed to present adequate extrinsic evidence concerning CX 8, the burden was on complaint counsel to do so. In light ofthat failure, we cannot find CX 8 to contain an implied message of aspirin content. We next turn our attention to the remaining print ad, CX 37. The headline of ex 37 characterizes Aspercreme as "a remarkable analgesic" and as "an effective alternative to pils for minor arthritis pain." The first paragraph of the text, which is in bolder print than the remainder, states: !!Aspercreme contains salycin, an effective non -aspirin pain reliever that concentrates relief right at the point of pain. You get hours of relief without the side effects pils may cause. Further on, the text additionally states: "Aspercreme has a non-aspirin pain reliever, so it can t cause aspirin s stomach upset or any of aspirin s side effects" and HAspercreme s active pain reliever-(31J salycin-is clincally proven. . ." (emphasis in original). All of these elements suggest a difference between the active ingredient in Aspercreme and the active ingredient in aspirin. Nor is there any element in the ad except for the brand name that even arguably could be viewed as implying that Aspercreme and aspirin are the same. It may be that the brand name alone would be enough to convince some readers that the product contains aspirin, and that the disclosures would be insuffcient to overcome this misimpression. However, we would require that such a conclusion be supported by extrinsic evidence.

Our conclusion diverges from the ALJ' s, who found that ex 37 did represent aspirin content (IDF 84). However, the ALJ's discussion of this particular ad is flawed by his analytic approach, which was to look at each element that was in one or more ads and then to decide whether or not the element, considered by itself, represented aspirin content. Not only do we believe this approach to be less analytically desirable than considering the net impression which different elements in an ad combine to give, but in this case the approach caused the ALJ to lump ex 37 together with several of Thompson s television advertisements, all of which are quite different from ex 37. Thus, the 27 We will not presume generally that consumers interpreting advertisements reasonably under the circum. stances wil always read a comparative ad to imply that the products being compared are the same. For example we seriously doubt that persons reading "Buy a racing car. It provides faster, more enjoyable transportation than a station wagon," would interpret the ad to imply that the products were the same Opinion 104 F.

ALJ incorrectly wrote that the affrmative statements in several ads including ex 37, were required by the television networks. Obviously, the networks had no control over print ads such as CX 37. Likewise where the ALJ cited extrinsic evidence as to why the affrmative statements in various ads were ineffective disclosures (IDF 84), the evidence itself pertained only to the television advertisements, not to ex 37 (32) The one point made by the ALJ that appears specifically to apply to CX 37 is his observation that the phrase "contains salycin, a nonaspirin pain reliever " found in both CX 37 and several television ads is ambiguous because it does not negate the impression that Aspercreme may contain both aspirin and "salycin." We find this point unpersuasive for two reasons. First, while it is literally true as a nicety of English semantics, that the statement "contains salycin would not absolutely negate the possibility that Aspercreme also contains aspirin, nothing in ex 37 expressly states that Aspercreme has more than one active ingredient. Nor do we find that the brand name Aspercreme" or any other element in ex 37 implies the product contains two active ingredients. Therefore, in considering how consumers probably would interpret the ad, rather than how it is technically possible for someone to read it, we find it irrelevant that the statement in CX 37 does not specifically exclude the possibility of two active ingredients. Second, even if one thought average adults were sticklers for English semantics, there is a statement in ex 37, one which the ALJ overlooked, that does negate the possibility of aspirin content. It is the statement that Aspercreme s "active pain (33) reliev- " is salycin. Since the ad refers to "pain reliever" in the singular it excludes the possibility of there being a second active ingredient. (5) The TV Ads With Audio Disclosures-exs 5, 9 and 21- The most recent television ads for Aspercreme, the ones on which most of the Aspercreme advertising dollars were spent (CX 25A), contain audio disclosures that either expressly or impliedly state that Aspercreme does not contain aspirin. Three of the ads, exs 5, 9 and , are testimonials whose main focus is to show various men and women endorsing the product. Among the individuals' statements are those that Aspercreme relieved their arthritis, did not upset their stomachs, was odorless, or relieved their pain for hours. These ads do not mention the word "aspirin " liken Aspercreme s relief to that provided by aspirin, or liken Aspercreme to "pills." Each also contains a disclosure in a middle frame ofthe ad (rather than at its beginning or end) stating that Aspercreme does not contain aspirin. In ex 5, it ,-" The name "Aspercreme " mOlY imply that Aspercremc contain ospjrin "s an "dive ingredient- Howevtof. lhe nan.., does not suggest whether or nut th", product has more than one active ingredient THOMPSON MEDICAL CO., ING 648 Opinion is the express audio statement ttaspirin free " coupled to a visual image of an Aspercreme tube. In exs 9 and 21, it is the statement Aspercreme contains salycin, a strong non-aspirin pain reliever. coupled to the visual image of an Aspercreme tube and the video super contains salycin.

In the remaining TV ad, ex 22, a woman compares Aspercreme with "pils " stating that Aspercreme provides long lasting relief without the side effects pils may cause. Insofar as "pils" is an implied reference to aspirin tablets, ex 22 is an (34) ad that contrasts the two forms of medication. In the middle frames of the ad the woman states "Aspercreme relieves arthritis pain because it contains salycin, a strong non-aspirin pain reliever." while the visual portion of the ad shows her applying Aspercreme to the back of one of her hands.

We would not find exs 5, 9 and 21-22 to convey an implied aspirin content message if we were to rely solely upon an examination of evidence from the ads themselves. The contradictory elements in these ads preclude us from determining with suffcient certainty what message(s) viewers would take away from the ads concerning the aspirin content of the product. On the one hand, there is the disclosure that Aspercreme contains a non-aspirin pain reliever, coupled to the absence of any language or visual imagery likening the product to aspirin tablets. On the other hand, there is the "Aspercreme brand name itself. The brand name was visually displayed in each ad as well as being repeated beween four and six times on the ads' voice tracks. Moreover, the brand name of a product is the most powerful single stimulus in an ad. eonsumers are more likely to recall brand names than specific copy points in advertisements (cohen, Tr. 559; Ross, Tr. 6317-19). A net impression based primarily upon a message derived from the brand name could be that Aspercreme contains aspirin.

In this situation, we require extrinsic evidence to help us determine how reasonable consumers actually perceive the ads. The record contains three surveys that we believe shed light on the question of what Consumers believe CX 9 to mean: the FRe Test (eX 35/RX 520), and the Lieberman Test (CX 32/RX 500), which (35) were both sponsored by Thompson, and the ASI Theater Test (eX 27), which was sponsored by complaint counsel. The results from each of these studies and significant details about their methodologies are summarized below. However, the conclusion we draw from the surveys can be stated at the outset: consumers interpreting ex 9 reasonably could and did think that Aspercreme contains aspirin.

Opinion 104 F.

(a) The FRC Test (CX 351RX 520) and the Lieberman Test (CX 321RX 500) In 1981 Davis and Gilbert, Thompson s counsel, contracted with Ken Warwick and Associates, Inc. to design a copytest ofCX 9. Based on background information contained in the FTC complaint and discussions with Davis and Gilbert, Warwick desigoed and conducted the FRe copy test (eX 35/RX 520). After the FRC study was completed and the results were in, Warwick redesigned the survey questionnaire used in ex 520 and completed a second study, the Lieberman copytest (CX 32/RB 500).

Both the FRC and Lieberman studies were intended to determine the percentage of respondents who, after viewing ex 9, believed Aspercreme contained aspirin. The methodology was the same for both studies and involved showing respondents, who were recruited and screened with shopping mall intercepts, the advertisement CX 9 and asking them a short series of questions. Respondents were shown ex 9 twice and no distracting materials such as other ads were included. This methodology increases the probability that respondents paid attention to ex 9. Thus, if(36) there was any bias in the methodology used in the Warwick studies, it was in the direction of increasing the percentage of respondents who remembered the information presented in ex 9.

Although there were few differences in the research design between the FRe and Lieberman studies, there were important substantive differences in the survey questionnaires used in each study. The questionnaire in the first study contianed four questions probing consumer perceptions as to the ingredients of Aspercreme and two other analgesic rubs:

Q1. First, what was the name of the ingredient in the product you just saw advertised? Q2. Based on the commercial1 you just saw, does the product in the commercial contain aspirin? Q5. Does Ben-Gay contain aspirin'? Q6. Does Mcntholatum contain aspirin? (RX 520; Warwick, Tr. 5332) The second study contained only one question probing such consumer perceptions:

Ql. What was the name ofthe ingredient in Aspercremc, the product advertised? (RX 500- Findings from the FRe and Lieberman studies are summarized in Table 2: r371 /., 648 Opinion Table 2 Unaided and Aided Recall Results to Questions About the Ingredients in Topical Analgesics Lieberman Ingredients Mentioned FRC Survey Survey With Unaided Recall Question 1 '" Question 1 ** Salycin 95 (45.9%) (25%) Aspercreme 43 (20.8%) ( 2%) Aspirin 6 ( 2.9%) ( 5%) Blank/Don t Know 47 (22.7%) (65%) Others 16 ( 7.%) ( 3%) n (totals) 207 (100%) 212 (100%) Aided Recall Perception of Whether FRC Survey Aspirin is an Don Ingredient in... * Ves Know Total Aspercreme (question 2) 22. 67. 10. 99. Ben-Gay (question 5) 39. 1 "' 54. 100. Mentholatum (question 6) 43. 51. 100. . RX 520- Warwick, Tr. 5321.

The results from the unaided recaJi questions (the first question contained in each survey) indicate that between 25 and 46 percent of the participants identified Salycin as the ingredient in (38) Aspercreme. Thus, less than half of the respondents were able to provide this response after seeing the commercial twice. However, in both surveys very few ofthe respondents mentioned aspirin as an active ingredient in Aspercreme. Thus, the unaided recaJi question in both surveys tends to support Thompson s position.

The results from the aided recall questions in the FRC survey (questions 2, 5 and 6) tell an entirely diflerent story. That is, when queried directly about whether Aspercreme contains aspirin, 22 percent ofthe respondents replied affrmatively. eonversely, only 6 percent of the respondents thought aspirin was an ingredient in Ben-Gay and less than 5 percent perceived that aspirin was contained in mentholatum. These results show that CX 9 did, in fact, cause average viewers to believe that the product being described contains aspirin. Specifically, the difference in the response percentages between Aspercreme, on J No infurmation i conlainnd in the record explajning why the percentage ofrespundfmL who identified Salycin as the active ingredient in Aspercreme varied so dramatically between the FRC and Lieberman studies Although the wording ofthp. unaided recall question (Ql) was changed between the two surveys to decrease the percentage of participants who responded "Aspercreme" (Warwick, Tr. 5325), there is no reason to expect that this change would influence the percentage of "Sa lye in" responses- We would expect that the change in wording in question I between the FRC and Lieberman studies would reduce t.he percentages of participants giving the response Aspercreme" from 20 to close to 0 percent and cause a corresponding increase in "dun t know" responses from 23 to 43 percent" What we find is that the Lieberman study obtained roughly the decline expected in Aspercreme responses e" from 20 to 2 percent, 20 percent more "don t know " responses, and 20 percent less Salycin responses than expected based on the FRC results Opinion 104 F.

the one hand, and Ben-Gay and Mentholatum, on the other, allows us to conclude with reasonable (39) confidence that the results of the FRe Test were not caused by survey bias. The questions about the content of Ben-Gay and Mentholatum acted as controls for the survey, and permitted an estimate of the "noise level" generated by various possible random factors, such as confusion on the part of consumers being surveyed or "yeasaying bias" in response to aided questions. The significant difference between the percentage of consumers who thought Aspercreme contains aspirin and those who thought it was contained in the other rubs clearly supports the conclusion that 9 generated a net impression of aspirin content among its viewers. (40) (b) The AS! Theater Test (CX 27) The effectiveness of the disclosure in ex 9 that aspirin is not an ingredient in Aspercreme was also examined in the ASI Theater Test (eX 27) sponsored by complaint counsel. The methodology involved recruiting two groups of subjects to a theater for the ostensible purpose of evaluating a pilot television show. Each group of subjects was shown the same t!pilot" show, a standard control commercial for Papermate Flair pens and one of two topical analgesic commercials one for Aspercreme and one for Mobisyl (CX 27-E). The next day as many viewers as could be contacted were telephoned and asked unaided and aided recall questions about the products advertised in the theater test and the ingredients they might contain (eX 27- H). It is likely that the responses to the ASI Theater Test show a bias toward relatively increased estimates ofthe confusion about the information contained in ex 9. This is due to the fact that CX 9 was shown only once, that it was shown along with programmatic material and another commercial, and that participants were not asked questions about the ingredients in these topical analgesics until the next day. Thus, it is likely that the ASI Theater Test resulted in an inflated '1 If roug"hly no more respondenl had answered that Aspcrcreme contains aspirin than answered that Ben-Gay or Mentholalum did, we would condude that the extrinsic evidence did not show ex 9 to give a net irnpres.ion ofaJpirin wntent. However, three times as many respondents perceived Aspercrcme to contain aspirin even after twice viewing a digelosure to the contrary, than perceived that aspirin was an ingredient in Ben-Gay or Mentholatum j! Dr- Warwick, the expert who originally designed the FRC Test for Thompson, argued that it was flawed through use of aided recall questions and that this is the reason why he designed a second test omitting them. However, we agree with the ALJ' s observations (IDF 109) on why the second study, the Lieberman Test, omitted the aided recall questions found in the earlier study: Although Dr. Warwick testified that this was an improvement over the FRC Test design, which he character. ized as "flawed " the evidence is also consistent with the conclusion that the direct ingredient question was dropped because it had produced results unfavorable to Thompson in the FRC Test. Dr. Warwick's testimony to the contrary is unpersuasive in light of the fact that the aided recall questions are widely used in advertising research and the fact that there are numerous ways of reducing the yeasaying hias Dr. Warwick arg-ed such questions can create. For example, one can list controls as Dr. Warwick himself had done in the FRC Test , , 648 Opinion estimate of the percentage of respondents who, as a result of ex 9 believe aspirin is an ingredient in Aspercreme (or conversely, a deflated estimate ofthe effect ofthe affrmative disclosure that Aspercreme does not contain aspirin). The results from the unaided and aided recall questions contained in the ASI Theater Test are shown in Table 3 below: (41) Table 3 Percentages of Participants Who Though Ingredients Were Contained in Aspercreme and Mobisyl Unaided Recall Mentions of Ingredients In Each Product:

Aspercreme Mobisyl Commercial Commercial Viewers: Viewers:

1 % 17%AspirinSalycinAll other ingreds. No aspirinDont know 68NoNone answer (100) (Number of Participants) (130)the Aspercreme commercial, as The unaided recall question was Now thinking about best you can recall, what ingredient or ingredients, if any did the commercial say Aspercreme contained?", followed by the probe, "Were there any other ingredients mentioned in the commercial for Aspercreme?" (CX 27-Z115) Percentages of Respondents to Aided Recall Agreeing That Various Ingredients Were Contained in These Products:

Aspercreme Mobisyl Commercial Commercial Viewers: rs: 38%AspirinSalycinHydrocortisone 20 32Lanolin 23Menthol(Number of Participants) 1 5 (110) The aided recall question was m going to read you(130)a list of ingredients. As I read each one, please tell me if the commercial indicated whether that particularingredient is present in the product even if you mentioned it to me before. " (CX 27-Z115) (42) The results of the aided recall questions are consistent with the hypothesis that a yeasaying bias exists, because a relatively large percentage of participants indicated that the control ingredients of Opinion 104 F.

hydrocortisone, lanolin and menthol were ingredients in both Aspercreme and the control product, Mobisyl. Nonetheless, the unaided recall results indicate that a sizable percentage of participants did not perceive or remember the disclosure that Aspercreme does not contain aspirin and that a much larger percentage of participants associated aspirin than salycin with Aspercreme. Similarly, the aided recall results indicate that a much larger percentage of those who viewed ex 9 believed Aspercreme contains aspirin than do those who saw the Mobisyl commercial. In addition, despite the likelihood of yeasaying bias which appears to exist in the ASI Theater Test data the overall results indicate that the net impression conveyed by ex 9 to at least one group of average listeners was that Aspercreme contains aspirin.

