This item is only available as the following downloads:
FUNCTIONALITY AND SECONDARY MEANING
TRADE DRESS OF BRAND NAME
PRESCRIPTION DRUG PRODUCTS
LARRY JACK SHAPIRO
A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY
UNIVERSITY OF FLORIDA
Larry Jack Shapiro
To my parents, my brothers, and Dora,
for being there when I needed them.
I would like to express my gratitude to Dr. William C. McCormick,
whose counsel and advice were invaluable throughout this investigation
and my graduate education.
Special thanks are extended to Dr. Richard A. Angorn, whose
influence and direction during my stay at the University will always be
I would like to acknowledge the other members of my supervisory
committee, Dr. Oscar Araujo, Dr. Douglas Bradham, Dr. Carole Kimberlin,
Mr. Max Lemberger, Dr. Ronald Marks and Dr. Pejaver Rao for their
I would like to acknowledge Harriet Douglas for her patience and
assistance in the preparation of this document and Bob Curtis for his
Finally, I wish to thank the Purepac Company for its assistance and
those pharmacies and pharmacists who cooperated in the study and made
this research possible.
TABLE OF CONTENTS
ACKNOWLEDGEMENTS . . . iv
TABLE OF CONTENTS. . . . v
ABSTRACT . . . vii
CHAPTER I INTRODUCTION . . 1
The Context of the Problem . . 1
The Purpose of the Study . . 4
Methodologic Framework . . 5
The Questions Addressed by the Study . 8
The Objectives of the Study . . 10
Statement of Hypotheses . . 11
CHAPTER II LEGAL BACKGROUND . . 14
Trademark and Unfair Competition Law . 14
Drug Product Copying Cases Prior to Ives v. Darby. 16
Ives v. Darby and Related Law . . 19
CHAPTER III REVIEW OF THE LITERATURE . 26
The Issue of Secondary Meaning . 26
The Issue of Functionality . . 29
Theoretical Foundations . . 33
CHAPTER IV METHODOLOGY . . 40
Overview . . . 40
The Survey Population . . 42
Selection of the Sample . .. 43
Laboratory Procedures . . 45
Instrumentation and Pilot Study . 46
Variables of Interest and Data Collection Procedures 47
Part I--The Personal Interview . 47
Part II--The Controlled Experiment . 50
Part Ill--The Self-administered Questionaire 53
Data Processing and Preparation . 55
Major Limitations of the Study . 55
CHAPTER V RESULTS AND ANALYSES . . 57
The Findings of the Study . . 57
Part I--The Personal Interview . 57
Part II--The Controlled Experiment . 69
Part III--The Self-administered Questionnaire 77
CHAPTER VI CONCLUSIONS . . 93
Summary and Discussion . . 93
Part I--The Personal Interview . 93
Part II--The Controlled Experiment . 96
Part Ill--The Self-administered Questionnaire 98
Implications and Recommendations . 100
Topics for Future Research . . 102
GLOSSARY . . . 105
A THE APPEARANCE OF THE DRUG PRODUCTS SELECTED
FOR THE INVESTIGATION . . 108
B PHARMACIES INCLUDED IN THE STUDY AND THEIR
LOCATIONS . . 112
C INFORMED CONSENT FORM . . 114
D THE INTERVIEW/QUESTIONNAIRE INSTRUMENT . 116
E VISUAL PROMPT DRUG FIRMS . . 122
F VISUAL PROMPT DRUG PRODUCT NAMES . 124
G DATA CODING SHEET . . 126
REFERENCES . . . 130
BIOGRAPHICAL SKETCH . . 136
Abstract of Dissertation Presented to the Graduate Council of
the University of Florida in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy
FUNCTIONALITY AND SECONDARY MEANING
TRADE DRESS OF BRAND NAME
PRESCRIPTION DRUG PRODUCTS
Larry Jack Shapiro
Chairman: William C. McCormick
Major Department: Pharmacy
A study was undertaken to examine several issues regarding the
functionality and secondary meaning of the appearance of brand name
A survey of 240 patients who were identified as chronically
taking at least one of 25 single-source, branded prescription drugs
was conducted in six community pharmacies in Gainesville, Florida.
The survey was accomplished utilizing a structured three-part inter-
In Part I, patients were shown samples of their prescription
drugs. A structured interview was conducted to determine by what
manner patients identify their medications and to ascertain whether
the patients were knowledgeable of the trademark name and source of
their drug products.
In Part II, a controlled experiment was conducted to determine
if patient acceptance of less expensive generic drug substitutes for
their brand name drugs is related to the look-alike appearance of
these substitute products.
In Part III, a self-administered questionnaire was utilized to
collect demographic data and other information related to the patients'
knowledge of their drug products.
In regard to secondary meaning, the results showed that signi-
ficantly more than a majority of patients recognized and initially
identified their drug product by its trademark name. Less than
one-fourth of the patients initially identified their medication by its
colors, functions, or effects. However, of those patients who were
knowledgeable of the trademark name, significantly less than a majority
associated that name with a specific company's product. Very few
patients were knowledgeable of their drug product's source of manufac-
ture. The findings support the claim that a majority of patients do not
associate the trade dress of their prescription drugs with a unique
In regard to functionality, there was insufficient evidence to
conclude that the trade dress colors of prescription drugs significantly
affect patients' choice among their brand name drugs and less expensive
generically equivalent drugs. This finding and additional results
regarding patients' attitudes toward drug product appearance refute the
claim that trade dress colors of brand name drugs are functional to
consumers for the reason that patients are less likely to accept non-
look-alike generic drug substitutes.
The Context of the Problem
The widespread adoption of state drug substitution laws during
the past decade, coupled with the increased availability and market
acceptance of less expensive generic versions of formerly patent-
protected prescription drugs, has substantially increased the demand for
generic drug products . However, the rapid growth of the generic drug
market has rekindled a number of hitherto dormant public policy and legal
controversies. One such controversy is exemplified by the recent rash
of unfair competition suits brought by certain brand name drug
manufacturers against other drug manufacturers who manufacture and sell
"look-alike" generic drugs which simulate the trade dress appearance of
their brand name products [2-3]. The issue for resolution is whether it
is legal to allow generic drug manufacturers to copy the nonpatented
features of color, shape, and size of brand name prescription drug
products which have become familiar to health care practitioners and
patients, or whether such copying results in sufficient public deception,
market misrepresentation, and false designation of origin as to consti-
tute unfair competition under both the common law and the various state
For a discussion of the original generic drug "substitution" problems,
see Galbally, J.J., "Substitution as Gross Immorality," Food Drug
Cosmetic Journal, 12:758, 1958; and Stamler, J.H., "Some Legal Aspects
of the Substitution Problem," Food Drug Cosmetic Journal, 8:643, 1953.
laws and to constitute unfair competition and contributory infringement
under the Lanham Act .