(6) What ttaspirin " means to reasonable consumers. In our above discussion, we found that most of Thompson s advertisements have implied representations that Aspercreme contains aspirin. As we noted to begin with, however, one argument Thompson put forward for the proposition that such representations are not deceptive is that consumers understand "aspirin" as a generic description for products that are pain relievers, not only products containing acetylsalicylic acid. Thompson s argument raises the question of whether ordinary or (43) average consumers would interpret the word "aspirin" to imply a specific ingredient or whether they only net impression they would receive would be that of a generic pain reliev- , leaving only an unrepresentative subset of consumers to think of aspirin as a specific drug.

The ALJ's initial opinion did not specifically address this point perhaps because Thompson did not raise it before him with any great degree of clarity (See, e. RMF 121, RAB 5). Thompson s argument as we interpret it is that consumers consider "aspirin" to be a generic reference to analgesics, so that representations that Aspercreme contains aspirin do not refer to acetylsalicylic acid. To support the allegations in Paragraphs lO(a), l1(a), 16 and 17 of the complaint-which allege that Thompson falsely represented Aspercreme to contain aspirin- Thompson s argument suggests complaint counsel must show that the product does not contain an em,ctive analgesic. The controlling question presented by this argument is what reasonable consumers understand "aspirin" to be. It is insuffcent for these purposes to note simply that scientists and medical experts understand "aspirin" to mean acetylsalicylic acid, since average COli- J' Claim that A percr"'IIe orTEAlS are effective analgesics are not alleged in the complaint to be false, but to be unsubstanli"ted- Thus there is little direct evidence offillsity, and we would have diffculty concluding that Aspercreme ur TEAlS h"ve been proved ineffective g, \_ .l.l. n.l .iW.U.l.nV.n.l v..... , .L.L.v. 648 Opinion sumers might not understand "aspirin" as would a scientist or doctor.33 The conflicting expert opinions (44) offered by the parties on this question, drawing on research designed for other purposes, pro- (45)vide little assistance.34 There are, however, three reasons why we are willing to conclude that at least some consumers within the mainstream do not consider aspirin" to be a generic reference to all types of pain reliever. In the first place, Thompson s own ads repeatedly distinguish between Aspercreme and aspirin. Indeed, several specifically state that Aspercreme does not contain aspirin, or that it contains salycin. Thus, the evidence from Thompson s own advertising indicates that asirin is not a generic reference to all analgesics. Second, for aid in interpreting the common meanings of a word we occasionally look to dictionary definitions. They are derived from the ordinary usage of words and as such, are of some use to us as indication of how (46) reasonable consumers would understand these words. In this case, we have 13 On the contrary, wr: note thatcientific and popular U!Jder tanding "re known to vary un on. ion. For example, an average pr:rson woulrl consider a spider to be ;In in eet. To an entomologist, however, spider a.re an order (Araneida) of animal in the da of Arachnida, whereas in acts are animal in an entirely eparate da (Insecta) of the phylum of articulate invertebrate (Art.hroporla). J1 Complaint Colli el' witness testified that the finriingR in ex 26 (whose mr:thodology we riiscuss below, at note 90), indicated con umers refer to aspirin as a specific type of internal analgr:sic rather thun a a generic term for ate pain relievr:r (Cohen, 1'1'. 163). However, CX 26's finding do not allow conclusions ofthi type to be rirawn. The objective of ex 26 was to "find out whether or nor the name Aspercreme led consumers to the inference that this product contained aspirin as an ingredient" (CX 26-B). lts objective was not to find out what consumers meant by ""spirin" when they stated that "Aspercreme " Ruggested aspirin content. There clcilrly is nothing in CX 26 which would allow inferences about whether ron.sumers consider "aspirin " to refer to one spr:cific type of internal analge ic or to be a generic phrase for many type of pain relievers with different chemical formulations Thompsun witness cited to the Mentholatum fucus groups a supporting the opinion that people use the term pirin " tu refer to a variety of analgesics with dii"en,nt d,emical formula"tioos (Ross, Tr. 5972-76). However, the Mentbolatum focus group results, while supporting his testimony, conDict with tbe results contained in the reports from the Schneid!'r (CX 52) and Nicholas (CX 53) focus groul' , which suppurt complaint cuunseJ's pusition because participants in th"'n believed Aspercreme cuntained aspirin and distingui hed between aspirin and other pain relievers. See discussion below. at note 82. In any event, focus groups are not a research tool whose methodology permit use of their result the bilsis for drawing generalizable conclusions.Id. Thompson s witne s "Iso cited to the ilnswers participants in the FRC Test (CX 35/RX 520) gave to the study Question:1 Cis evidence that consumers understood "aspirin " as a generic pain reliever (Ru8\, Tr. 5976-79). Question 3 asked "What is m;pirin I" The most frequent responses characterized it a " pain reliever (55%), a pain killer 01%). or as something that relieves headaches (7%) or pain ("%). Relatively few participants (2%) described it by giving a particular chemicill name(e. salicylic acid) ur otherwise characterized it as a chemicill, suhstance or compound (2%) (RX 520B.J). Tbompson s witne s argued from this that fewer than 8% of consumers associate the word "aspirin " with a specific ingredient (RX 520C-D; Ross, Tr. 5978). However, Question 3 was too ambiguou ly framed to permir drawing such conclu ions from it. The fact that many participants first associated the word aspirin " with pain rr:lir:fdoes not nece sarily mean that they thought of the drug as a generic reference to all pain reli( vers. any more than the fact that people might fir t describe iI Buick Cis a "car " would necessarily mean that they think "Buick" i " generi" .reference to automobile . Further probing would have been necessary tohow whether or not consumer distinguish between aspirin and analgesics based on other chemical compounds Indeed the answers to Question" uf the same study slrongly suggest that person who characterized aspirin as a pain reliever in response to Que tion 3 did not necessa.rily think of the word as a generic reference to analgesics. Question 4a asked participant if they knew tJ,e chemical name for aspirin. Tbe relatively small numher (14) who answered "yes" were asked by Question "b what the name was. Several ofthe participants who correctly described aspirin s specific chemical name as acetylsalicylic acid were among those who had responded to Question.1 by silying aspirin was a pain reliever (RX 5201.; RX 517A- J(; Definitions arc less reliable than survey research a an indicator of how consumers underst.and advertisements because they can only provide the meaning generally used for words, rather than the specific meanings of the wurds in a particultu cootext. The lImge ofa word in an advertisement may, of course, communic"t.r: iI meaning at. variance with the word's diction"r)' definitions, such as when it is used as slang ('" Vou can drive this lovely, (footnot.e cont'd) g., , Opinion 104 F.

examined two widely used dictionaries, both of which define "aspirin either as acetylsalicylic acid or as aspirin tablets. Neither of the dictionaries defines it as a generic reference to pain kiJJers.36 In the third place, we note the extensive advertising contemporaneously with Aspercreme of analgesics whose active ingredient is something other than aspirin acetaminophen (CX 54-Z042 to Z047). Ifthere ever was a time when the only analgesics being marketed had aspirin as (47) their active ingredient, so that reasonable consumers might only think of "aspirin" as synonymous with pain killer, that time appears to us to be long past.

We recognize that consumers today might not be aware that aspirin s proper chemical name is, in fact acetylsalicylic acid. " Nor do we exclude the possibility that reasonable consumers interpret the word aspirin in more than one way, with some thinking of it as a generic pain killer in the manner suggested by Thompson. Indeed some consumers probably both are unaware of aspirin s chemical name and think of it as a generic pain kjIer. What we conclude, however, is that those persons who viewed or read the Asprecreme commercials and interpreted them to mean the product contains "aspirin " and who further thought "aspirin" is a specific chemical, fact within the range of persons who would be average or ordinary members ofthe adult population and, as such, are reasonable consumers. We therefore conclude that one net impression conveyed by Thompson s commercials is that Aspercreme contains acetylated salicylate. 3. Recent Development of Aspercreme-Paragraph 10(b) The ALJ found that Thompson has represented Aspercreme to be a recently discovered or developed drug product (IDF 126, ID 135-36). Thompson has not argued on appeal that this finding was in error. (48) We agree with the ALJ that the claims were made, but view them as implied rather than express. The representation that Aspercreme is a new product is contained in the headlines of five print ads, exs 6-8 10-11. Four of those ads, exs 6-7 10- , contain identical lanlate model car hOITI' for JURt twu thousand five hundred bananas. ") However, we will consider dictionary definilion to be indicators of" ward' H meaning unless other extrinsic evidence or our own examin"tion of the advertise. ment gives us reason to think that the context of a word IT"kes dictionary definitions inapposite. It appears that the word "aspirin " is being used in CXs 3 and 4 in an ordinary manner, Therefore, reference to defjnjtion to learn how consumers might interprel the word appears appropriate. 3" Web t"r Third ew International Dictionary 130 (1976) define pjrjn" as follows. 1: a white crystaUine compound CI-L,COOC,; COOH of salicylic acid used e p, in tablet form as an antipyretic and analgesic like the alicylO1tes but producing fewer undesirable efTcct&-alled also acetylsalicylic acid 2: a tablet ofaHpirin A Supplement to the Oxford Engk;h Dictionary, Vulume J 135 (1972) describes "aspirin" in this manner. A white cry talline compound, acetylsalicylic acid, u ed c p. as an analge ic ,md antipyretic; withan and a dose of this in tablet/arm. Also"Urili . , ._. , . .

648 Opinion guage: "At last! A remarkable breakthrough for arthritis pain: Aspercreme." The headline of the other ad, CX 8, states: "There s always been aspirin. . . Now there s Aspercreme." Both headlines create an implied representation that Aspercreme is a new product by use of phrases that suggest Aspercreme was not hitherto available for purchase. The logical inference is that it was unavailable for purchase because it did not exist. Because the language used in the headlines clearly conveys a message of newness, and because there are no other elements in the ads which might alter the net impression created by the headlines, we find that the implied claims were made. (49) 4. Aspercreme Is More Effective Than Aspirin Tablets- Paragraph 12(c).

The ALJ found aji of the nationally distributed Aspercreme ads to represent that Aspercreme is more effective than aspirin tablets because it works faster than aspirin tablets, or works without aspirin side effects such as upset stomachs, or both (IDF 133). Thompson has asserted on appeal that its advertisements contained no such comparative superiority claims (RAB 5 , 33).

Our own review of the record leads us to agree with the ALJ that seven of the ads, CXs 1-4, 7- , and 37, made superiority claims. However, we conclude that there is inadequate information on the record from which to determine whether the remaining ads cited by the ALJ CXs 5-6 10-11, and 21- , would be understood by ordinary consumers to state that Aspercreme is more effective than aspirin tablets. The seven ads that we find to make comparative effcacy claims do so expressly. They contain either or both of two different superiority claims: (1) Aspercreme provides faster relief than aspirin tablets; and (2) Aspercreme does not cause the side effects of aspirin tablets. One ad, the print ad CX 8, makes both claims. The headline ofCX 8 states that Aspercreme "Works faster, safer than aspirin-relieves pain in minutes. " (50) Three other ads, the TV ads exs 1 and 3 and the print ad ex 37 claim that Aspercreme lacks the side effects of aspirin tablets. In exs 1 and 3 , an announcer tells viewers that Aspercreme has none of aspirin s possible side effects." In ex 37, the text of the ad states that Aspercreme has a non-aspirin pain reliever, so it can t cause aspirin s stomach upset or any of aspirin s side effects. In yet another three ads, the TV ads CXs 2 and 4 and the print ad ex 7, the claim is made that Aspercreme works more quickly than aspirin tablets. In CXs 2 and 4, it is made expressly by a combination of visual and audio elements. As an announcer states "When you take regular aspirin, it goes throughout your body like this. " the visual portion of the ads show the outline of a man s body with a few rays Opinion 104 F.

streaming outward from the-stomach. One of these rays is shown reaching the man s left shoulder. The announcer then continues But, in seconds, Aspercreme starts concentrating aU the temporary relief of two aspirin directly at the point of minor arthritis pain while the video portion of the ad shows a heavy concentration of rays streaming outward from a point on the left shoulder to the rest of the shoulder. Given the announcer s claim that the outlines compare the ability of Aspercreme and aspirin to provide relief, the fact that the visual images show Aspercreme providing more rays to the shoulder area than aspirin tablets at the same point oftime makes an explicit message that Aspercreme provides reliefto the area of arthritis pain more quickly than aspirin tablets. In the print ad, ex 7, the text states that Aspercreme (51) "is actuaUy faster and more efiective than aspi- It also states that Aspercremerin in relieving minor arthritis pain. works faster than aspirin because you rub it in right where you hurt, The remaining ads, exs 5- , 9- , and 21- , contain no explicit comparisons between Aspercreme and aspirin. Instead, they only make statements about the attributes of Aspercreme itself; without contrasting these attributes and those of aspirin. The ads state that Aspercreme does not upset the stomach (CXs 9 and 21-22), or that it has "no side effects" (exs 10-11), or both (eX 6). None of the ads expressly states that aspirin tablets do upset the stomach or that they have unspecified side effects.

It may be that some consumers think of aspirin in association with stomach upset and that, upon hearing in an ad that Aspercreme does not cause stomach upset or other side effects, they would connect the two thoughts and conclude that the ad implies that Aspercreme is superior to aspirin. However, itis equally plausible that consumers would not make this connection. A simple examination ofthe ads does not provide us with suffcient information to determine whether reasonable consumers come away from these ads with the impression complaint counsel suggest. Therefore, this is a situation where we require extrinsic evidence before we can conclude with confidence that the ads imply Aspercreme is superior to aspirin. Having examined the record, we do not find suffcient probative evidence to support complaint counsel's position that the ads make superiority claims. The evidence cited in the (52) initial decision (IDFs 134-37) was the opinion of marketing experts to the effect that various Aspercreme ads state the product to be more effective than aspirin. Having looked at the testimony ourselves, however, we note that it was not in reference to exs 5- , 9- , or 21-22. Rather, it referred to those Aspercreme ads that we already have found to make express sunerioritv claims. Therefore. we conclude that the record is insufI- ), g., l't1U1VH"' uVJ au-"..

648 Opinion cient for us to find that exs 5- , 9- , and 21-22 make an implied representation of greater effcacy for Aspercreme than aspirin. 5. Thompson Had Substantiation for Its Claims That Aspercreme Is Effective Paragraphs 10c, 14. The Commission requires that advertisements containing objective product claims be supported by a reasonable basis. Ifthe advertisements contain express representations regarding a particular level of support that the advertiser has for the product claim (e. tests prove ) or when the ad implies to (53) reasonable consumers that the firm has a certain level of support, the eommission expects the firm to have that level of substantiation. If the ad does not expressly or impliedly refer to a particular level of substantiation, the eommission determines the adequacy of the advertiser s existing substantiation using a number of factors-such as the ease of obtaining substantiation or the cost ofa false claim-identified in Pfizer, Inc. 81 F. e. 23 (1972) and subsequent cases.

The complaint in this matter charges Thompson both with representing that it had a particular level of substantiation for claims that Aspercreme is effective (Paragraph loc) and with making objective product claims (the effcacy claims) that imply the existence of substantiation without representing a particular level for it (Paragraph 14). We conclude that Thompson did make both types of represent ation. In our discussion, we first wil explain why we conclude that several Aspercreme ads represented the existence of a particular level of substantiation. Then we identify, for the remaining ads, the objective product claims implying the existence of a reasonable basis. (54) The ALJ concluded that two Aspercreme ads, exs 7-8, represent that valid studies have scientifically proven that Aspercreme is more effective than orally-ingested aspirin for the relief of minor pain of arthritis and rheumatic conditions (IDF 129). We agree with the ALJ concerning exs 7- , and further find that another print ad, CX 37 also represents that effcacy claims for Aspercreme have been scientifically proven. Moreover, we find the claims in all three ads to have been expressly made.

37 For example, in three other recent analge ic& cases we have required that efficacy daim be supported by a reasonable basi . See Bristol-Myers Co., supra note 8;Sterling Drug, Inc. 102 F. C 395 (1983). afrd No, 83-7700 (9th Cir. August 28, 1984);American Home Products Corp. 98 F. C. 136 (1981),afrd 695 F.2d 681 (3rd Cir, 1982), We discuss the operation of our reasonable basis requirement in more detajJ iD.IVPartof this dedsion. However, we note here that the rationale for it under a deception theory is that objective product claims C!JI'Y with them aD. e"pre s or impEed statement that the adverti er has Home amount of support for the claim- Consumers find these representations of support to be important information in evaluating the reliability of the product claims, Therefore, injury is likely if the advertiser lacks support for the daims J! Pfizeru.sed 3D unfairness atJalysis to reach the conclu jon that the failure to have a reasonable basis violates Section 5(a) ofthe FTC Act. We do not rely upon such an analysis here. Rather, we find Thompson s failure to have a reasonable basis is deceptive by using the analysis first usedNationalin Dynamics Corp. 82 F. C, 488 (1973), affdand remanded on othergrounrh' 492 2d 1333 (2d Cir, art. deni€d419 UB 993 (1974); reisued, 85 F, 391 (1976) Opinion 104 F.

Two of the ads, exs 7- , use similar language. Both state that Aspercreme was "tested" by a leading arthritis specialist on his own patients, with the results indicating that Aspercreme is actually faster and better than aspirin. CX 8 additionally refers to the test as a controlled clinical test." The references to tests by a medical specialist, or " clinical tests " are an express reference to the type of test acceptable to the medical scientific community. Because the ads contain these express claims, we find it reasonable for consumers to expect that the claims that Aspercreme is faster and more effective than aspirin would be substantiated in a manner acceptable to the (55) medical scientific community.39 The third ad, CX 37, does not compare Aspercreme s effectiveness to that of aspirin, but does expressly claim that Aspercreme s active pain reliver "is clinically proven to give strong, effective relief at the point of arthritis pain. Therefore, we operate on a presumption, which may be rebutted by extrinsic evidence (but was not in this proceeding), that reasonable consumers would expect scientifically acceptable evidence to support the claim that Aspercreme provides effective relief. Because the representations that scientific substantiation exists are express, we need not consider whether the product claims for which the substantiation supposedly exists are objectively verifiable ones. If an advertiser states that he has substantiation of a given sort for assertions about his product' s characteristics, we presume reasonable consumers would believe the substantiation capable of having been acquired by the advertiser. (56) Having identified three Aspercreme ads that expressly represent the existence of a particular level of substantiation, we next examine whether any additional advertisements contain implied representations that Thompson had a particular level of substantiation for its Aspercreme effcacy claims. We conclude that the record is insuffcient for us to find that any do.

Most of the remaining Aspercreme advertisements contain no elements whatsoever that might give reasonable consumers a net impression about a particular level of substantiation. However, four of the ads, exs 2-4 and 6, contain visual elements that might create in the minds of reasonable consumers a net impression that Thompson was claiming to have scientific or medical substantiation for Asper- 39 On appeal, Thompson has argued that Commission consideration of the fact that ex" 7 and 8 refer to a scientific test fls the basis for effcacy claims should be influenced by the fact that "these two print advertisement: were dissemillated to sllch a limited extent as to have had virtually no impact il1 the marketplace" and because the claims were "not prominent" in the body copy ofthe advertisements (RRB 18). We reject both there contentions. In the tirgt place, our inspection of the ads shows us that the claims Thompson characterizes as being "not prominent" were no more or less prominent than other claims in the hody ofits 7 and 8. In the secood place we reject as fundamenlaHy erroneous the implicit suggestion that an advertiser may avoid responsibility for express representations by later claiming that the representations were not widely distributed. Such arguments may have some bearing on the extent ofthe relief ordered by the Commission. They have no bearing 00 the issl1e of liability for rlecept.ive acts or practjce g..

1 nV1Vlruvr lVllCUlL,J\L L-V. , U'IL-. O..iJ 648 Opinion creme s effcacy claims. These visual elements are models that some consumers might consider scientific looking (CXs 2 and 4) or line drawings ofthe human body that also might carry with them a scientific aura (CX 6).'0 eonsumers might further draw from such elements, at least where used in drug advertising, the inference that Thompson had performed tests acceptable to the scientific and medical community. On the other hand, it might be that ordinary persons viewing the ads would think no further than that the models and drawings were visual aids to assist the announcer in making points. Because we are unable to interpret these visual elements (57) as implied claims of scientific support and in the absence of extrinsic evidence to assist us in understanding how consumers actually understand these ads, we cannot say whether or not one net impression the ads leave is of the fact that there is scientific support for the assertions that Aspercreme is an effective pain reliever. Given our conclusion that only three Aspercreme ads expressly represent the existence of a particular level of substantiation, and given our further conclusion that the record does not demonstrate any other Aspercreme ads to make such a (58) representation impliedly, we finally must consider whether the effcacy claims in the remaining Aspercreme ads are objective product claims impliedly representing an unspecified level of substantiation, or whether they The issue is not a diffcultare "puffng" representations that do not.'2 one to decide. The record in this proceeding contains ample evidence that effcacy claims for OTC pain relievers can be and are objectively verified (See, e. exs 268-9). Indeed, Thompson did not argue that its effcacy claims were puffery, but rather that it had adequate substantiation for them (RAB 9-23). We therefore conclude that all ef- 40 These sorts of visual elements also are found in CXs 7-8 and 37 , ads which we have already found to represent expressly through other elements contained in them that Thompson had scientific evidence to support its effcacy claims for Aspercrfme.

41 Our approach here is consistent with that we followed in our two other recent analgesics cases Bristol-Myers and Sterlin; Drut;,where we concluded that similar visual image were insufcient to show an implied claim of scientific proof. For example, we stated inBristol-Myers, supra note 8, at 323: rIln CX 61 , 63 and 64., a computer typewriter prints out a column made up of the words "Buferin " and aspirin" on graph paper at the same time as the announcer speaks about scientific tests. . . . Although the computer typewriter enhance the implication (of scientific support created by phrase such as "scientific tests show ).... we do not think that it alone can create lhe impres. ion ofscient.ific support for the claim. Similarly, we do not think that glass models of people with Bufferin and aspirin tablcLG crumbling intomach.'their and reforming in their heads indicates that Bufferin s superior speed has been scientifically established. The difference between our opinions here and in BriBlol- rs, one of emphasis, is due to a refinement in our analysis since we decided that case. There we concluded the visual elements did"ot convey an impression of !!ientific support. Here we merely say that complaint counsel failed to provide extrinsic evidence demonstrating that they created a net impression which did. We do not attempt to use our judgment to reach any substantive conclusion. Where the implied meanings of an advertisement are unclear absent extrinsic evidence, our expertise is no more reliable in pennitting conclusions that an interpretation is unreasonable than that it is reasonable 42 "lTJhere is a category of advertising themes, in the nature of pufng or other hyperbole, which do l8icj.not amount to the type ofaffnn ative product claimsfor which either the Commission or the coosumer would expect documentation. Pfizer, su.prap. , at64. Opinion 104 F.

ficacy claims in the Aspercreme ads are objective product claims. III. WERE THE CLAIMS MATERIAL? After having determined which claims were made by Thompson advertisements, we next turn to examine whether or not those claims were material. material" misrepresentation or practice is one that is likely to affect a consumer s choice of or conduct regarding a product. In other words, it is information that is important to consum- The eommission considers certain categories of information pre-ers. (59)sumptively material. First, the eommission presumes that express claims are material. Similarly, when evidence exists showing that a seller deliberately made an implied claim, the Commission presumes materiality. The underlying rationale in both instances is the assumption that the wilingness of a business to promote its products reflects a belief that the consumers are interested in the advertising. In this case, Thompson itself has acknowledged making claims about Aspercreme s basic effcacy in relieving the pain and other symptoms associated with minor pain of arthritis and rheumatism. Because this acknowledgement clearly demonstrates Thompson s intent, we presume that all basic effcacy claims for Aspercreme, whether express or implied, are material. We also have found that several Aspercreme ads expressly represent Aspercreme to be more effective than aspirin tablets and that (60) another set of ads expressly represents Aspercreme s effcacy to have been proven by the results of scientific tests. Therefore, we presume these claims are material.45 However, the two remaining claims at issue in this proceeding are implied claims that Thompson has not acknowledged deliberately making. These are the claims that Aspercreme contains aspirin and that it is a new product. In assessing the materiality of such implied claims, we are required to make our own evaluation of whether or not reasonable consumers would consider the information in the claims important. One aid to us in doing so for many claims is the fact that over the years our cases have established several categories of claims pertaining to the central characteristics of a product or service, such See our discussion of effcacy claims below ;lt pp 78-5 44 By "basic effcacy claims " we refer to the claims listed in Paragraphs 12a .12c and 12e of the complaint. 45 In considering the matp.riality ofcJaims, we are mindful of the Supreme Court's observation in FTC v Co/gate-Pa/rrwliue Co.380 U,S. 374 , 392 (1965), thalt.he Commission s inference of materialily must be "within the bounds of reasun" Accordingly, we do not use our prlosumption as an inflexible rule that eliminates our need to look at materiality on a case.by-case basis. On thlo contr;lry, the presumption simply r"flects our general judgment that substantive claims in advertisements (in other words, claims other than "puffery" or windowdressing) would not have been made except to affecl a consumer s choice or conduct regarding a product Thus, the very exislence of U,e claim ordinarily is suffcient evidence fur us to conclude it is material. However respondent is always free to counter this evidence either with arguments pertaining to the content of the ad itself Or with extrinsic evide!):". Moreover, the presumption does not preclude us from exercising our own judgment and concluding from evidence in t.he advertisement (or extrinsic evidence) th"t a claim is not. material even if the r,-snnnrlpntrlopsnnl, rJisDlJt,!materialitv 648 Opinion as those relating to its purpose, safety, effcacy, or cost. We now presume that any implied claim in one ofthese categories is material. (61) However, neither the claim that Aspercreme contains aspirin nor the claim that it is a new product appears to fall into any of the above-mentioned categories of claims. Therefore, we must look at the facts on the record to determine whether or not reasonable consumers would consider the claims important.

The ALJ already has analyzed the materiality of the claim that Aspercreme contains aspirin (IDFs 162-64).46 He found that consumer research in the record shows a significant portion of test subjects preferred an aspirin product over a non-aspirin product for pain relief (IDF 163). For example, in the Lieberman Study 53% ofthe arthritics tested expressed a preference for an aspirin product (RX 500F, Warwick, Tr. 5333-34).47 The ALJ also found that Thompson s expert witnesses did not dispute the proposition that aspirin is a drug of choice for treatment of minor arthritic pain (IDF 162). Having reviewed the AI.,J's findings, we agree with them and consider them a suffcient basis for concluding aspirin content claims to be material. (62) The ALJ did not, however, explicitly discuss why claims that Aspercreme is a new product are material.48 Accordingly, we must examine that issue ourselves.

We conclude that these newness claims are material because they imply product effcacy to arthritis sufferers. Rheumatoid arthritis and osteoarthritis are chronic diseases having no cure (O'Brien, Tr. 3946-7). Arthritis diseases cause suffering to their victims and cripple tremendous numbers of persons each year (eX 268, p. 35 455). As a result, arthritis sufferers are constantly looking for cures, with more money being spent on unproven arthritis remedies than on legitimate arthritis research (Roth, Tr. 1537). Testimony by arthritis specialists (Roth, Tr. 1540; Ehrlich, Tr. 4109-11) and medical literature cited in the record (See, e. Brien, Tr. 3775-76) show that in this context, where no existing remedy is fully adequate, arthritis sufferers enthusiasticaJJy try new remedies in the hope that these remedies will provide relief beyond that they are obtaining from existing remedies.

;!i The aspirin contlmt claim is the only one whose materiality the ALJ discussed- Because a finding as t.o materiality is integral to a determination that a representation is deceptive, such findings should be made with respect to any claim upon wnich a respondent is found liable for deceptive advertising. 47 By citing to a specific percentage fig1re (which happens to be over the fifty percent level), we do not ffelln to imply that any particular level of expressed preference must be reached before we would conclude a claim like this one to be material. A lower level thm, that found in thi!\ case clearly could al!\o !\uffce, dependinr; upon the circum!\tance!\ !\urrounding the claim anri the inten ily of the preference expres erl by consumers. 1" We pre ume from I.he general tenor of the ALJ' s discus ion ofthi claim (IDF 126- , ID 1.15-36) thlt he did in fact. believe newne s clajm lo be material. Opinion 104 F.

These facts warrant the conclusion that the Aspercreme ads representing the product as "new " CXs 6- and 10-11, also impliedly represent it as a product effective in providing relief of arthritis pain. Effcacy claims are one category of claims (63) that our previous cases have found to concern central product characteristics.49 Therefore effcacy claims are presumptively material.