The drug product copying cases which were litigated prior to the
current trend which favors generic drug substitution clearly indicate
that the imitation by a drug manufacturer of the "nonfunctional" trade
dress features of another manufacturer's drug product constitutes unfair
competition if those features are shown to have acquired a "secondary
In order for nonfunctionality to exist, the drug product's trade
dress features could have no functional purpose or utility aside from
identifying the product and its source . If certain colors or other
distinguishing features were found to have therapeutic value, or if those
features could be shown to aid in the solubility, dispensing, or
commercial success of the drug product, then such characteristics could
be found to be functional and are therefore copyable [4:5].
In order for secondary meaning to exist, the drug product's trade
dress features must be shown to have acquired a new significance to a
substantial number of patient-consumers as an indication of the product's
source or uniqueness in the marketplace in a "trademark" sense
[5:516-22]. Once a nonfunctional drug product color or feature was shown
to have acquired a secondary meaning, it could be protected from copying
on the basis of common-law trademark .
In the 1980 case of Ives Laboratories, Inc. v. Darby Drug Co. et
al., a Second Circuit trial court refused to enjoin defendant
pharmaceutical firms from copying the trade dress coloring of plaintiff's
2488 F.Supp. 394 (E.D. N.Y. 1980).
drug product CYCLOSPASMOL . The court found that the plaintiff did
not meet its burden proving nonfunctionality and secondary meaning in the
trade dress color appearance of its drug capsules.
Testimony by several of defendants' physician witnesses claiming
that patients were more likely to associate the drug product's colors
with the drug itself or its therapeutic effect, rather than its source,
was accorded substantial weight. In addition, the court was persuaded by
defendants' arguments suggesting that the capsules' trade dress colors
were copyable because they had functions aside from source identifi-
cation. The defendants claimed that their use of the identical blue and
red color combination of CYCLOSPASMOL aided in the identification of the
drug entity in emergency medical situations and facilitated consumer
acceptance at the retail level which was necessary for the drug product's
commercial success. No hard data, however, were presented to support the
claims of functionality or claims of secondary meaning in the trade dress
colors of CYCLOSPASMOL.
In sharp contrast to Ives, a Court of Appeals for the Third Circuit
affirmed the lower court's decision in SK&F Co. v. Premo Pharmaceutical
Laboratories, and enjoined defendant drug firm from manufacturing and
distributing trade dress copies of plaintiff's drug product DYAZIDE .
Here, the court expressly rejected similar arguments that the appearance
of plaintiff's capsules was functional with respect to patients or to
health professionals. In addition, substantial weight was accorded to
plaintiff's contention which suggested that the capsule's appearance was
"arbitrary" and "distinctive" and that consumer recognition of DYAZIDE
625 F.2d 1055 (3rd Cir. 1980).
was sufficiently "widespread" to establish secondary meaning. As in
Ives, no hard data were presented in support of the claims of secondary
meaning or the claims of functionality in the trade dress of DYAZIDE.
The conflicting decisions in Ives and SK&F highlight the failure of
the judiciary to bring about uniformity in unfair competition law as it
relates to drug product copying and generic drug product substitution.
Inasmuch as the elements of unfair competition are not susceptable to
precise definition, and that a balance relative to drug substitution
laws, generic drug copying, and unfair competition has yet to be struck,
it appears that the outcome of "look-alike" drug product litigation is
less predictable than for other types of unfair competition cases.
Nevertheless, the records in Ives and SK&F reveal that the legal
conclusions on the issues of functionality and secondary meaning were
based on the tenuous opinion testimony of litigants' witnesses, circum-
stantial evidence, and uncontrolled studies rather than on any impartial
The Purpose of the Study
The principal purpose of the study was two-fold. First, in regard
to the issue of secondary meaning, the purpose was to investigate the
manner by which patients identify their brand name prescription drugs
when they are shown only the product's size, shape, and color trade dress
appearance and to ascertain whether patients have knowledge of their drug
product's trademark name and source. Second, in regard to the issue of
4Id. at 1059.
functionality, the purpose was to discover whether the trade dress
colors) of prescription drug products significantly impact on the
patients' choice between their brand name drugs and less expensive
generically equivalent drug substitutes.5
The issues of functionality and secondary meaning in drug product
trade dress encompass a substantial part of the legal questions posed in
the "look-alike" drug product cases. While the scope of this investi-
gation is not intended to cover every aspect related to the issues of
functionality and secondary meaning, it was undertaken to better clarify
several major points of contention and, perhaps, set the stage for
studies broader in character and in scope.
Frequently at issue in cases of trademark infringement and unfair
competition are the subjective mental associations, perceptions, or
reactions of a class of purchasers in regard to an alleged trademark .
Consequently, the parties in litigation have the burden of establishing
the "true" state of mind of purchasers so that the judge or jury can make
a decision based on the facts of the case .
The use of survey evidence in unfair competition cases as a means of
substantiating claims raised in the pleadings has become increasingly
popular for trademark owners . Such evidence, however, is considered
It appears that the duplication of colors) is at the heart of the
functionality issue, rather than the copying of the product's size and
shape. This can be pointed out by the fact that brand name
manufacturers have not pursued unfair competition suits against generic
drug manufacturers who have changed the trade dress colors of their
heresay. Since its admittance would result in the introduction into
evidence of statements made by interviewees who are not sworn or subject
to cross examination, the courts are careful in examining their technical
adequacy and prudently accord them weight .
The sample surveys introduced in trademark litigation have primarily
been accomplished through personal interviews, telephone calls, and mail
questionnaires . They have incorporated various research techniques,
including flash-card experiments and observations at the point of sale
. However, the cases clearly indicate that the most broadly appli-
cable and reliable survey format for use in trademark litigation is one
which employs a fixed form of interview seeking the unbiased reactions of
a representative sampling of relevant purchasers [14:156].
Although survey evidence may be admitted in court and given
substantial weight, there are no fixed requirements in trademark law as
to the number or percentage of purchasers who must be shown to think or
react in a specified manner in order to constitute adequate proof. The
cases indicate that a "substantial" or "appreciable" number of relevant
purchasers are necessary . However, a "substantial" or "appreciable"
number do not necessarily mean a majority of purchasers, and in a given
trademark case a showing of less than a majority may suffice [15:512].
As a general rule, a survey which shows that at least a majority of
purchasers (majority test) think or react in a specified manner will
Prior to the 1950's, survey evidence was frequently barred from use in
the courts by the hearsay rule. Several cases during the 50's set
precedent allowing the introduction of such evidence in court as an
exception to the hearsay rule. For an in-depth discussion of this
issue, see Licht, L., "Public Opinion Polls as Evidence in Unfair
Competition Cases," The Trademark Reporter, 46:1462-68, 1956.
usually suffice as proof as long as the survey is found to be technically
Social scientists have recorded a number of fundamentals of
designing and conducting surveys which, if adhered to, increase their
reliability and validity and thus promote their technical adequacy in
litigation [16-18]. These fundamentals include
(1) selecting an unbiased, representative sample from the relevant
(2) surveying the subjects in the proper setting and context;
(3) utilizing an unbiased and reliable measuring instrument;
(4) employing a fair and uniform method of questioning the
(5) gathering, processing, and tabulating the data accurately and
(6) applying accepted statistical procedures; and
(7) expressing and explaining the results and all relevant terms in
clear and unambiguous language.