Evidence from the ads themselves confirms our conclusion that Thompson was making implied effcacy representations when it represented Aspercreme to be a new product. Four of the ads, exs 6-7 and 10-11, use the words "remarkable breakthrough" in describing the product. The headlines in the four ads state: " At last! A remarkable breakthrough for arthritis pain: Aspercreme" (emphasis added). A "breakthrough" is something new in the sense that it did not exist for more than a short time before. But it is more than that. A breakthrough is something resulting from a significant advance in scientific knowledge. Use of the word "remarkable" to modify the word breakthrough" makes even stronger the implied message that the product is the result of a major scientific advance. The net impression left with reasonable consumers may be not merely that Aspercreme like other products, is an effective arthritis remedy, but that it is a more effective arthritis remedy than the others. However, we find that the newness claims constituted, at a minimum, implied representations of basic effcacy. (64) IV. WERE THE REPRESENTATIONS ONES LIKELY TO MISLEAD CONSUMERS ACTING REASONABLY? 1. Factually Inaccurate Claims, Generally. To this point, our analysis has shown that Thompson made a variety of express and implied claims that were material to reasonable consumers. However, material claims are not deceptive if the messages they convey to reasonable consumers are accurate. Thus, to make a case that advertising is deceptive, the eommission has the burden of showing that the material claims communicated to reasonable consumers by the advertising are false in some manner.50 In other words, deceptive representations must be "likely to mislead. There are two different analytic routes by which complaint counsel can prove advertisments are likely to mislead. One is to carry the '9 See, e.g.. Bris/"l-Myers, supra note 8 Sterling Dru.lf, supraDote 37 B. Williams Co., 68 F. C. 481 (1965). (lfrd 381 F.2rl 8114 (6th Cir. 1967).

5V See. e.g, Bristol-Myers supmnote 8, at320: "If an ad conveys more than one meaning to reasonable consumers and one of those meanings is false, that ad may be condemned, National Commission on Egg Nutrition u. F. T. 570 F.2d 157, 161 nA (7th Cir. 1977),eert. denied 439 U,S. 821 (1978). 51 Not only reprcscnlatio!ls are capable of being "likely to mislead. " Commission precedent alao treats as likely to mislead both practices conveying a material false impression and omissions of materia! information doing the same. This case involves neither su h practices nor such omissions. Therefore, the discussion in the text refers n'"lv tn rpnr,, pnt. ;nn - Hnwp.vp.r. thp. analvsis auulies eouallv to other forms of de eptive conduct ..

.1nV1Vlr vni lVlLUll., I.V. , 11't1.. Vi"" 648 Opinion burden of proving that the express or implied message conveyed by the ad is false. The other is to show that the advertiser lacked a reasonable basis for asserting that the (65) message was true.52 For example, ifan advertisement claims that a new brand of orange juice is more nutritious than others on the market, the eommission could put on its own evidence showing the claim to be false or it could show that the substantiation the advertiser had to support the ad did not provide a reasonable basis for the claim of greater nutritional value. Because the two forms of proof are different, we keep them analytically separate even where a complaint charges, as does this one, both that the advertiser made false claims and that it made claims lacking a reasonable basis.

In this case, Paragraph 11 alleges that three of the claims made by Thompson were "false, misleading, or deceptive." In other words, the complaint signals that complaint counsel must carry the burden of proving the claims to be false. These three claims are: (1) that Aspercreme contains aspirin (Paragraph loa); (2) that it is a recently discovered or developed drug product (Paragraph lob); and (3) that valid studies have scientifically proven it to be more efiective than aspirin tablets for the relief of the symptoms of arthritis and rheumatic conditions (Paragraph loc). We therefore are required to determine whether the evidence put on by the eommission shows the claims to be false. (66) For the claim that Aspercreme contains aspirin, our above discussion has already demonstrated falsity. The active ingredient in "aspirin" is acetylated salicylate, whereas the active ingredient in Aspercreme is TEAlS. The two are not the same. Therefore, the claim that Aspercreme contains aspirin is false.

The second allegedly false claim is that Aspercreme is a recently discovered or developed drug product. The ALJ' s disposition of this issue was not explicitly listed by Thompson as a ruling on which the ALJ erred (RAB 5). However, the company did argue on appeal that most consumers generally had no knowledge of TEAlS products prior to 1979, when Aspercreme national advertising commenced, because Aspercreme itself had only been marketed in Ohio and two other OTC drug products containing TEAlS were: (1) more expensive; and (2) sold mainly through physician recommendation (RAB 3). We infer from this that Thompson believes it was legitimate for exs 6-8 and 1G-11 to claim that Aspercreme was a "new" drug product. The ALJ found that Thompson had made the representations al- 52 Tils method of proof is only avaijable to the Commission for objective product claims. For such claims, the representation "X is true" carries with it the implied representation that "The claim 'X is true' is supported by a reaB.nable basis. " The Commil'lion proves that the advertising is likely to mislead by proving that it isnot supported by a reasonable basis. This docs not preclude the possibility that the claim "X is true" is correct, although the pO!libility typically is an unlikely one. Opinion 104 F.

leged in Paragraph lob (IDFs 126-28, ID 135), but was persuaded that such claims were not likely to mislead because they were made "during the period when Aspercreme was being nationally advertised for the first time" (ID 136). Despite finding that Thompson s newness claims were not deceptive when made, the ALJ nonetheless included in his proposed order a ban on making any (67) further such claims stating, "In any event, Thompson should not make new drug claims for Aspercreme in the future" (ID 136). We disagree with the ALJ' analysis, though not with his result.

As we described above, Thompson s advertising described Aspercreme as a recently discovered or developed drug product, thus representing that the product itself or its drug ingredient was new. In contrast, paragraph 11b of the complaint alleges: Aspercremc is not a recently discovered or developed drug product; it has been available for purchase since at least 1971 and its active ingredient has been in existence since at least 1954.

Thompson s arguments and the ALJ's decision both appear to focus only on the "new product" aspects of the complaint allegations in discussing the propriety of the company s claims. The record shows that Aspercreme was not marketed nationally prior to 1979 and that it had only limited availability between 1971 and 1979. In light ofthis evidence, we cannot conclude that the new product" claims for Aspercreme are false as alleged.

However, we believe that an order provision prohibiting new drug" claims is appropriate on the basis of a different paragraph Paragraph 12. Paragraph 12 states that Thompson made express and implied effcacy representations for Aspercreme. As our discussion in Part III ofthis opinion shows, the newness claims for Aspercreme also were implied effcacy claims. All (68) Thompson s effcacy claims are deceptive for failure to have a reasonable basis. Therefore, we conclude that an order provision barring the newness effcacy claims is warranted.

The third claim the complaint alleges to be false is the claim that Thompson possessed a particular level of substantiation for its effcacy claims, that level being scientific tests. This claim was made in three ofthe Aspercreme ads (exs 7-8 and 37).54 To prove these claims false, complaint counsel had to establish the standards a test must meet to pass muster in the view ofthe medical community as support for the types of claims Thompson was making, and then show that Thompson s tests did not meet these standards. This burden was met. 5J Our discussjon below, at pp- 81--3, explains why we conclude that Thompson lacked a reasonable basis for it. effcacy claims.

THUMPSON MEDICAL CO., INC.

648 Opinion Both our discussion below and the ALJ' s findings (IDFs 214-2) reach the conclusion that the standard generally adhered to by the medical scientific community for testing the effcacy ofa drug is the possession of two well-controlled clinical tests. The ALJ' s careful and methodical examination of every test introduced into the record by Thompson (IDFs 246-333) establishes that the tests fall short of this standard. It follows that Thompson s claims to possess one or more tests acceptable to the scientific community were false. (69) 2. Claims Lacking a Reasonable Basis.

In addition to charging that Thompson made false claims the complaint alleges (in Paragraph 13) that Thompson lacked a reasonable basis for the objective product claims listed in Paragraph 12. These are the various claims that Aspercreme is an effective drug for the relief of arthritis pain.55 Thompson s admissions or our own findings have established that the company did, in fact, make such claims. Now we must determine what level of substantiation Thompson should have had for them and consider if Thompson did possess such substantiation for the advertisements in question.56(70) Starting with Pfizer and National Dynamics,58 our reasonable basis cases have identified several factors that we wi1 weigh in determining the appropriate level of substantiation for objective advertising claims. We recently summarized the factors, as developed by our prior cases, in a policy statement on advertising substantiation. That policy statement is attached to this opinion as an appendix. The factors it summarizes include: (1) the product involved; (2) the type of claim; (3) the benefits of a truthful claim; (4) the ease of developing substantiation for the claim; (5) the consequences of a false claim; and (6) the amount of substantiation experts in the field would agree is reasonable.59(71) 55 Paragraph 12 itself identifies three different Borts of effcacy claims found in Thompson s ads: (1) claims that Aspercreme is an effective drug; (2) claims that Aspercremc is as effective II drug as aspirin tablets; and (3) claims that Aspercreme is a more effective drug than aspirin tablets As a legal matter, it might have been possible for Thompson to possess a nmsonable basis for only one or two of the three claims, hut not for all of them. For example, Thompson might have tests showing that TEAlS is effective, but not tests comparing TEAlS' s effcacy to that of aspirin tablets. In this case, however, none ofthe evidence offered by Thompson was adequate to support anyone of the three types of emcacy claims identified in the complaint. Our discussion therefore docs not distingush among them 0I The advertisements in question are exs 1-6, 9- , and 21-22, We already have found three other Aspercreme ads, CXs 7-8 and 37, to have expressly represented the existence of8Cientific proof for Aspercreme effcacy claims Accordingly, the effcacy claims in those ads would have to be supported by scientific proofregardle!J of whether or not our analysis here concluded iluch substantiation was necessary for the remaining Aspercreme ads. 57 Supra p. 53.

0I Supra note 38 S9 This case is similar to three other recent Commission cases involving GTC analgesics American Home l'product. , Silpra note 37 Sterling Drug, supra note 37, and Bristol-Myers, supra note 8. Those cases speak of estabjishment claims" for GTC analgesics and state that if a claim is an "eslablishment claim " it must be substantiated by two well.controlled clinical teste, We do not use the term "establishment claim " here. However our analysis is consistent with that we employed earlier. " Estahlishment claims" are claims that the effcacy of a drug his been scientifically proved established. " Tn our three recent cases, we stated that we require such (footnote coni'd) ); ), ); Opinion 104 F.

Applying those factors here leads us to conclude that Thompson should possess two well-controlled clinical tests to have a reasonable basis for its Aspercreme effcacy claims. The first criterion we consider is the type of product. In this (72) case, the product is a drug whose application is supposed to improve the physical welfare of its users by reducing pain and the other symptoms of minor arthritis. In past cases involving health or safety issues, we have required a relatively high level of substantiation, typically scientific tests. The second factor we consider is the type of claim. Our past cases have identified at least two types of claims that require a high level of substantiation, such as scientific or engineering tests. One is a claim that refers to specific facts or figures, rather than making generalized descriptions of the product's capabilities..' The other is a claim whose truth or falsity would be diffcult or impossible for consumers to evaluate by themselves"2 This case involves the latter, (73) claim.s to be substantiated by evidence suffcient to satisfy the relevant scientific community of the claim s truth. We further stated that the approprillte level of substantiation for otherdaims would be determined by considering factors such as the harm to consumers if the claim were false. rrhese arc the factors developed Pfzerandin 8uh8eqmmt cases. See, e.g.. Bristol-Myers, supra note 8, at 321. OUf analysis here does not employ the term "",establishment claim" to avoid creating the impression that claims for an advertiser s posse ion of scientific proof wil be treated by us as a unique category of claims. There is nu conceptual or practical rea!)n to single out such claims for special treatment. They are but one example of an express or implied claim that an advertiser possesss a particular level of::ubstantiation. Other such claims might include claims that a particular flower s ability to grow in hot, dry weather had been field-tested (we might require that such claims be substantiated by field tests conducted according to recognized horticultural standards) or thllt surveys show consumers prefer one brand of orange juice to another (we might require thllt such claims be backed by appropriate survey research).

Considered from a rigorously analytical perspective, none ofthes. claims actually falls within the advertising substantiation principles set forth Pfizerin and subsequent cases. Pfizer holds that the Commission itself may identify the appropriate level of substantiation for ads that do not expres.ly or impliedly claim a particular level ofsuhstantiation. It also lays out the factors we will consider in setting the appropriate level ofsubslantiation. We do not have to perform such an evaluation where an advertisement itslf makes express or implied substantiation claims. We treat such claims like any other representations contained in the ad. We verify itthatis reasonable to interpret the ad as making them, that the claims are material, and that they are false. lfso, they are deceptive under Section 5(a) of the FTC Act.

Such an analytic approach is easier for us to employ than if we have to evaluate a case usingPfizerfactors.the However, the end result in either event, assuming we find liability, is an order requiring the advertiser henceforth to have substantiation for the objective product claims being made. From the perspective of the final result therefore, all cases ending upin a substantiation order can be considered ad substantiation cases. This case involves both ads expressly claiming a particular level of substantiation and those to which we must apply the factors outlined Pfizer.in The express substantiation claims are thosein CXs 7-8 and 37, ads representing that scientific tests prove Aspercreme to work faster and more effectively than aspirin table!;. We discussed those claims at pp. 5S-0, above. However, the remaining Aspercreme advertisements did not make express or implied claims to a particular level of substantiation. Therefore, in this section of the opinion we employ the l'fizerfactors to identify the proper level ofsubstantiation for those advertisements 60 See, e.g., Firestone Tire Ru.bber Co. 81 FT.C. 398, 475 (1972),afrd 481 F. 2d 246 (6th Cir. r:rt. denied 414 US. 1112 (1973) (safety and performance claims for automobile tires must be substantiated by "competent scientific tests National Commission on EgR Nutrition 88 F. C. 89, 192 (1976),afrd 570 F.2d 157 (7th Cir. r:rt. denied 439 S. 821, reissued 92 F. C. 848 (1978) ("Parties making claims about the attributes ofproducts. and particularly about the safety of product.'---we to the public a high degree of precision and care. Porter & Dietsch, Inc. 90 F. C. 770, 885 (1977), afrd 605 F. 2d 294 (7their. 1979), cert. denied 445 II.S. 950 (1980) (claims that any food, drug, or device can help a user achieve any result, such as weight 1081, must be substantiated by competent scientific or medical tests or studies 61 See, e.g, National Dynamics C'orp. , supra note 38 (valid laboratory tests would provide reasonable basis for claims concerning specific attribu s of battery additive, such as claims for "quick start in - 40 degrees" or increases brightne81 of lights by 25% "

62 Bristol-Myers. su.pranote 8; Sterling Drug, supra note 37; American Home Products, supra note 37 g, tlUlVt'bUN M.i'J1CAL CU. , lnc. 648 Opinion not the former type of claim.63 As the ALJ' s opinion explains in some detail (IDFs 196-206), arthritis pain is a phenomenon that is not constant. The pain ebbs and flows, making it hard for individual consumers to assess for themselves whether the relief they feel at a given moment is due to a particular treatment they are taking or is a natural phenomenon. This diffculty would be compounded for a new product such as Aspercreme by the "placebo effect " a phenomenon whereby patients' hopes and expectations cause them to believe often for extended periods oftime, that a clinically ineffective medication is providing them with real relief64 (The same placebo effect is also capable of influencing doctors' judgments about drugs they are testing or (74) prescribing ifthey know a drug s identity.) Thus, effcacy claims for Aspercreme are precisely the sort that consumers would not be able to verify easily for themselves. We often consider the third and fourth Pfizerfactors in conjunction with each other. The third factor is the benefit ofa truthful claim. The fourth factor is the ease of developing substantiation for the claim. Our concern in analyzing these factors is to ensure that the level of substantiation we require is not likely to prevent consumers from being told potentially valuable information about product characteristics. In this case, the benefit to consumers from the advertising messages in dispute would be significant if Aspercreme provided both faster relief and relief with fewer side effects than aspirin tablets. However, the record in this case does not suggest that requiring two well-controlled clinical tests as substantiation of effcacy claims for this product (or similar analgesics) would significantly reduce the likelihood of consumers being told about effective remedies for the relief of arthritis pain. The market for such remedies is large, being in excess of 18 milion persons (75) in the United States alone.66 The f, Some Aspercreme ads refer to the product as providing reliefin "seconds." It could be argued that these ads refer to specific figures for the speed with which the product acts. On balance, however, we do not find these references to be the sort of specific figures(e, provides 50 amps starting power at zero degrees Farenheit referred to in the cases holding advertisers who makl! such claims to a higher level of substantiation than otherwise.