Because of the nature of the issues to be examined in this
investigation, the research questions were framed in compliance with the
requisites of the majority test. In addition, the survey fundamentals
listed above were used as a framework for the design and implementation
of this study.
The Questions Addressed by the Study
This study was designed to provide answers to the following
questions related to the issues of functionality and secondary meaning in
drug product trade dress:7
1. (a) Do a majority of patients, who are shown only the trade
dress features of size, shape, and color of their prescrip-
tion drugs, initially identify their drug product by its
trademark name or source of manufacture in a trademark
(b) Do less than a majority of patients initially identify
their drug product by its colors, functions, or effects (e.g.,
"my red heart pill") without regard to trademark?
2. Is there a difference between the percentage of patients who
initially identify their drug in a trademark sense and the
percentage of patients who initially identify their drug in a
functional sense without regard to trademark?
3. If patients do not initially identify their drug product by its
trademark name, are a majority of patients knowledgeable of its
trademark name after verbal probing and visual prompting?
4. If patients do not initially identify their prescription drug
by its source of manufacture, are a majority of patients
knowledgeable of its source after verbal probing and visual
Questions one and two relate to both issues of functionality and
secondary meaning, questions three through six relate specifically
to secondary meaning, and question seven relates specifically to
5. (a) Given those patients who are knowledgeable of their drug
product's trademark name, do less than majority associate that
name with the trademark of a specific company's product?
(b) Do a majority of patients associate that name with the drug
entity itself without regard to the trademark of a specific
6. Is there a difference between the percentage of patients who
associate the trademark name with the trademark of a specific
company's drug product and the percentage of patients who
associate the name with the drug itself?
7. If patients are offered less expensive generically equivalent
drug substitutes in place of their brand name drugs, is there a
difference between the percentage of patients who choose
"look-alike" generic drug substitutes and the percentage of
patients who choose "nonlook-alike" generic drug substitutes?
In the context of this study, "look-alike" means a drug product of the
same size, shape, and color as the brand name drug. "Nonlook-alike"
means a drug product of the same size and shape but different color.
The Objectives of the Study
In attempting to answer the research questions, the following
objectives identified for the study were
1. To ascertain the percentage of patients who initially identify
their prescription drug by its trademark name or source of
manufacture in a trademark sense.
2. To ascertain the percentage of patients who initially identify
their prescription drug in a functional, nontrademark sense.
3. To discover whether the percentage of patients who initially
identify their prescription drug in a trademark sense is
significantly different from the percentage of patients who
initially identify their prescription drug in a functional
sense without regard to trademark.
4. To ascertain the percentage of patients who can identify the
trademark name of their drug product after assistance through
verbal probing or visual prompting.
5. To ascertain the percentage of patients who can identify the
source of manufacture of their drug product after assistance
through verbal probing or visual prompting.
6. To ascertain the percentage of patients who are knowledgeable
of their drug product's trademark name and who associate it
with the trademark of a specific company's product.
7. To ascertain the percentage of patients who are knowledgeable
of their drug product's trademark name but who associate it
with the name of the drug itself.
8. To discover whether the percentage of patients who are
knowledgeable of their drug product's trademark name and who
associate it with a trademark of a specific company's product
is significantly different from the percentage of patients who
are knowledgeable of their drug product's trademark name but
who associate it with the name of the drug itself.
9. To discover whether there is a significant difference between
the percentage of patients who would switch to a look-alike
generic drug substitute and the percentage of patients who
would switch to a nonlook-alike generic drug substitute.
Statement of Hypotheses
The following nine research hypotheses were formulated to address
the aforementioned research questions related to functionality and
secondary meaning in prescription drug product trade dress:
1. More than 50 percent of the patients initially identify their
drug product by its trademark name or source of manufacture in
a trademark sense.
2. Less than 50 percent of the patients initially identify their
drug product by its colors, functions, or effects without
regard to trademark.
3. There is a significant difference between the percentage of
patients who initially identify their drug product in a
trademark sense and the percentage of patients who initially
identify their drug product by its colors, functions, or
effects without regard to trademark.
4. More than 50 percent of the patients who do not initially
identify their drug product by its trademark name are
knowledgeable of its trademark name after verbal probing and
5. More than 50 percent of the patients who do not initially
identify their drug product by its source of manufacture are
knowledgeable of its source of manufacture after verbal probi
and visual prompting.
6. Less than 50 percent of the patients who are knowledgeable of
their drug product's trademark name associate that name with
the trademark of a specific company's product.
7. More than 50 percent of the patients who are knowledgeable of
their drug product's trademark name associate that name with
the drug entity itself without regard to the trademark of a
specific company's product.
8. There is a significant difference between the percentage of
patients who associate the trademark name with the trademark
a specific company's drug product and the percentage of
patients who associate the trademark name with the name of th
drug entity itself.
9. There is a significant difference between the percentage of
patients who choose "look-alike" generically equivalent drug
products and the percentage of patients who choose
"nonlook-alike" generically equivalent drug products when
offered a less expense generic drug substitute in place of
their brand name drug.
Content of the Remaining Chapters
Chapter II presents the background information on trademark and
unfair competition law, the relevance of the drug product copying cases
litigated prior to Ives v. Darby, and the evolution of the Ives v. Darby
Chapter III presents the literature review regarding the secondary
meaning and functionality issues and provides the theoretical foundation
of the study.
Chapter IV constitutes the methodology utilized to conduct the
Chapter V gives the results and statistical analyses of the study in
respect to the objectives and the specific research hypotheses outlined
in Chapter I.
Chapter VI, the final chapter, presents a summary and discussion of
the study results, the implication of the findings, and suggestions for
Trademark and Unfair Competition Law
Protection of ownership interests in diverse forms of intangible
personal property has long been recognized under the common law . In
addition, Congress, under authority of Article I, Section 8, Clauses 3
and 8 of the Constitution of the United States, has accorded legal
protection in the ownership rights of inventions, artistic works, and
trade symbols by enacting patent, copyright, and trademark laws which
protect these rights and reserve to the owners the exclusive use of their
business property [5:121-45].
Commercial excesses in the deliberate copying of a competitor's
product or packaging in order to achieve a greater quantum of success in
the marketplace date back many centuries [6:361]. While this practice
has to some degree been controlled by our patent laws, it is primarily
the laws of trademarks and unfair competition, based on common law
rights, which are most relevant to the drug product copying cases .
Historically, Anglo-American trademark law arose from a series of
court cases which over the years established common law protection of
trademark property [19:310-26]. Recovery for trademark infringement
was recorded in England  as far back as 1824 in Sykes v. Sykes.I
3 B. & C. 541 (1824). This English chancery court held in favor of
trademark protection because of defendant's conduct in palming off his
goods as those of the plaintiff.
Modern trademark law in the United States is derived partly from its
common law basis and partly from specific statutory enactments, most
notably the Lanham Trademark Act of 1946 . Under the modern common
law concept, trademark infringement has come to be viewed, in effect, as
unlawful trading on the goodwill and reputation established by another
The law of unfair competition emerged out of the common law
protection accorded to trademarks [19:336]. In the earliest case law,
unfair competition was recognized when a seller was shown to have "passed
off" or "palmed off" goods of his own manufacture when the goods of
another manufacturer were requested by the purchaser [15:170]. The act
of palming off was most often facilitated by defendant's intentional
copying of the plaintiff's product labels, product features, or overall
product appearance .
An action for unfair competition was inherently broader in scope
than an action solely for trademark infringement [5:44-45]. Every
physical feature of the defendant's goods could be relevant, including
the product's packaging, labeling, trademark, configuration, features and
colors [5:45]. Consequently, in a suit for unfair competition, the court
had no need to focus only on one aspect of the plaintiff's product as it
did in the case for trademark infringement. However, in both trademark
infringement and unfair competition suits the keystone legal objective
was the same; it was to prevent commercial deception due to confusion or
a likelihood of confusion in the minds of the purchasing public [5:46].
Drug Product Copying Cases Prior to Ives v. Darby
The United States Supreme Court in 1924 rendered its decision in the
leading drug product copying case of William R. Warner & Company v. Eli
Lilly . The case involved defendants' imitation of Lilly's
nonpatented chocolate-flavored quinine product COCO-QUININE which was
first marketed in 1899. The Court found that Lilly had no right to the
exclusive use of chocolate in its quinine preparation on the grounds that
chocolate was functional as a flavoring and suspending agent, aside from
imparting a distinctive color, and thus free for any manufacturer to
incorporate in its products.
Testimony in the case revealed that several of the defendants'
selling agents made statements to pharmacy retailers suggesting that
Warner's less expensive drug product could be substituted for the Lilly
product without danger of detection. In addition, deceptive
substitutions by druggists were shown to have occurred on numerous
occasions. The evidence of passing off and predatory marketing tactics
practiced by defendants' selling agents were sufficient to find Warner
and codefendants contributorially liable for the unfair competition
practiced by pharmacy retailers.
The Court granted Lilly limited relief. It enjoined defendants'
selling agents from suggesting to pharmacy retailers that they could
substitute without notice and required the defendants to place a
disclaiming statement on its drug product packages indicating that their
preparation was "not to be sold or dispensed" as those of Lilly. The
2265 U.S. 526 (1924).
31d. at 533.
Court, however, expressly refused to enjoin the manufacture and sale of
defendant's quinine preparation which was similar in appearance to the
The drug product copying cases which followed Warner v. Lilly
generally relied on its rulings and adhered to its remedies,4-7 but with
some variations. If the copied features of the drug product were found
to be both nonfunctional and to have acquired a secondary meaning, and if
defendants were found to be contributorially liable for encouraging or
inducing "passing off" activities practiced by pharmacy retailers, then
Smith, Kline & French Laboratories v. Clark & Clark et al., 157 F.2d 725
(2nd Cir. 1946).
Smith, Kline & French Laboratories V. Waldman, 60 F.Supp. 646 (E.D.
Upjohn Company v. Schwartz, 246 F.2d 254 (2nd Cir. 1957).
Norwich Pharmacal Company v. Sterling Drug, Inc., 271 F.2d 569 (2nd Cir.
the courts would enjoin the manufacture and sale of the drug product
similar in appearance to that of the plaintiff. 812
In some courts an exception to the traditional rules was recognized.
If the plaintiff, in seeking relief, could show actual passing off or a
likelihood of consumer deception as a result of defendant's copying and
predatory marketing tactics, then the plaintiff was freed from having to
meet the requirements of secondary meaning and/or nonfunctionality in its
drug product trade dress features.13-16 The cases indicate that such
predatory marketing tactics or inducements to illegally substitute could
exist in the form of salesmens' statements, postcard advertisements,
deceptive packaging and labeling, and comparison promotional literature
. However, in Marion Laboratories Inc. v. Michigan Pharmacal
Smith, Kline & French Laboratories v. Heart, 90 F.Supp. 976 (S.D. N.Y.
Smith, Kline & French Laboratories v. Lipton, 89 U.S.P.Q. 418 (N.D. Ohio
10Ross-Whitney Corp. v. Smith, Kline & French, 207 F.2d 190 (9th Cir.
1Smith, Kline & French Laboratories v. Broder, 125 U.S.P.Q. 298 (S.D.
2Smith, Kline & French Co. v. Premo Pharmaceutical Laboratories, Inc.,
625 F.2d 1055 (3rd Cir. 1980).
13Martin H. Smith Company v. American Pharmaceutical Co., Inc. et al.,
200 N.E. 779 (C.A. N.Y. 1936).
14E.R. Squibb & Sons, Inc. v. Premo Pharmaceutical Labs, Inc., 195
U.S.P.Q. 545 (S.D. N.Y. 1977).
15Merrell-National Laboratories, Inc. v. Zenith Laboratories, Inc.
et al., 194 U.S.P.Q. 157 (D.C. N.J. 1977).
16Pennwalt Corporation v. Zenith Laboratories, Inc., 472 F.Supp. 413
(E.D. Mich. 1979).
Corporation,17 the defendant's sales catalog which compared plaintiff's
drug product to its own product was ruled not to constitute contributory
unfair competition because it was found not to have induced pharmacists
to engage in deceptive substitution.
Ives v. Darby and Related Law
In 1978, Ives brought its first suit against Darby Drug Company and
other drug firms for trademark infringement of its registered mark
CYCLOSPASMOL. The action also charged contributory trademark
infringement and unfair competition under the Lanham Act, and unfair
competition under New York state law due to defendant's sale of
identically colored generic imitations of cyclandelate capsules [27).
Aside from the traditional claim of common law unfair competition, Ives
claimed that defendants should be liable for contributory infringement
and unfair competition under the Lanham Act on the ground that
defendant's "look-alike" capsules and promotional catalogs encouraged
retail pharmacists to illegally substitute or deceptively label generic
cyclandelate with the CYCLOSPASMOL trademark. Apart from the pleadings,
New York's drug product selection law, which mandated generic drug
substitution, opened up a public policy issue for which a balance between
unfair competition and drug product copying had to be struck.