64 Two additional factors noted hy the ALJ also would make it djffcult for consumers to evaluate the effcacy of Aspercreme in a nonclinical settng. Many consumers use other medications as well as Aspercreme. This would make it hard for them to separate out which product is the source of the reliefthey feel (IDF 198). Furthermore the method of applying Aspercreme b by rubbing. Rubbing itself is known to have a soothing effect upon musculoskeletal pain (lDF 199). The relief generated by rubbing might be attributed to a nonexistent medical effect of the product being rubbed, masking the product's irwffectiveness. oS The fact that consumers and doctors in uncontrolled environments cannot readily evaluate the effcacy of Aspercreme is a principal reason why we rejl!ct Thompson s claims that effcacy of the product can be suhstantiated by evidence such as the clinical observations of doctors or marketing data (RAB 20). 66 The report of the FDA Panel on OTC Internal Analgesics (CX 268, p. 35 455) provides the following statistics' Incidence of rheumatic disease larthritisJ in the United SUites during 1974 Number of persons Rheumatic disease (millons) Osteoarthritis_ Gout_Rheumatoidarthritis (footnote cont , Opinion 104 F.

potential sales for such a remedy are correspondingly large, as is evidenced by the fact that Thompson s 1981 sales of Aspercreme approached $6 milion (lDF 74). By comparison, the cost of a well-controlled clinical trial of an analgesic s effcacy would be between $10 000 and $15 000 (Adriani, Tr. 1176-77 , Roth, Tr. 1562). This means the total cost of complying with the reasonable basis requirement we establish for Aspercreme would not exceed approximately $30 000. In view ofthe large potential market and likely high demand for analgesics suitable for treating arthritis symptoms, these costs should not deter the development or advertising of new arthritis remedies.

The fourth factor we consider is the consequences of a false claim. In this connection, the ALJ' s opinion stated that use of an ineffective drug (i. one not significantly different from a placebo) to treat arthritis would be both injurious to health (lDFs 207-09) and economically harmful to consumers (IDF 210). Having reviewed the evidence, we agree with the ALJ' s finding that the failure to treat arthritis with effective remedies can (76) cause significant economic harm to the consumer. Those costs result from the repeated purchase of an ineffective product by consumers who are unable to evaluate drug effcacy in an easy manner. However, we differ with the ALJ's findings regarding the health effects of Aspercreme.

The ALJ notes (IDF's 208-09) that failure to diagnose and treat rheumatic diseases with effective medication can seriously harm an individual' s health. Where an OTe product is represented or used as a long-term treatment or cure for arthritis, there is a potential for substantial consumer injury because OTC products do not prevent the progression ofthe two principal forms of arthritis --steoarthritis and rheumatoid arthritis. Where an OTe product is advertised or used for the temporary relief of minor arthritis pain, in contrast, there is little potential for the product to cause serious injury to consumers health if (77) claims about its effectiveness in relieving pain prove false. In this case, we find that the health risk from using Aspercreme as represented to relieve minor arthritis pain is uncertain and should 0.4 to 0. Systemic lupus erythemat.osus Juveni!, rheumatoidarthritis .,7 Our conclusions about potential phy ;ca! harm to consumers rest. in large part upon the characterizations of arthritis provided by the monograph of the FDA's Panel on OTC lntermd Analgesics (CX 268). The panel nutes - 0.that each of the two principal forms of arthritis--Qsteoarthriris and rheumatoid arthritis-has a different cause a diff",rent prognosis, and a different. method of treatment. (CX 268, p. 35453) Active treatment of osttmarthritis requires physical meaSllres and surgical management to retard progression of the disease- As the panel report notes lnJo medication has bt en shown to retard the development or progression of degenerative joint disease (Citation omitted.) I'harmal'ologir agents ldrugsJ playa relatively minor role in the management of osteoarthritis " (CX 268, p. 35456) Aspirin is the mainstay of therapy for rheumatoid arthrit.is, but is must be administered in dosages much higher than those listed on labels orOTC products sold for t.he relief of pain in onier to obtain the anti.inflammalory effects that r",t.ard progression ofthe djm ase- OTC drugs, including Aapercreme, do not prevent the progression of these common forms of arthritis, ami tI,eir contim",d use as self-medication could result in serious health consequences to consumers.

'ltlULVlt'L:l.ll" lfU.:.u-''-LU... 648 Opinion be minimal if the product is used according to the warning on the label that a physician should be consulted if pain persists beyond a short period of time.

The sixth factor we consider is the amount of substantiation experts in the field would consider reasonable. As has been made clear in our past decisions68 and by the ALJ' s initial decision in this matter (IDFs 214-22), the substantiation standard generally applied by the scientific and medical community to claims for the effcacy of an analgesic is that they must be based on the results of at least two well-controlled clinical trials. Evidence from the record of this matter showing that this is the general standard includes the fact that regulations issued by the Food and Drug Administration69 apply such a standard to all OTe drugs and the fact that the standard was applied to analgesics in particular, by the panels of independent experts who evaluated OTe internal analgesics and external analgesics (including TEAlS) for the FDA (CXs 268 and 269). (78) The FDA regulations recognize the possibility of exceptions to this general rule70 In fact, the panels reviewing OTC external and internal analgesics each approved a few OTC drugs without requiring two clinical tests, as Thompson has pointed out (RAB 14-15). However, the existence of these exceptions is not inconsistent with a general rule in the medical community requiring two well-controlled tests to show effcacy, any more than would be the existence of a minority body of opinion holding to some other standard.

In this case, we have not only evidence of the standards generally applied by the medical community to effcacy claims for OTe drug products, but specific evidence that impartial experts do not believe TEAlS' effcacy to have been demonstrated (79) according to appropriate standards. The "active" ingredient of Aspercreme was reviewed by the Panel on OTe External Analgesics in 1979. It held that the effcacy of TEAlS had not been established (eX 269, p. 69 856). After the panel's report was submitted to the FDA , agency personnel independently evaluated the panel's findings. During FDA's evaluation period, interested persons were entitled to submit additional 01 See, eg., Bristol-Myers wpnlf1ote 8, at 338, 376-77. m 2\ C.F.R. 314.111(a), 330. 10(a)(4)(ii) (1983) 71' See, e,g. 21 CF.R. 330. 1O(a)(4)(ii).

. . Proof ofl!fTe tiveness shall cOnSigt of controlled clinical investigations as defined114. 11l(a)(5)(ii)io of the cbapter, unless this requirement is waived on the basis of a showing that it is not reasonahly appjicabJe to the drug or essential to the validity of the investigation and that an alternative method of ;nvel:tigation is adequate to subsbmtiat" dfecl.ivenes . Investigations may b, corroborated by partially controlled or I.meontrolled studies, documl'nted clinical studies by qualified experts, and reports ofsignifieam human experience during marketing. lsolat,d case reports. random exp rience, and reports !"cking the details which permit scientific ev.,l"ation will not be considered. Geoera! reeognitinn ofeffectiveo.cs shall ordinarily oe hascd upon published studies which mOlY be corroborated by unpublished studies and other data It We do 110t believe it n()cessary in deciding this case to attempt to id()notify all the situations when exceptions hDm th", general rule wiIJ be permitted The positions of l\W Panel on OTC ExtertJaJ Analgesics and of the FDA On Tt:A/S demonstrate that 1m exception is not appropriate for this particuJllr ctwrnical substance Opinion 104 F.

evidence supporting the effcacy of ingredients that the panel had found not to be proven effective. Accordingly, Thompson submitted to the FDA the studies it asserts in this proceeding to be well-controlled clinical tests. The results of FDA's deliberative process and its evaluation of Thompson s evidence are found in the "tentative final monograph" on OTC external analgesics, which was published for public comment in 1983. The tentative final monograph reiterates the panel' s conclusion that TEAlS has not been proven effective and dismisses each of Thompson s proffered tlc1inical tests" as inadequate (CX 443, p. 5 855).72 Although the tentative final monograph is subject to public comment and possible revision before the FDA publishes a final monograph, the tentative final monograph reflects the agency considered judgment and current position on the merits. Therefore, based upon our review ofthe six Pfizer factors, we conclude that the proper level of substantiation for Aspercreme effcacy claims is two well-controlled clinical (80J tests. We are additionally persuaded to use this level of substantiation because our above discussion indicates that this is the standard currently being required of TEAlS by the Food and Drug Administration. We believe that advertisers of drug products subject to the joint jurisdiction of the FTe and the FDA wil benefit from greater regulatory certainty if they can act with reasonable assurance that the two agencies will accept the same evidence to demonstrate the safety and effcacy of a particular ingredient. Thus, we state that advertisers who comply with the FDA' requirement of well-controlled clinical tests to demonstrate effcacy have adequate substantiation to make such claims in their advertisements. Although we do not preclude ourselves from also permitting advertisers to use other types of evidence to comply with our substantiation requirement, nothing in this record suggests any rationale for our permitting a different form of substantiation for effcacy claims in Aspercreme advertisements than the FDA is prepared to establish for the product's labeling.73 On the contrary, the inability of consumers to evaluate analgesic effect by themselves in an uncontrolled environment is a persuasive reason for consumers to expect (and us to require) appropriate scientific testing before effcacy claims are made. (81) In reaching our conclusion that Thompson lacked a reasonable basis for its Aspercreme effcacy claims, we reject Thompson s arguments to the contrary. The first argument, relying on the testimony ,. The FDA's position contradicts Thompson s assertion (RAB 17) that" . three double-blind clioical studies supported effcacy (ufAspcrcrcmcj."

7) Consist,mt with U,is view, O\lf ordn would permit Thomp on to advertise Aspercreme as"n effective analgesic . .. ifthe Fuud and Drug Administration promulgates any final standard that establishes conditions under which such product is safe and cfTccUvc under the Food, Drug, and Cosmetic Act " whether or not the standard requires two well-controlled clj!Jicals (Wp. have no re,, on to think FDA would di8pcn e with the requirement of two wdl-cuntrol1ed clinicals.) , p !HJn.L.L 648 Opinion at trial of Thompson s expert witnesses, is that the scientific and medical community does not consider just one type of evidence (i. e., well-controlled clinical tests) suffcient to demonstrate drug effcacy (RAB 13). While the argument would be true enough ifit Were merely an assertion ofthe point hitherto acknowledged by us, that the medical community may on occasion permit an exception to the general rule of well-controlled clinical tests, this is not what Thompson would have us find. Rather, the company wishes us to accept the proposition that the medical community does not, as a general matter, require well-controlled clinical tests to support claims of drug el!cacy. We find this assertion contrary to the preponderance of the evidence in the record of this case and, accordingly, reject it. Thompson next asserts that not even the FDA requires the existence of two adequate and well-controlled studies as the basis for an OTe drug effcacy claim (RAB 14 16), relying principally on the fact that FDA panels reviewing various OTC drugs have not invariably required such evidence. As we stated earlier, however, the existence of exceptions does not, in and of itself; disprove the existence of the general rule. For example, the three external analgesics approved by the FDA Panel on External Analgesics without two well-controlled clinical tests came from a group of over forty drugs reviewed by that panel (eX (82) 269 , p. 69 790). Moreover, complaint counsel have presented us with persuasive explanations of why these particular exceptions were made.

Third, Thompson argues that it did have three double-blind clinical studies on Aspercreme, studies that met the applicable standards of the FDA and the medical community (RAB 17-18). We find it remarkable that Thompson would make this assertion in light ofthe FDA' present position that Thompson s studies are (83) inadequate (eX 443 p. 5 855). Moreover, we are independently persuaded by the evidence on our own record (discussed in IDFs 246-312) that each of Thomp- Complaint counsel' s answering brief slates (CAB J4): Two of the drugs Thompson points to are couoler-irritants, which are a specific class of drugs that. exert their analgesic action in a unique way, by producing t.he seosation of warmth or coolness on tbe skin. Accordingly, since these drugs' mechanism of action is understood, clinical tria)s of various counterirritant, would be adequate to document the effcacy of counterirritants as a class. (Roth, Tr. 1763-1) Moreover, in the case of all three of these external analgesics, the External Analgesic Pane! spccificiiUy states that there are reports about each in the published literature and cites to authoritative compendia Oll drugs In contrast TEAlS an obscure drug, and is not listed in any of these compendia. (Y. 345). (IDF 345) AlsoexpJanatory of the exceptions are the comments orthe Panel on OTC Extp.rnaJ Analgesics itself. For example in djscus iI1g turpentine oil, one of thetwo counterirritants it found effective, the pane! stated' Due to the i!1gredient's wide use and clinical acceptance and on the basis of published report. in the literature, the Pan(!l concludes that turpentine oil is effective far use as an OTC external analgesic. No scientifically controlled studies;; concerning the use ofturpentino oil alone for the treatment ofrheuma. tism. arthritis, and muscular aches and pains were found. Howover, the use of turpentine oil for self-medication is almost an American folk traiiition, and fuji-strength turpentine oil has been employed with impunity as a topical counterirritant. (CX 269 R40) Opinion 101 F.

son s tests had one or more serious flaws that prevents it from complying with the standards for a web-controlled clinical test. Finally, Thompson argues that its possession of a reasonable basis for its Aspercreme ads is underscored by the quality ofthe individuals who came forward to testify on its behalf(RAB 20-21) Unfortunately for Thompson, however, none of the factors established by Pfizer our other reasonable basis cases suggests that the credentials of a respondent' s witnesses should be a substitute for factual evidence as a basis for objective product claims. Moreover, the testimony of experts, no matter how well-qualified they may be, is no substitute for controlled clinical testing as a means of substantiating the drug efficacy claims at issue in this case.