The remedies for infringement of a registered trademark under the
Lanham Act appear in section 32.18 The Act prohibits the use in commerce
of any reproduction, counterfeit, copy, or colorable imitation
17338 F.Supp. 762 (E.D. Mich. 1972).
1815 U.S.C. 1114.
of a registered mark without the consent of the registrant. Infringement
exists if defendant's use of the mark is likely to cause confusion,
mistake, or deceive consumers [27:235]. This section has also been
construed so as to make any party contributorially liable for the
infringing action of another party if he supplies the means by which a
consumer may be deceived [27:235]. The contributing infringer, however,
must be shown to have deliberately induced, encouraged, or knowingly
facilitated the infringing action of another party in order for liability
to attach [26:39-42].
Under section 43(a) of the Act, false designations of origin and
false descriptions are .forbidden.19 This section has become recognized
as the federal law of unfair competition [281. Claims made under this
section frequently stem from a competitor's copying either a prominent
product feature or the product's overall appearance, apart from any
claims of trademark infringement [27:236]. However, to sustain an unfair
copying claim under section 43(a), the plaintiff must show existence of
secondary meaning under the common law because he lacks the presumptive
source association of a registered trademark [27:237]. That is, the
plaintiff must establish "trademark" significance in the product's
contested feature in order to show a likelihood that consumers can be
confused. In addition, the product feature alleged to have trademark
significance must meet the requirement of nonfunctionality [27:237].
1915 U.S.C. 1125a.
In Ives, the lower court refused to grant a preliminary injunction
against defendant's drug product copying.20 In regard to the unfair
competition claim under section 43(a), the court found that Ives did not
unequivocally establish that the colors of its CYCLOSPASMOL capsules were
nonfunctional or had acquired a secondary meaning. In addition, it found
insufficient evidence that defendants had conspired with pharmacists or
suggested that they improperly substitute or label generic cyclandelate
with Ives' registered trademark CYCLOSPASMOL in violation of federal or
state unfair competition law. The lower court indicated that the Lanham
Act could not be employed to prevent manufacturers of generic drugs from
copying the colors of brand name drugs and that granting a monopoly in
the colors of plaintiff's drug product would "inhibit legitimate
substitutions which the New York legislature has found in the public
On the appeal from the denial of preliminary relief, the Second
Circuit refused to overturn the lower court's decision.22 However, it
differed as to the criteria required to establish contributory
infringement claims under section 32 of the Lanham Act. The court held
that a drug product imitator may be liable for contributory infringement
if he suggested, even if only by implication, that a retailer fill a
bottle with generic capsules and apply Ives' nark CYCLOSPASMOL to the
20Ives Laboratories, Inc. v. Darby Drug Co. Inc. et al., 455 F.Supp. 939
(E.D. N.Y. 1978).
1Id. at 951.
22601 F.2d 631 (2nd Cir. 1979).
601 F.2d 631 (2nd Cir. 1979).
label.23 Nevertheless, the appellate court was constrained to agree with
the lower court's findings of fact that Ives had failed to provide that
quantum of evidence necessary to sustain a preliminary injunction under
either section 43(a) or section 32. It remanded the case to trial under
its analysis and indicated that if defendants could show evidence which
would suggest that copying was an essential function to the commercial
success of the generic drug, then that evidence would be especially
On remand, the trial court rejected Ives' argument that defendant's
"look-alike" drugs and promotional catalogs suggested, even by
implication, that pharmacists fill bottles with defendant's generic
cyclandelate and apply Ives registered trademark CYCLOSPASMOL.25 In
regard to Ives' unfair competition claim under section 43(a), the court
once again found that Ives did not meet the traditional requirements of
secondary meaning or nonfunctionality. On the contrary, it found that
the color of Ives' capsules might have identification functions in
medical emergencies and commercial success functions. The court also
indicated that consumers associate the drug colors more often with the
drug itself or its effect, rather than a particular source or in a
23Id. at 636.
4Id. at 644.
25488 F.Supp. 394 (E.D. N.Y. 1980).
26Id. at 400.
Additional evidence presented by Ives indicated that of 35 pharma-
cists who substituted a generic cyclandelate product in place of CYCLO-
SPASMOL, 10 labeled the prescription with the CYCLOSPASMOL trademark in
some manner.27 However, only one pharmacist was found to have passed off
the generic drug as CYCLOSPASMOL and to have charged the brand name
price. The court dismissed the mislabeling evidence as instances where
"druggists have misunderstood the precise requirements of New York drug
substitution laws" and stated that pharmacists "appear not deliberately
to have been attempting to pass off the generic drug product as
The trial court's decision on the claim of contributory infringement
was appealed on the ground that defendant's drug copying, comparison
catalogs, and additional evidence of mislabelings did, undisputedly,
establish a violation of section 32 of the Lanham Act. On the second
appeal, the Second Circuit concluded (in a 2 to 1 decision) that the
evidence clearly established defendant's liability for contributory
infringement.29 The court reasoned that defendants should have
anticipated that the use of "look-alike" capsules and comparison catalogs
would result in illegal substitutions and trademark mislabelings by a
substantial number of pharmacy retailers. No judgments were reached on
Ives' unfair competition claims under section 43(a) of the Act. The
court was of the opinion that the simplest way to reduce illegal
7Id. at 397.
29638 F.2d 538 (2nd Cir. 1981).
substitutions and mislabelings would be to require defendants to sell
their drug products in capsules which did not resemble CYCLOSPASMOL.30
Defendants appealed to the Supreme Court on the ground that the
Appellate Court erred in finding them in violation of section 32 of the
Lanham Act given the evidence at trial.31 Certiorari was granted and the
case was heard in February, 1982. The key questions presented to the
Court in defendant's petitions for certiorari are summarized below.32
1. Can the trade dress appearance of a pharmaceutical firm's
nonpatented drug product gain sufficient "trademark"
significance in the marketplace to grant pharmaceutical firms
an exclusive property right in size, shape, and color?
2. Does the intentional trade dress copying of an innovator
pharmaceutical firm's nonpatented drug product constitute
sufficient market misrepresentation and contributory
infringement, false designation of origin, and unfair
competition under federal and state law to preclude generic
pharmaceutical firms from the manufacture and sale of generic
drug product "look-alikes" in light of drug product selection
laws and public policy which fosters generic substitution?
The opinion of the Court was delivered on June 1, 1982.33 The Court
did not address the broader issues related to drug product copying as
presented in the defendants' petitions for certiorari. Instead, it
unanimously reversed the Court of Appeals on the narrow issue of
0Id. at 545.
31BNA's Patent Trademark & Copyright Journal, 568:A-1, February, 1981.
3250 U.S.L.W. 3266 (October 1981).
33BNA's Patent Trademark & Copyright Journal, 582:103, June, 1982.
BNA's Patent Trademark & Copyright Journal, 582:103, June, 1982.
contributory infringement (section 32) and remanded the case back for
review of Ives' section 43(a) claim (false designation of origin) which
was not reached at the appellate level.