In addition to the above arguments, Thompson asserts that it is inappropriate for the Commission to take action against Aspercreme before the FDA reaches a final decision on whether or not TEAlS has been proven effective (RAB 28-31).7 In other (84) words, Thompson suggests that it is inappropriate for the eommission to reach a final decision on whether or not TEAlS has been proven effective when the issue is stiJ an open question at the Food and Drug Administration. Thompson further implies that evidence it submitted to the FDA too late for that agency to review prior to publication of the tentative final monograph on OTe external analgesics wil persuade the FDA to conclude that TEAlS has been proven effective when it publishes the final monograph on OTC external analgesics. It is true that the FDA's proceeding is stiJ open and that the agency could reverse its tentative decision on TEAlS. However, we have no reason to believe this will happen. As the ALJ' s initial decision points out (IDFs 387-99), the new material submitted by Thompson to the FDA does not appear to contain two (or even one) wen-controlled clinical tests. Therefore, ifthe FDA continues to apply the standards it heretofore has applied to TEAlS it should not find the new materials any more persuasive than the old.

In any event, our decision to issue an order in this proceeding does not rely upon a guess as to how the Food and Drug Administration ultimately will come out on the question of TEAlS' s efficacy. Rather it is based upon Thompson s failure to provide us with evidence that we think provided a reasonable basis for Aspercreme s effcacy claims. Moreover, the order we issue contains language allowing Thompson to rely for substantiation of Aspercreme effcacy claims ); Thig is a favof;lbk reading ofTbompson s argument. As written, Thompson s briefmereJy asserts the ALJ to have "erred in holding that. the Ext.ernal AnalgC!\ics Panel and the FDA have foundTEAlS to be ineffective OIS a topical analgesic ingredient." However. (wthing the ALJ stated in the initial decision suggests thflt he believed the FDA to have made a final determination. He correctly characterized the position expre98ed in the f"DA' tentative final monograph for external analgegics (CX 443) as the FDA' s "current position (IDY 395), He stated his further belief Lhilt iL is "highly unlikely " the FDA will reverse its position (IDF 400). He never expressly or impliedly stated that the FDA had finally det.ermined TEAlS to be inefI"active 'lH nn.1 u\.n. lnd.1.1'-.U-' '-'-. , J-',"'. 648 Opinion upon any evidence conforming to a final standard for effcacy promulgated by the (85) FDA. Thus, ifthe FDA changes its position the result amounts to an automatic modification of our order. Indeed, we see very good reason for us to take action on Aspercreme despite the pendency of the FDA's OTe external analgesics review. FDA' s proceeding is a rulemaking that must focus simultaneously on many different drugs, one of which is TEAlS. As Thompson itself has noted (RAB 30), while the date set for close of comments on the FDA' external analgesics tentative final monograph was April 9, 1984, it is uncertain how much additional time FDA wil need before resolving all of the issues presented to it by the rulemaking. In contrast, the proceeding before us is an adjudicative one focused specifically on Aspercreme. Perhaps for this reason, we are in a position to reach our final determination before the FDA is able to reach its final determination. Given our conclusion that the Aspercreme advertising in evidence on our record is deceptive for failure to have a reasonable basis (among other reasons), it is in the public interest for us to act expeditiously to prevent further harm from the continuation of such advertising. (86) v. SCOPE OF RELIEF This part of the opinion discusses the order we enter against Thompson to prohibit and prevent it from engaging in deceptive acts or practices. Our order differs in several respects from the one proposed by the ALJ. Accordingly, we first discuss the rationale for each of our changes. We then discuss why we believe it appropriate for the order to apply not only to Aspercreme, but also to other drug products marketed by Thompson. Finally, we explain why we decided not to adopt an order provision that was urged upon us by complaint counsel, one that would have required Thompson to excise the "Aspercreme" brand name.

Our first change removes from the first paragraph of Parts I and II the word "labeling." The eflect of this deletion is to limit application of the order to the "advertising, offering for sale, sale or distribution" of drug products. Our revision is intended to ensure that the order cannot be interpreted as applicable to any information the Food and Drug Administration either permits or requires Thompson to place on the labels of its drug products. For example, if the FDA permits Thompson to characterize Aspercreme as an analgesic during the pendency ofthe FDA's OTC drug review process, our order would not bar such labeling. On the other hand, any advertising placed on 76 We have includt d thi language in our urder hecau e of our aforementioned view t.hat it is advisable for us and the FDA to t;!ke a unified regulatory appro"ch to issue brought before us where tbe is.'\ues appe"r jdentical or quite sjmilar Opinion 101 F.

the label (i. language not covered by FDA regulations) would be covered (87) by the order. The order as revised by us is identical with the orders in our three other recent analgesics cases American Home Products, Bristol-Myers and Sterling Drug in this respect. The ALJ apparently added the word "labeling" to Parts I and II to address a concern of complaint counsel. They felt the disclosure mandated by Part LA (stating Aspercreme does not contain aspirin) had to be made both in Thompson s advertising and on the product label to be effective and that, therefore, the order had to apply to labeling (ePF 487). We agree with complaint counsel's belief that the disclo sure needs to be on the product label to be eflective. However, Part LA of our revised order stiU requires that Aspercreme labels contain the desired disclosure if Thompson wishes to continue using the Aspercreme" brand name in advertising or sales or otherwise to represent in ads that the product contains aspirin. Thus, the net effect of our revision is to make continued advertising and sale of a product named "Aspercreme" conditional upon a disclosure on its label that it does not contain aspirin, but not to make the order applicable to any labeling regulated by the FDA.

Our second change in the ALJ' s proposed order broadens coverage of its Part I from "over-the-counter analgesic drugs" to "over-thecounter drugs" generally. We have broadened Part I's coverage because we agree with complaint counsels' argument on appeal that the facts of this case warrant such " fencing- " to prevent deceptive acts or practices by Thompson in the future. We discuss in more detail below our reasons for reaching this conclusion. (88) Our third change modifies the language of Part LA by replacing the last H " in it with the words provided, however and by making other conforming changes. We made this technical revision because the langugage approved by the ALJ could literally be read as permitting Thompson to represent expressly (or impliedly) that Aspercreme contains aspirin as long as Thompson s advertising and labeling contained a disclosure that it does not. Our revision makes clear that Thompson may engage in no such conduct (other than use of the brand name "Aspercreme ) regardless of whether or not it uses the disclosures required by the order.

Our fourth change is a revision to Part LA.1 of the order. That provision sets out the requirements for disclosures that Aspercreme does not contain aspirin when the disclosures are used in television advertisements. As approved by the ALJ, it not only specified that the advertisements include both a clear and prominent vocal statement at the end of each ad informing consumers that Aspercreme does not contain aspirin and a clear and prominent video super advising consumers of the same fact, but also specified that the super must be g., TtiUIVIY;:Ul LVIJ1UILi-H.J LV.

648 Opinion displayed throughout the length of each advertisement. The requirement that the super be perpetually displayed throughout an entire ad strikes us as overkil. We have not imposed a similar requirement in any of our past affrmative disclosure cases and do not believe it is appropriate to do so here. Instead, we have revised Part LA. 1 to require that the visual super be displayed at the end ofthe ad simultaneously with the vocal disclaimer. (89) Our fifth change from the AI.' s proposed order revises the language of Part LB., the provision prohibiting false newness claims. We have made several changes in this provision. First, we have deleted the reference to new "mechanical" principles. Nothing in this case suggests Thompson has claimed (or plausibly could claim) that the effcacy of an OTe drug is based upon new mechanical principles. Second, we have deleted the prohibition on generally representing that a product is new (as opposed to representing that it involves new scientific principles) because the materiality of newness claims in general has not been established by the evidence before us. Third, we have deleted the language excepting from operation of the provision claims of newness made during a product's introductory period. We see no reason why Thompson, or any other advertiser, should be entitled to use the excuse of a "new " product (e. an aspirin tablet marketed under a new brand name) to represent directly or indirectly for any period oftime that the product is the result ofa new scientific principle (e. a new active ingredient) when it is not. Finally, to clarify the operation of Part LB we have replaced the phrase "generally available for purchase" with the phrase "available for purchase as an over-the-counter drug. " (90) Our sixth change replaces a part of the order deleted by the AI., one that prohibits Thompson from misrepresenting the active ingredient(s) in any OTC drug products. The AI.' s initial decision does not explain why he omitted this provision, which had been included in the notice order accompanying the complaint. Whatever his rationale, we agree with complaint counsels' argument on appeal that the deletion was incorrect. Part LD. of our order, the provision deleted by the AI., is based on one of the basic deceptive practices this case has shown Thompson wiling to engage in-the misrepresentation of active ingredient. It prohibits such misrepresentations for any OTe drug products. Although Part LA of the order prohibits the specific misrepresentation involved in this proceeding, misrepresentation of aspirin content, we believe Part LD is necessary as fencing-in for the same reasons that prompt us to broaden the coverage of Part I from OTe analgesic drugs to OTe drugs generally. We discuss these reasons below.

Our seventh change is to Part ILA of the order. That part prohibits Opinion 104 F.T.

Thompson from representing any OTe analgesic drug tb be effective unlessfor the relief of symptoms of musculoskeletal disorders, Thompson has a reasonable basis for such claims two well-controlled clinical tests. Our revision removes a "delayed" effective date (April 9, 1984) that no longer would be in the future, given the passage oftime since the AW issued his initial decision. An alternative would have been to extend the delay until such time as the FDA makes its final determination about TEAlS' effcacy. Unlike the AW, howev- , we question the (91) wisdom of such a course. Thompson already has had ample time during the pendency ofthis litigation to conduct the sort of tests that could establish effcacy to our satisfaction. If it has not done so, either out of conviction that it is not required to or out offear that such tests would show Aspercreme to be no better than a placebo " it wi1 have to stop making effcacy claims in its advertising and offering for sale of Aspercreme until those tests are performed. While the inability to advertise that Aspercreme is effective wiu make it diffcult (although not absolutely impossible) to market Aspercreme for a period of time, Thompson can always resume a full-scale selling campaign if and when it obtains the evidence of eftcacy that it should have had all along.

Our eighth and final change is to add back to the order Parts IV and , standard provisions for aU our orders. Part IV requires Thompson to notify us prior to any proposed change in the corporation. Part V requires the company to fie a compliance report within 60 days after service of the order upon it. These provisions appear to have been inadvertently omitted from the ALJ's proposed order. (92) The order we adopt contains fencing-in provisions it would place restrictions on Thompson s ability to market products other than Aspercreme. Part II of the order requires a reasonable basis consisting of two well-controlled clinical tests if effcacy claims are made not just for Aspercreme, but for any OTe analgesic drug. Part I applies not only to analgesics, but also to any other OTC drug products Thompson might sell. We believe such fencing-in provisions are warranted in this case. The ability ofthe eommission to issue orders containing fencingrequirements is clear. See, e.g., FTC v. Ruberoid Co. 343 U.S. 470, 473 (1952); FTC v. Colgate-Palmolive Co. 380 U.S. 374, 394-95 (1965). The eommission has wide latitude in fashioning orders to prevent inventive respondents from pursuing a course of conduct similar to that 1" Such a rear rnigf,t well be warranted. A the ALJ' s opinion note (IDFs 376-9) five studips in evidence in this proceeding find no statistically significant differ"nce between TEAlS and a placebo- The ALJ concluded that thp.se studies, like Thompson s. wpre not well-ontrolled (lDF 374). Howpver, the relative quality ofthe",se studies versus those rdi( d on by Thumpson is suggested by the filet that the FDA' s Tentative Final Monograph Or! OTC External Analgesic Products cit"d to the Hndings of four of these studies as evidence that TEAlS has not been proven effective, am! did so without critical comment, while criticizing the methodology of Thumpson s three supposed clinical test.s (CX 44:J, p. 5 855) THOMPSON MEDICAL CO., INC. b;j;j 648 Opinion found to have been unfair or deceptive in the past. However, the Commission s discretion is subject to two constraints. One is that the order must be suffciently clear and precise to be understood by the violator. See, e. , Colgate.Palmolive, 380 U.S. at 392. The second is that the order must bear a reasonable relationship to the unlawful practice found to exist. See, e.g., Jacob Siegal Co. v. FTC, 327 U.S. 608 612-13 (1946).

To ensure that a multi-product fencing-in order such as this one bears a reasonable relationship to the unlawful practice found to exist, the eommission considers three factors. They are: (1) the deliberateness and seriousness of the present violation; (2) the respondent' s past history of violations; and (3) the transferabilty of the unlawful practices to other (93) products. The more egregious the facts with respect to a particular element, the less important it is that another negative factor be present. See, e.g., American Home Products Corp. o. FTC, 695 F.2d 681 , 706 (3d eir. 1982); Sears Roebuck Co. v. FTC, 676 F. 2d 385, 392 (9th eir. 1982). In considering these three factors, we find that two ofthem, the deliberateness and seriousness of the present violation and the transferability of the unlawful practices to other products, warrant adoption of the order we enter today.

We look first at Thompson s present violations of the law. We have concluded that they are deliberate and serious. The seriousness of Thompson s violations is evidenced by the size and duration of Thompson s deceptive advertising campaign. As our above discussion shows, Thompson has been making deceptive effcacy and aspirin content claims since it began advertising Aspercreme in 1977. These claims have been made in numerous different ads on TV and in print in a campaign backed up by a multi-milion dollar advertising budget (IDF 74). Such a persistent, long-term pattern of deceptive advertising evinces a (94) massive long-standing effort"79 to persuade consumers that Aspercreme contains aspirin and that there is a reasonable basis for claiming it is effective.

The ads claiming that Aspercreme is a new breakthrough and that its effcacy is supported by scientific tests were run only in print ads rather than in the TV commercials that commanded the bulk of Aspercreme s budget (RX 573). Even so, these ads were not insignificant. For example, the CX 6 print ad ran twice in the Reader s Digest and once in the Saturday Evening Post in 1979 (eX 25A). Print ads 70 Complaint ounsel 3T1iued on appeal that Thompson s history of past violation also provides grounds for a fencing-in order. However, the history cited by complaint counsel consists of a single consent order against Thompson js ued in the early 19605 Thompson Medical Company. lnr, 59 F, C. 287 (1961HCAP 17-18). RecauR consent orders do not constitute a legal admission of wrongdoing, we wiJJ not use a single consent order as a basis for concluding that Thompsun has a history of part violations. We express no opinion on whether or not a pattern of con ent orders would be a sufficient hi tory of p st violations to warrant fencing- See American Home Products Corp. v. FTC, supra note 37, 695 F. 2d at 707 g.. ), g., Opinion 104 F.

in such magazines of nationwide distribution would result in the objectionable representations being seen by large numbers of persons. The seriousness of the violations also is affected by the fact that consumers could not readily judge the truth or falsity of the claims Thompson was making.80 As our discussion above has noted, consumers cannot readily determine on their own whether or not an analgesic is effective. It would be even harder for them to determine whether or not a product advertised as a new "breakthrough" actually is one or to figure out whether supposed scientific tests showing that a product is faster and safer than aspirin would pass muster according to the general standards of the scientific community. Therefore, the claims Thompson emphasized in its advertising were ones to which consumers were particularly susceptible. (95) Just as troubling as the seriousness of Thompson s violations is their apparent deliberateness. Deliberateness is shown by the consistency of Thompson s advertising themes over the years, supporting a conclusion that they were no accident or "isolated instance. See, e. Jay Norris Inc. u. FTC, 598 F.2d 1244, 1250 (2d eir. cert. denied, 444 U.s. 980 (1979).