The Court ruled that the Court of Appeals made a procedural error by
setting aside the lower court's findings of fact without showing them to
be clearly erroneous. The majority opinion indicated that (1) the lower
court's findings were not unreasonable; and (2) that the Court of Appeals
had unjustifiably diluted the test for contributory infringement under
Unless the Court of Appeals finds the lower court's findings clearly
erroneous, it appears that Darby and codefendants will be exonerated of
all charges. Nevertheless, the Supreme Court ruling clearly indicates
that brand name drug manufacturers can still seek to prevent copying of
their products under section 43(a) of the Lanham Act, as well as state
unfair competition law, if they can present stronger evidence in support
of nonfunctionality and secondary meaning in their drug products' trade
dress. Since functionality claims and secondary meaning claims will
continue to be debated by drug manufacturers in future "look-alike"
cases, this research may be valuable to the parties because it
constitutes an evaluation of the issues based on objective, controlled
REVIEW OF THE LITERATURE
The Issue of Secondary Meaning
One undisputed consequence of state generic drug substitution
legislation has been the increased frequency with which generic drug
look-alikes have entered the market. The benefits, drawbacks, and legal
actions resulting from the marketing of such products have been widely
discussed in the literature [2,3,6,29-36].
Two fundamental legal issues regarding look-alike prescription drugs
are examined in this study. The first is whether the size, shape, and
color appearance of brand name drugs are sufficiently distinctive to
identify medications in the same way a trademark aids a consumer in
recognizing a specific product. In trademark law, the proof of
distinctiveness is equated with the finding that the product's colors and
features have acquired a secondary meaning to consumers thereby
identifying a unique product source.
As might be expected, brand name drug manufacturers claim that drug
product colors and shapes develop a secondary meaning to patients with
extended use of the medication. Generic drug manufacturers, on the other
hand, claim that such features identify only the drug entity to patients
rather than a drug product produced by a specific company. If generic
drug manufacturers' claims are upheld, features such as size, shape, and
color will have no trademark protection and may be copied at will.
The literature reveals a lack of empirical research regarding the
manner in which patients primarily identify their prescription drugs or
how patients perceive prescription drug brand names. The lack of studies
in this area may be due to pharmaceutical companies' hitherto traditional
practice of refraining from advertising or promoting prescription brands
directly to the public . Nevertheless, the subject of brand
awareness, brand perception, and brand preference has been extensively
studied in respect to other consumer goods.
The marketing literature indicates that an individual's awareness
and knowledge of brands is correlated with the nature of the product, its
purchase frequency, mass media exposure (advertising and promotion) and
opinion leadership [38-40]. In addition, predispositional factors such
as sex, age, education, income, and social class have been found to be
associated with an individual's awareness, knowledge, and perception of
specific brands [41-42].
A common methodology used to evaluate the ability of consumers to
identify and distinguish brand products is controlled experimentation.
Allison and Uhl tested 326 beer drinkers with unlabeled beer products in
order to determine if they could identify and distinguish by taste
specific brands with which they were familiar . Charton and
Ehrenberg used a consumer panel of 180 housewives to examine brand
identification and brand loyalty of frequently purchased consumer goods
. Participants were asked to purchase once a week one brand of four
unmarked detergent products and one brand of three unmarked tea products
which were priced at different levels but which did not differ in
Another method used to evaluate consumer knowledge and perception of
branded goods incorporates self-administered questionnaires. Keiser used
a questionnaire instrument which incorporated visual prompting with
product brand names . Subjects were asked to indicate the type of
product which was associated with a specific brand name. Peter and
Tarpey used questionnaires on 217 college students to measure brand
identification, perception, and preference of automobiles . The
subjects were prompted with profile cards of automobiles that contained
different kinds of information on each brand and asked to indicate their
opinions on seven point semantic differential scales.
In addition to controlled experiments and questionnaires, personal
interviews have frequently been used to measure consumer brand knowledge.
Woodside and Fleck conducted in-depth interviews with consumers at their
homes to evaluate brand knowledge and brand choice of beer . Rao
conducted interviews with 50 subjects to examine whether profile
information of automobiles affected brand perception and brand preference
. Bogart and Lehman interviewed 400 housewives in their homes to
evaluate brand awareness of goods . The subjects were paid five
cents for every brand they could recall in a product class without aid or
prompting. It is interesting to note that brand name recalls for
pharmaceuticals (prescriptions and otc) represented only 1 percent of all
brands mentioned by the subjects.
In summary, a review of the literature indicates that very little
empirical research has been conducted in regard to consumers' brand
awareness, brand knowledge, or brand perceptions of prescription drugs.
Nevertheless, the marketing literature reveals that the subject has been
studied extensively for other types of consumer goods. Methodologies
incorporating either controlled experiments, questionnaires, and personal
interviews are used most often to study consumer knowledge and perception
of branded products. Apparently, most studies utilize a combination of
two or more of the methods.
Even though pharmaceutical companies have not traditionally
advertised and promoted prescription brands directly to the public, at
least one manufacturer has broken tradition and started to do so with one
of its prescription drugs . Recent literature indicates that this
practice may become more prevalent in the near future [51-52]. As a
consequence, an increase in reported research related to patients'
knowledge and perception of brand names of prescription drugs can be
expected in the literature.
The Issue of Functionality
The second legal issue examined in this study is that of
functionality in the trade dress colors of brand name drugs. Generic
drug manufacturers contend that drug product colors are functional and
therefore are copyable because patients would be less likely to purchase
a less expensive generically equivalent drug product which looks
different from the brand name drug. Brand name drug manufacturers rebut
this argument by contending that drug product colors are arbitrary,
inherently nonfunctional, and do not affect a patient's willingness to
purchase a generic drug. If generic drug manufacturers' claims of
functionality are upheld, they will then be free to market "look-alike"
generic versions of patent-expired brand name drugs.
There have been relatively few studies reported in the literature
regarding consumer attitudes and acceptance of generic drug products.
Nelson and Gagnon mailed out questionnaires to 1000 housewives and
reported that only 20 percent of the respondents absolutely preferred
that the pharmacist not substitute a generic drug . More than 65
percent of the respondents indicated they would accept a less expensive
generic substitute on most prescriptions unless the physician had
specifically ordered a certain brand name be dispensed. The
investigators also found that willingness to accept a generic product
increased with the individual's income and education but decreased with
age. In a separate study, the same investigators found that a majority
of consumers were willing to accept a generic drug product if the savings
were as little as 5 percent [541.
Mason and Bearden conducted interviews and questionnaire studies
with pharmacists, physicians, and consumers in order to determine salient
factors of generic drug product prescribing and acceptance . The
interviews revealed that dimensions of quality, price, and effectiveness
were important to consumers. The dimensions of safety, side effects,
reputation of manufacturer and efficacy were found to be most important
to health practitioners.
In another study, Bearden, Mason and Smith found that willingness to
accept a generic drug substitute in an elderly population was associated
with the perceived risk . Older patients were found to perceive more
overall risk toward generic drug prescribing than younger patients.