Thompson has attempted to rebut the notion that it engaged in a deliberate campaign of deception by claiming that it did everything possible to ensure it had a reasonable basis for its Aspercreme effcacy claims. 81 We do not find it so easy to exonerate Thompson s conduct. Thompson excuses itself by asserting that it relied on the opinions of scientists it hired (96) as consultants, who advised it that Aspercreme was effective. Thompson seeks to avoid the charge that it should have realized those scientists were not relying on well-controlled clinical tests by arguing that it does not conduct in-house clinical investigations and lacks the expertise to evaluate such studies itself (RAB 9). However, the facts on the record in this proceeding show a pattern of outside sources repeatedly warning Thompson that the effcacy of TEAlS had not been established. The FDA's Panel on OTC External Analgesics expressed this view in 1979. In early 1980, the National olt BJ See. e. Respondent's Answering Brief, at 7: The record shows that the slate of mind lJfThornpson, from tht' tim",it first purchased the product Aspp.rcreme (and even before), was directed towards;Itjsfying iL'Ielf, and accumulating ample cientific proof, that Aspercrernp. was both safe and effective for the temporary reli.,fofminor musculoskeletal pain, including that minor pain associated with arthritis and rheumatism. Towards this goal, Thomp o!1 accumuJat . at great expense, extensive medical and Mcicntiflc opinions, reports and clinical documentation from numerous out. standing, well recognized authorities and experts. Thompson also argues that its money-back 6'1arantee evidences jt good faith reliance on the evidence that Aspercreme is effective.ld. OIt 10. However, a muney-back g-uarantee is not a defense to a charge of deceptive advertising. See, e.g.. Montgom('ry Ward & Co. u. ftc :179 F.2d 666, 672 (7tn Cir. 1967). In this case, the placebo dated would result in sume consumers purchasing the product many tim.,s before discontinuing its use and/or asking for a refund un the last butt!. bought. Therefure, 01 rnolley-back b'1OIrantee would nut eliminate sulmtantial sales revenues "ven io the unlikely event that all consumers eventually invok"d it g., g., THOMPSON MEDICAL CO., INC. lI;jb 648 Opinion Association of Broadcasters and each of the television networks advised Thompson that the documentation for Aspercreme s effcacy claims was inadequate and that the tests Thompson relied on were flawed. (See, e. CXs 88, 89, 90-91). Also in and around 1980, Thompson became aware of four studies-Roth (CX 344Z-195), Ehrlich (eX 344Z-157), charles (ex 344Z-168), and Brown (eX 344Z-182)-showing no statistical difference between TEAlS and placebos. In 1981, the director ofthe FDA's Division ofOTe Drug Evaluation told Thompson that TEAlS had not been shown effective. His comments are particularly significant because he reviewed the clinical studies Thompson had submitted to the FDA after publication of the OTe external analgesics panel's report, the same studies Thompson has submitted to us. He wrote to Thompson that these studies were all deficient to prove Aspercreme s effcacy (eX 342; see also ex 343). Again in 1983 with the adoption of its tentative final monograph (eX 443), the FDA concluded that TEAlS has not been (97) proven efiective. Through all of this, Thompson steadfastly argues that the scientific experts it employed (e. the persons who performed the defective studies) believed TEAlS to be effective, that its studies are valid, that its other evidence (such as a 1981 non-projectable survey of pharmacists by a magazine, a survey critiqued by the ALJ in IDFs 369-71) is adequate to support effcacy claims, and that it is acting in good faith when it continues to rely on this evidence. Our reading of the record is that Thompson has known or should have known for some time now that its effcacy claims for Aspercreme are unsubstantiated. We further conclude that Thompson has deliberately continued making effcacy claims despite this fact.

Likewise, it seems clear that Thompson deliberately sought to lead consumers into the belief that Aspercreme contains aspirin. Again Thompson had good reason to know that one reasonable interpretation of the product's labeling and advertising was that it contains aspirin. The company was warned of this possibility by the results of two focus group studies that it had available to it as early as 1978-the Nicholas and (98) Schneider focus groups.8" consumer belief in aspi- R2 The Nicholas foeUi; gTOUp study (CX . 2) was conducted in 1973 with two groups ofelcven women, all of whom suffered from arthritis or some form of muscular aches and pains and all of whom were users ofcither topical analgesic or aspirin- The Schneider focus group study (CX 53) was conducted in 1978 with three groups of respondent. (two groups of females and one group of males). aJl of whom used some sort of internal tablets or external rubs fur arthritis relief. Participants in both the Nicholas and Sdmcider focus group studies were recruited and asked to use Aspercrernc for at least ten days prior to the focus group ses:ioo Because focus group studies are conducted with very few r€spondent.'\ obtained through nonprobahilily ;mmples and because the interviews are conducted in an unstructured group format, it is diffcult to draw generalizable conclusions from them. Indeed, it is not unusual to obtain conflcting results from fucus Seegroups. discussion at note 34 above. (One explanation for such conflcting results is that the moderator s control over discussions can skew them toward a particular result. It mig1lt he, therefore thlt 90% of the participiint.'\ in a given focus group (9 out of 10 people) thought of "aspirin " !IS a particular chemical, and in another only 10% (lout of 10), while a survey would show 60% of the general population to hold that belief: Accordingly, we do not. expect Thompson to have treated these focus groups by themselves as a fully reliable indication of consumers' beliefs about aspirin (footnotecont , Opinion 104 F.

rin content also clearly was indicated by the results of the Mapes & oss copy test in 1979 (eX 50), .which examined the earliest Aspercreme (99) TV ads (exs 1 and 2).83 The possibility of such a net impression was further communicated to Thompson by the television networks and NAB in 1979 and 1980 (See, e. CXs 79, 80, 88DJ. The fact that the ads created a net impression of aspirin content appears to have been no accident. The document "Aspercreme Brand Review " prepared for Thompson by Ogilvy & Mather in July, 1980 (eX 54), shows the purpose of Aspercreme advertising was to communicate that Aspercreme is aspirin in a rub. Moreover, when the networks insisted on a revision of the early TV ads to reduce the incidence of mistaken beliefs, and Thompson accordingly approved ads with supers stating contains no aspirin " and relief without aspirin" (exs 3 and 4), it appears that Thompson continued to use the audio phrase "relief of aspirin" in the revised ads hoping that the phrase would (100) offset the disclaimer.85 Thompson also made no attempt to test whether or not the disclosures in exs 3 and 4 were effective.86 In addition, Thompson apparently continued to use ex 9 and related TV commercials (the ones with the disclosure "contains salycin, a strong non-aspirin pain reliever ) after the completion in early 1981 of consumer research that it had sponsored and from which it should have realized that the ads communicated a message that Aspercreme contains aspirin.87 All of this conduct is evidence that content in Aspercreme. However, the focus groups did provide Thompson with exploratory information about tho.'e beliefs. The result ofhoth the Nicholas and the Schneider focus group studies were consistent in demonstrating that there wa a high prob"bility that pilt-icipants, would believe Aspercrcme contains aspirin- Indeed, the final reports of both studies contain quotes that emphasize an a pirin 8S8ociation For example, the Nichol"s report states that one participant said When I sow it and saw 'Asper, ' I right away thought it had aspirin in it "(CX 52-I.) and the Schneider study states "in addition, others felt they were attracted by 'Asper! Aspercreme' because it has aspirin in it' or is ' full of aspirin.' " (CX 53-Z56) Thus, both the Nicholas and Schneider focus group studies indicate that even after using Aspercreme for at least ten days several of the focus group participants believed that Aspercreme contained aspirin- The con istency ofthi finding across both udies should have beel1 a warning signal to Thompson that potential consumers might be confused about the ingredient of Aspercreme 1\1 As discussed above, at pp 20 - , results from the Mapes and Ross Test indicated that between 21 and 35 percent of persons viewing exs 1 and 2 believed Aspercreme contained aspirin i4 See, e. CX .54Z--D05.

Creative Slrater:Y The creative objective is to convince arthritis sufferers - men and wOlTen over 50 - that Aspercrerne is the arthritis medicine that puts all the relief of aspirin directly at the point of pain The reason why is that Aspercreme contains the pain relieving ingredient of aspirin in a penetrating carrier so that it is taken quickly through the skin into painful joints and muscles 8, A 19.80 conference report summarizing a meeting between Thompson and Ogilvy & Mather (CX 66) to discuss rurming various new ads states: "The client lThompsonj agreedlto TV ads with supersJ and will pursue approval of the aspirin equivalency claim. In the meantime the agency iOgilvy & MatherJ will puruse the ' strong relief of aspirin' claim to of1set the contains no aspirin super HI Thompson has suggested to us that "the networks accepted these supers because they were suffcient to cJ"rify any possihlc amhiguity" (RAB 47). This is not the case. At least one network warned that the supers were ineffective. SeeCX-.80 (letter from Director ofCommerical Clearance for cns to Thompson): " In the ncw tape, the super (relief without aspirin) accomplishes little and may only serve to confuse the issue. For that reason, we have decided nul to accept the revision hut to remain with the original for a period of two months, until January 15 1980. By that time, we wil hope for some hctter-ddined message which can avoid the present diffculty 648 Opinion Thompson has known full well for some time that consumers misunderstood the identity of the principal ingredient in Aspercreme and has continued to advertise in a manner that creates more such misunderstanding.

In addition, the "adaptability or transferability" of Thompson violations to other products convinces us that a fencing-in order is appropriate. The scope ofthe order in this (101) case is similar to those and Sterling Drug. Thosein American Home Products, Bristol-Myers were cases that also involved claims for analgesic products. Our comments in American Home Products 98 F. e. at 405, are equally apposite here: "The effort to misrepresent the nature of . . . (anJ ingredient is a technique that could easily be applied to advertising of OTe drug products other than (this one)." Likewise, claims that tests prove a product' s superiority or claims that its active ingredient is a breakthrough could readily be employed for any non-prescription drug product. Accordingly, it is necessary that Part I of this order apply to alj such OTe drug products.

Part II of the order is narrower in scope than Part I, being limited to OTe analgesic drugs. It is so limited because our factual findings go only so far as to conclude that two well-controlled clinical tests are necessary as a reasonable basis for analgesic effcacy claims. Conversely, however, this analysis makes clear that no lesser standard than two well-controlled clinical tests is appropriate as a general rule TEAlSfor any analgesic product, whether its active ingredient is something else. It also is clear there is nothing inherently unique about Aspercreme or TEAlS to prevent Thompson from transferring the practice of claiming effcacy without such proof to other OTe analgesics. Thus, a fencing-in provision is warranted. Finally, we comment on one order provision that complaint counsel urged upon the ALJ as well as upon us and that both he and we have rejected. It is an order provision that would (102) require Thompson to excise the "Aspercreme" brand name. complaint counsel have asserted that brand name excision is a remedy the eommission has employed in the past when a brand name was deceptive and when no less restrictive alternative would suffce to eliminate deception (eap 23-26). complaint counsel further argue that such is the case here. We agree with complaint counsel that brand name excision is a remedy available to us for use in extreme circumstances.88 We do not find, however, that complaint counsel have made a suffcient case to warrant employing the remedy here. To order brand name excision we would have to be persuaded that a less severe remedy, such as 08 See, e. , FTC v. Algoma Lumber Co. 291 U,S. f) (934): Resorl Car Syglems, Inc. (I, FTC 518 F.2d 962 (9th Cir.). cert denied 423 S. 827 (1975); Continental Wax Corp v. FTC 330 F2d 475 (2nd Cir. 1964);Bakers FrOfL'hi-le Corp v, FTC. 302 F.2d 258 (3rd Cir. 1962);Mora Cigar Co. u. FTC, 107 F.2d 429 (4th Cir. 1939) , Opinion 104 F.

affrmative disclosures, could not correct the misimpression that Aspercreme contains aspirin. complaint counsel have argued that the evidence in' this case leads to such a conclusion (CAP 34-38). However we do not find that this evidence justifies brand name excision. The evidence consists in part of the testimony by complaint counsel' marketing expert. He stated that the brand name is the most salient part of a commercial to consumers and that, therefore, the misperceptions generated by the brand name "Aspercreme" cannot be overcome by any disclaimers included in ads (CAP 34-36). This line of reasoning appears to prove too much. It leads to the conclusion that the eommission must ban any brand name suggesting an (103) ingredient not contained in the product. As Thompson has pointed out, however there are numerous products on the market whose names suggest an ingredient they do not contain.B9 While no evidence is before us showing whether or not consumers are confused by those names, we think it probable that a properly designed ad campaign for such products or for Aspercreme, could convey to consumers the message that the product is similar to but not identical with the ingredient suggested by the brand name. In any event, we are not wiJJng to discount this possibility based upon one expert's opinion. The other evidence cited by complaint counsel (CAP 36-38) consists of the surveys on the record showing Thompson s TV disclosures that Aspercreme does not contain aspirin were (104) ineffective.9o Complaint counsel argue this shows disclosures cannot work. However complaint counsel agree with the ALJ and with us that Thompson disclosures were "woefully insuffcient" (CAP 37). Evidence that a B" See RAE 41:

The marktJplacc abounds with products whose marks suggest but do not describe a character or quality of the goods, as for example Bacos (no bacon), Sugar Twin (no suger), Egg Beater (no egg), Cremora (no cream), Silkience (no silk), CottonelJe (no cotton). Tuna Twist (no luna), Chock Full D' Nuts (no nuts), Chicken orthe Sea (no chicken). Apple Beer (no beer), and Rubbermaid (no rubber) (Ross, Tr. 6083-085; Silver, :r. 5662- 5664). The common sense "messge" inherent in these names is "similar to, i. , similar to bacon, similar to sugar etc.