Lambert, Doering, Goldstein, and McCormick used a questionnaire
instrument to measure the impact of two levels (high and low) of brand
name drug prices on generic drug acceptance . They found that
approximately one-third of the 510 respondents were willing to accept a
generic substitute for either high priced or low priced brand name drugs.
Age and perceived drug effectiveness were linked most consistently with
drug choice. In addition, generic drug rejection was found more likely
to be associated with females, with persons having lower annual incomes,
with persons having low general drug knowledge, and with persons having
beliefs that serious consequences could result from the use of less
In an exploratory study, Carroll and Jang reported that 66 percent
(35 of 53 patients) of their sample preferred to have a generic drug
. About 77 percent of the subjects felt that generics were the same
as brand names in terms of effectiveness, 75 percent felt they were the
same in terms of safety, and 95 percent felt they were less expensive
than the brand name drug.
LaBarbera conducted a questionnaire study to examine the acceptance
of generic foods and generic drugs and found that approximately 60
percent of the respondents were willing to accept generic products .
She reported that 41 percent of respondents had purchased generic foods
or drugs in the past. Of those subjects who had not purchased generics,
52 percent indicated they were willing to try generics in the future.
White and Geiger used a questionnaire instrument to measure consumer
attitudes and behavior toward drug product substitution . The
results from 448 respondents indicated that the subjects' beliefs about
the quality of the generic drug, their years of education, the number of
prescriptions they purchased annually, and their annual expenditure for
drugs were most associated with their willingness to purchase generic
drug products. No significant relationships were found between willing-
ness to accept generic drugs and predispositional factors such as age,
income, and who paid for the prescription. Approximately 40 percent of
the sample indicated they had purchased a generic drug sometime in the
Baldwin and Berger used a cartoon technique to evaluate consumer
acceptance of drug product substitution activities of pharmacists .
Subjects were asked to supply the speech of a patient (cartoon) character
responding to four different versions of a pharmacist suggesting a
generic drug substitute. The results from 528 subjects indicated that
slightly over 50% would have accepted the generic drug substitute.
Perceptions of quality, equivalency, and price were most associated with
generic drug acceptance. Demographic factors were not studied in this
The differences in results of the cited studies can be partially due
to the diverse designs and techniques used to measure generic drug
acceptance. A basic methodological problem addressed by several of the
investigators was the manner in which "generic drug" was defined.
Carroll and Jang reported that only 59 percent of their sample understood
the term "generic" . LaBarbera reported that only 50 percent of her
sample were actually knowledgeable of generic drugs although 58 percent
had indicated they thought they knew the meaning .
In summary, the literature indicates that perceptions of quality,
price, and effectiveness may be the most important factors related to
patient willingness to accept generic drugs. In addition, the studies
indicate that predispositional factors such as an individual's sex, age,
income, education, drug product consumption, drug expenditures, and
previous experiences with generic products may affect acceptance of
generic drugs. Research related to the impact of drug product appearance
on the acceptance of generic drug substitutions was not, however,
investigated in any study. In considering the legal debate regarding the
functionality issue and the lack of information on this subject, the
investigator felt this issue was important to address through controlled
study so that the results might be useful to decision makers.
The purposes of this study are to examine how patients identify
their prescription drug products, to ascertain whether patients are
knowledgeable of their drug product's name and source, and to evaluate
patients' choices between their brand name drugs and less expensive
generically equivalent look-alike and nonlook-alike drugs. The
theoretical foundations of the study arise from psychology, behavioral
science, and in particular from consumer behavior.
Two fundamental theoretical positions underlying consumer behavioral
research have achieved prominence. The first is termed "drive-habit"
theory, which is adapted from the classic work of stimulus-response
learning theorists such as Thorndike, Pavlov, Skinner, and Hull .
Drive-habit models of consumer behavior posit that consumer purchasing is
more or less a learned response. The interplay of internal drives
(needs), external stimuli (cues), and rewards (reinforcements) are
assumed to play major roles in determining purchase behavior . A
variety of stochastic models of brand choice based on learning theory are
present in the literature .
The second theoretical position is termed "expectancy-value" theory
and is associated with the work of psychologists such as Tolman and
Lewin . According to the general theory, man is seen as responding
to the expected consequences of a given behavior . Expectancy-value
models of consumer behavior emphasize goal-directed action tendencies.
The action tendency is defined to be a function of the perceived
magnitude or utility of the goal object and its probability of
attainment. One model which has received a great deal of attention in
the marketing literature is the expectancy-value model adapted from a
formulation proposed by Fishbein . The "extended" version of
Fishbein's behavioral model as adopted by Ryan and Bonfield  is
illustrated conceptually in Figure I.
According to this theory, an individual is assumed to form a
specific behavioral intention which directly determines his subsequent
overt behavior. Two major factors are assumed to determine behavioral
intention: (1) the individual's attitude towards the behavior, and (2)
the individual's subjective norm (social influence) towards the behavior.
The attitude component is defined to be a function of the individual's
beliefs about the consequences of performing the behavior and the
individual's evaluation (e.g., goodness or badness) of the expected
consequences. The subjective norm is defined to be a function of the
individual's normative beliefs (e.g., the person's belief that a referent
group or individual thinks he should or should not perform the behavior)
and the individual's motivation (e.g., desire or lack of desire) to
comply with those beliefs.
u- 8c D
0- 0 ^
The Fishbein model has been extensively investigated and found
useful in explaining and predicting an individual's brand preference
[66-70]. In a consumer behavior context, the model takes the form as
illustrated in Figure II.
The assumption underlying the application of the model to consumer
purchasing behavior is that an individual develops purchase attitudes
towards specific brands of products which determine his purchase
intention . The individual's purchase intention is then assumed to
direct his purchase behavior.
According to the model, an individual will process information in
forming specific beliefs about salient product attributes (e.g., style,
appearance, price, texture, taste, etc.) to arrive at a purchase attitude
toward each known brand. The strength or magnitude of the attitude
formed towards each brand is assumed to shape purchase intention and
thus, direct purchase behavior. In addition to attitude, the theory
holds that the individual's social or subjective norm (e.g., the strength
of beliefs regarding brand choice of referent individuals or groups) will
shape his purchase intention. However, there is some disagreement about
the usefulness and predictive power of this component of the model .
It is clear from the literature that attitudes and even purchase
intentions are not completely reliable predictors of actual buying
behavior. Predispositional factors, such as demographic variables 
(e.g., age, sex, education, income, social class) and personality
variables  (e.g., traits and life-styles) have been found to be
associated with brand choice. In addition, situational factors (e.g.,
importance of purchase, financial state, product availability) are known
Q QI v
3 "5s c/
to shape purchase intentions and purchase decisions [751. These factors
are illustrated in the model in Figure II.
In this study, the manipulated independent variable and attribute of
interest is the color appearance of a generic drug product substitute.