W" have discussed these surveys, CXs 27, 32 .mrl 35, above, It is arguable that another survey sponsored by complaint counsel, the ASI Interlock Experiment (CX 26), provides informlltion about whether the brand name Aspercreme inherently leads consumers to believe Aspercreme contains aspirin However, we do not believe that the survey support. uch a conclusion, The objective ofCX 26 was "to find out whether or not the name Aspercreme led consumers to the inference that this produd contained aspirin as om ingredient. " (CX 26-B) The research design involved showing eouch respondent a single stimulus which included the name Aspercre, Ben.cay or Mobisyl as well as the phrase "for the temporary relief of minor anhritis pain " (CX 26-). They were then asked What ingredient or ingredients are Buggested by the nllme _ 1" (CX 26-Z32). We find this question to be ambiguous given the nature of the objective ofCX 26 because the wordingis not equivalent to asking people whether they belive a topical analgesic for the temporary relief of minor arthrtis pain contains aspirin, This ambiguity can be ilustrated with the following example. If people were asked what ingredients weresu.gestedby a cigarette with the name "Old Gold" tbe response "gold" would be expected from many. If they were instead asked what ingredients the productconlu.ined we would expect that very few would reply "gold. " The question in CX 26 is similarly flawed. Additionally, the ambiguity in ex 26 was probably heightelld because respondenl were not told that Aspercreme is a topical analgesic rather than ODe thatis taken intemally. Therefore CX 26 docs not provide probative evidence regarding whether the brand name Aapercreme causes reasonable consumers to believe that aspirin is an ingredient in Aspercreme. In addition, nothing in CX 26 tests whether or not any incipient potential fiwm;snprcPo!,ion coulrl hpovprN1mp bvdisclosures , ,j. g., .. IlV1Y..l oJV1 1Wl..' .JI"-1"1.. "-V., U'j"-. uva 648 Opinion poorly design d disclosure is ineffective, an unsurprising result, does not prove the inabilty of a well-desigoed disclosure to communicate a message to consumers. (105) VI. CONCLUSION For the reasons set forth above, we affrm the administrative law judge s finding ofliability and modify his initial decision as described. APPENDIX FTC POLICY STATEMENT REGARDING ADVERTISING SUBSTANTIATION Introduction On March 11 , 1983, the Commission published a notice requesting comments on its advertising substantiation program.! To facilitate analysis of the program, the notice posed a number of questions concerning the program s procedures, standards, benefits and costs, and solicited suggestions for making the program more effective. Based on the public comments and the staffs review, the Commission has drawn certain conclusions about how the program is being implemented and how it might be refined to serve better the objective of maintaining a marketplace free of unfair and deceptive acts or practices. This statement articulates the Commission s policy with respect to advertising substantiation. (2) The Reasonable Basis Requirement First, we reaffrm our commitment to the underlying legal requirement of advertising substantiation-that advertisers and ad agencies have a reasonable basis for advertising claims before they are disseminated. The Commission intends to continue vigorous enforcement of this existing legal requirement that advertisers substantiate express and implied claims, however conveyed, that make objective assertions about the item or service advertised. Objective claims for products or services represent explicitly or by implication that the advertiser has a reasonable basis supporting these claims. These representations ofsubstantiation are material to consumers. That is, consumers would be less likely to rely on claims for products and services if they knew the advertiser did not have a reasonable basis for believing them to be true.2 Therefore, a firm s failure to possess and rely (3) upon a reasonable basis for objective claims constitutes an unfair and deceptive act or practice in violation of Section 5 of the Federal Trade Commission Act. Standards for Prior Substantiation Many ads contain express or implied statements regarding the amount of support the advertiser has for the product claim. When the substantiation claim is express (e. tests prove doctors recommend", and "studies show ), the Commission expects the firm to have at least the advertised level of substantiation. Of course, an ad may imply more substantiation than it expressly claims or may imply to consumers that the firm has a certain type of support; in such cases, the advertiser must possess the amount and type of substantiation the ad actually communicates to consumers. Absent an express or implied reference to a certain level of support, and absent other I 48 l"R10471 March 11 , 1983.

2 Nor presumably would an adverti.'ur have made such claims unJe!L the advertiser thought they would be material to consumers.

840 FICDERAL TRADE COMMISSION DECISIONS Opinion 104 F.

evidence indicating what consumer expectations would be, the Commission assumes that consumers (4J expect a "reasonable basis" for claims. The Commission s deiermination of what constitutes a reasonable basis depends, as it does in an unfairness analysis, on a number of factors relevant to the benefits and costs of substantiating a particular claim. These factors include: the type afclaim, the product, the consequences of a false claim, the benefits of a truthful claim, the cost of developing substantiation for the claim, and the amount of substantiation experts in the field believe is reasonable. Extrinsic evidence, such as expert testimony or consumer surveys, is useful to determine what level of substantiation consumers expect to support a particular product claim and the adequacy of evidence an advertiser possesses. One issue the Commission examined was substantiation for implied claims. Although firms are unlikely to possess substantiation for implied claims they do not believe the ad makes, they should generally be aware of reasonable interpretations and will be expected to have prior substantiation (5) for such claims. The Commission will take care to assure that it only challenges reasonable interpretations of advertising claims.3 Prucedures !()r Obtaininf- Su.bstantiation In the past, the Commission has sought substantiation from firms in two different ways: through industry-wide "rounds" that involved publicized inquiries with identical or substantially similar demands to a number of firms within a targeted industry or to firms in different industries making the same type of claim; and on a case-by-case basis, by sending specific requests to individual companies under investigation. The Commission s review indicates that "rounds" have been costly to both the recipient and to the agency and have produced litte or no law enforcement benefit over a case-by-case approach. l6J The Commission s traditional investigatory procedures allow the staff to investigate a number offirms within an industry at the same time, to develop necessary expertise within the area of investigation, and to announce our activities publicly in circumstances where public notice or comment is desirable. The Commission intends to continue undertaking such law enfcJrcement efforts when appropriate. However, since substantiation is principally a law enforcement tool and the Commission s concern in such investigations is with the substantiation in the advertiser possession, there is little, if any, information that the public could contribute in such investigations. Therefore, the Commission anticipates that substantiation investigations wil rarely be made public befbre they are completed.

Accordingly, the Commission has determined that in the future it will rely on nonpublic requests for substantiation directed to individual companies via an informal access letter or, if necessary, a formal civil investigative demand. The (7) Commission believes that tailored, firm-specific requests, whether directed to one firm or to several firms within the same industry, are a more effcient. law enforcement technique. The Commission cannot presently foresee circumstances under which the past approach of industry-wide rounds would be appropriate in the ad substantiation area. Relevance uf Post-Claim Evidence in Substantiation Cases The reasonahle basis doctrine requires that firms have substantiation before disseminating a claim. The Commission has on occasion exercised its discretion, however to consider supporting materials developed after dissemination.4 The Commission has iIndividual Commissioners have expressed differing views tis lo how claims should be interpreted so that advertisers are not held r.o outlandish or tenuousinterprdations Kotwithstanding these variations in approach thefucusofaJlCommLssionerson reClsonabie interpretClr,ionsofclairns is intended tuensure lhi!ladvertisersare not required to substantii!tecli\imsthClt were notmClde 'The Commission s evident.iary rule, J6 C, R. 3. , hils sometimes j"",n interpreted as precluding introduct.ion of nost, r.i\Lm Hubstantiatioo. In Cae!. it doe not. Section3AO only provides a sanction again t the int.production of 648 Opinion not previously identified in one document the circumstances in which it may, in its (8) discretion, consider post-claim evidence in substantiation ca.o;cs.5 Such guidance can serve to clarify the program s actual operation as well as focus consideration of postclaim evidence on cases in which it is appropriate. The Commission emphasizes that as a matter of law, firms lacking a reasonable basis before an ad is disseminated violate Section 5 of the FTC Act and are subject to prosecution. The goal of the advertising substantiation requirement is to assure that advertising is truthflll, however, and the truth or falsity ofa claim is always relevant to the Commission s deliberations. Therefore, it is important that the agency retain (9) the discretion and flexibility to consider additional substantiating evidence, not as a substitute for an advertiser s prior substantiation, but rather in the following circumstances:

. When deciding, before issuance of a complaint, whether there is a public interest in proceeding against a firm;

. When assessing the adequacy of the substantiation an advertiser possessed before a claim was made; and . When deciding the need for or appropriate scope of an order to enter against a firm that lacked a reasonable basis prior to disseminating an advertisement. First, using post-claim evidence to evaluate the truth of a claim, or otherwise using such evidence in deciding whether there is a public interest in continuing an investigation or issuing a complaint, is appropriate policy. This does not mean that the Commission will postpone action while firms create post-claim substantiation to prove the truthfulness of claims, nor does it (10) mean that subsequent evidence of truthfulness absolves a firm of liability for failing to possess prior substantiation for a claim. The Commission focuses instead on whether existing evidence that claims are true should lead us in the exercise of our prosecutorial discretion to decline to initiate a law enforcement proceeding. Ifavailable post-claim evidence proves that the claim is true issuing a complaint against a firm that may have violated the prior substantiation requirement is often inappropriate, particularly in light of competing demands on the Commission s resources.

Second, post-claim evidence may indicate that apparent deficiencies in the pre-claim substantiation materials have no practical significance. In evaluating the adequacy of prior substantiation, the Commission wil consider only post-claim substantiation that sheds light on pre-existing substantiation. Thus, advertisers will not be all(Jwed to create entirely new substantiation simply because their prior substantiation was inadequate. (l1J Finally, the Commission may use post-claim evidence in dctcrmining the need for or appropriate scope of an order to be entered against a firm that lacked a reasonable basis. Thus, when additional evidence offered for the first time at trial suggests that the claim is true, the Commission may frame a narrower order than if there had been no post-claim evidence.

The Commission remains committed to the prior substantiation requirement and further believes that these discretionary factors will provide necessary flexibility. The Commission will consider post-claim evidence only in the circumstances listed above. But, whether it will do so in any particular case remains within its discretion. Self Regulation Groups and Government Agencies The Commission traditionally has enjoyed a close working relationship with self regulation groups and government agencies whose regulatory policies have some bear- The distinction between pre-claim tlnd post-claim evidence is only relevant when the charge is lack ofsubstantia. lion. For other charges, such ,1R falsity. when evidence was developed is irrelevant to its admissibility at trial Final Order 104 F.

iog on our law enforcement initiatives. The Commission wil not necessarily (12) defer however, to a finding by a self-regulation group. An imprimatur from a self-regulation group will not automatically shield a firm from Commission prosecution, and an unfavorable determination will not mean the Commission will automatically take issue, or find liability if it does. Rather the Commission wil make its judgment independently, evaluating each case on its merits. We intend to continue our useful relationships with self rcgulation groups and to rely on the expertise and findings of other government agencies in our proceedings to the greatest extent possible. By direction of the Commission.

FINAL ORDER The matter has been heard by the Commission upon the appeal of counsel for respondent Thompson Medical Company, Inc. and complaint counsel and upon briefs and oral argument in support of and in opposition to the appeals. The eommission, for reasons stated in the accompanying Opinion, has granted a portion of complaint counsel's appeal and denied that of respondent. Therefore It is ordered That the initial decision of the administrative law judge be adopted as the Findings of Fact and Conclusions of Law of the Commission except as is otherwise inconsistent with the attached Opinion.

Other Findings of Fact and eonc1usions of Law of the Commission are contained in the accompanying Opinion.

It is further ordered That the following Order to cease and Desist be entered:

ORDER It is ordered That respondent, Thompson Medical company, Inc. a corporation, its successors and assigns, and respondent's offcers representatives, agents and employees, directly or through any corporation, subsidiary, division or other device, in connection with the labeling, advertising, offering for sale, (2) sale or distribution of any over-the-counter "drug" as that term is defined in the Federal Trade Commission Act, in or affecting commerce, as "commerce" is defined in the Federal Trade eommission Act, do forthwith cease and desist from:

A. Employing the brand name "Aspercreme" for such products or otherwise representing directly or by implication that an active ingredient of such product is aspirin, unless such product contains aspirin in therapeutically significant quantities; prooided, however that the brand name "Aspercreme" may be used for such product if its .

InVLVLruVl lUCd..l..fiLJ ""\,.1. 648 Final Order advertising and labeling clearly and prominently disclose that the product does not contain aspirin.

(1) In television advertisements, an explicit and simple aspirin disclaimer statement (such as "ASPIRIN-FREE") shall be superimposed on the television screen simultaneously with a vocal aspirin disclaimer statement (such as "Aspercreme does not contain aspirin ) at the end of each advertisement.

(2) In radio advertisements, an explicit aspirin disclaimer statement (such as "Aspercreme does not contain aspirin ) shall be made at the end of each advertisement.

(3) In print advertisements, an explicit aspirin disclaimer statement (such as "ASPEReREME DOES NOT contain ASPIRIN" shall be displayed prominently and conspicuously in relation to each such advertisement as a whole.

(4) In labeling, an explicit aspirin disclaimer statement (such as DOES NOT contain ASPIRIN") shall be prominently and conspicuously printed in the front package panel (or in the front of the container if no package is used).

B. Representing, directly or by implication, that such product involves a new scientific principle, when such product or one involving such principle has been available for purchase in the United States as an over-the-counter drug for more than one year. (3) C. Misrepresenting the contents, validity, results, conclusions, or interpretations of any test or study.

D. Misrepresenting the identity ofthe active ingredient(s) in such product.

It is further ordered That Thompson Medical company, Inc., a corporation, its successors and assigns, and respondent's offcers, representatives, agents, and employees, directly or through any corporation, subsidiary, division or other device, in connection with the advertising, offering for sale, sale or distribution of any OTC analgesic drug," as that term is defined in the Federal Trade eommission Act in or affecting commerce, as "commerce" is defined in the Federal Trade eommission Act, do forthwith cease and desist from: A. Representing that such product is effective for the relief of minor pain and other symptoms of any musculoskeletal disorder (such as arthritis, tendonitis, bursitis, or rheumatic disorders). B. Representing that such product is as fast as or faster than, or is as effective as, or more effective than any other drug or device in the Final Order 104 F.

relief of minor pain and other symptoms of any musculoskeletal disorder (such as arthritis, tendonitis, bursitis, or rheumatic disorders); unless at the time of the dissemination of any such representation respondent possesses and relies upon a reasonable basis for such representation consisting of competent and reliable scientific or medical evidence. For analgesic drug products competent and reliable scientif ic or medical evidence shall include at least two adequate and wellcontrolled, double-blinded clinical ,tudies which conform to acceptable designs and protocols and are conducted by different persons independently of each other. Such persons shall be qualified by training and experience to conduct such studies. Provided howeoer with respect to any representation covered by this part other than claims of supe,'ior or comparative effectiveness or safety, if the Food and Drug Administration promulgates any final standard which establishes conditions under which such product is safe and effective under the Food, Drug and cosmetic Act, then in lieu of the above, respondent may rely upon scientific evidence which fully conforms to such final standards as a reasonable basis for said representation. (4) It is further ordered That so much of the complaint as relates to Paragraph 12 (I) be, and the same hereby is, dismissed. It is further ordered That respondent Thompson Medical Company, Inc. shall notify the Commission at least thirty (30) days prior to any proposed change in the corporation such as a dissolution, assignment or sale resulting in the emergence of a successor corporation, the creation or dissolution of subsidiaries or any other change in its corporation which may affect compliance obligations under this Order. It is further ordered That the respondent herein shall within sixty (60) days after service of this Order upon it and at such other times as the eommission may require, fie with the Commission a written report setting forth in detail the manner and form in which it has complied or intends to comply with this Order. UHJ.L .L v.L.Lv un'-U.L , u.v.

845 Complaint