The dependent variable is the patients' purchase choice of either the
brand name drug which they are familiar with or a less expensive
look-alike or nonlook-alike generic drug substitute. The controlled
experiment involves exposing patients to different appearing samples of
fictitous generic drug products to test whether the looks of the generic
product affect patients' purchase intention. Because the generic drug
samples are fictitious, patients can have no prior brand beliefs or
attitudes toward these products. Thus, the Fishbein model as applied to
this research involves only belief and attitude formation, not change.
In designing the study, the investigator attempted to control for
salient product attributes of generic drugs which were identified from
the literature. These attributes included generic drug equivalency,
effectiveness, and economic savings. All subjects were given verbal
information indicating that the generic drug was the same medication in
the same strength and potency as their brand name drug. Patients were
also told that the dispensing of the generic product would likely result
in approximately a one-third savings on their prescriptions.
The most important social (normative) influences identified from the
literature came from the physician and the pharmacist. Information
received from these sources were also controlled. Patients were told to
assume that their physician had given his consent on the prescription for
a generic drug substitute and that the pharmacist was willing to dispense
a less expensive generic drug with confidence if they chose to purchase
In addition to the factors above, a number of salient predisposi-
tional variables identified from the literature which were thought to be
associated with the willingness of patients to accept generic drugs were
evaluated for equal distribution among the look-alike and nonlook-alike
experimental groups. Specifically, the demographic variables of sex,
age, annual income, and highest education level were examined.
Additionally, other factors thought to impact on generic drug acceptance,
such as total prescription drug use, the pharmacologic class of the brand
name drug, the length of time the patient had been taking the brand name
drug, who pays the cost of the prescription, past discussions of generic
drugs with health practitioners, and past experiences in taking generic
drugs were evaluated for equal distribution among the experimental
The above text concludes the review of the literature. The next
chapter, Chapter IV, presents the methodology used to conduct the study.
A survey of 240 patients was conducted in six community pharmacies
in Gainesville, Florida. Patients in the study were identified as taking
at least one single-source, brand name prescription drug over the
extended period of time which was chosen for the investigation. The
survey was accomplished utilizing a structured interview/questionnaire
instrument designed and pretested by the investigator. It was divided
into three separate parts (Part I, Part II, and Part III) for
administration. The first patient survey was taken on November 9, 1981,
and the last survey was taken on May 20, 1982.
In Part I, patients were shown a sample of their single-source,
brand name drug product with its trademark side not visible. A fixed
form interview was conducted to uncover the manner in which patients
initially identified their drug product and to ascertain whether they
were knowledgeable of the drug product's trademark name, source, and
function. In addition, patients were asked what they thought the
trademark name of their drug product meant. This part of the study was
primarily designed to evaluate secondary meaning in prescription drug
product trade dress appearance.
In Part II, a controlled experiment was conducted to investigate
whether the trade dress colors of prescription drugs have significant
impact on patients' purchase choice between their brand name drugs and
less expensive generic drug equivalents. This part of the study was
designed to discover whether the trade dress colors of brand name drugs
are functional in respect to patients' acceptance of generic drug
Patients were randomly divided among two experimental conditions.
In the control group, patients were shown mock samples of "look-alike"
generic drugs which had the same size, shape, and color as their brand
name drug. In the treatment group, patients were shown mock samples of
"nonlook-alike" generic drugs which had the same size and shape as their
brand name drugs, but which differed in color. Patients in both groups
were initially informed that a less expensive generic version of their
brand name drug would soon be available on the market. In addition, the
patients received a brief explanation about generic drug substitution as
required under Florida's generic drug substitution law. The patients
were shown the mock samples and asked if they would want their pharmacist
to continue dispensing the brand name drug or if they would want to
switch to the less expensive generic drug equivalent. In addition,
several open-ended questions were asked in order to gain insight into the
patients' decision-making rationale.
In Part III, patients were asked to respond to questions posed in a
self-administered questionnaire. The data obtained from the question-
naire were used to describe the demographic characteristics of the
resulting patient sample, to evaluate certain predispositional factors
associated with patients' acceptance of generic drugs, and to compare the
distribution of these characteristics and factors between the two experi-
The Survey Population
The survey population consisted of community pharmacy patients, at
least 18 years of age, who were currently taking on a long-term (refill)
basis one of twenty-five single-source, brand name prescription drugs
selected for the investigation. Only those patients who paid out-of-
pocket in whole or in part for their medications were pertinent to this
investigation. Patients who did not pay for their medications (e.g.,
Medicaid recipients) were purposely excluded from the study because they
would have no motivating monetary incentive to switch to a less expensive
generic drug. In addition, these patients commonly have no real choice
among equivalent prescription drugs because regulations under government
insurance programs (e.g., MAC regulations) dictate that the pharmacist
dispense the lower cost product.
The drug products selected for investigation were identified in
respect to four pharmacologic drug categories for which long-term
medication therapy is usually indicated . These categories are (1)
cardiovascular drugs; (2) nonsteroidal anti-inflammatory drugs; (3) mild
tranquilizing drugs; and (4) gastrointestinal drugs.
The investigator identified all single-source, brand name drugs in
tablet or capsule dosage form which fell into one of the selected
pharmacologic categories and which was reported in the list of the "Top
200 Drug Products" for 1980 . Twenty-five single-source, brand name
drug entities were identified for potential investigation. A number of
the drug entities came in more than one dosage strength. Thus, the 25
drug entities represented a total of 54 different drug products.
Appendix A illustrates the 54 drug products identified for the study.
The single source status of each brand name drug was validated
through information obtained from the Food and Drug Administration's
"Approved Prescription Drug Products" publication . The study was
limited to single-source drugs in order to assure that patients had not
previously been offered a less expensive generic substitute for that
particular drug entity.
Selection of the Sample
In the initial phase of the study it was determined that a minimum
sample size of 238 patients (119 patients per experimental group) would
be necessary in order to find significance in the results. It was
decided that a total of 240 patients would be included in the study.
A judgmental sampling procedure was utilized to select patients for
the survey. Six community pharmacies in Gainesville were selected on the
basis of ownership classification and location in order to obtain a
reasonable cross-section of patients. Accordingly, three chain-owned
The sample size was determined in respect to the number of patients
required in each of the two groups in the research experiment. The
underlying assumptions of the sample size determination were (a) that
at least 80 percent of the patients would accept the generic drug; (b)
that a difference of ten or more percentage points between the two
experimental groups would be significant; and (c) that the level of
significance of the statistical test would be 0.05. See Fleiss, J.L.,
Statistical Methods for Rates and Proportions (New York: John Wiley &
Sons, 1973), p.194.
pharmacies and three independently owned pharmacies were identified as
desirable practice settings in which to conduct the survey. These
pharmacies were contacted and they agreed to participate in the study.
The pharmacies and their locations appear in Appendix B.
The sampling plan called for a quota sample of 40 patients per
pharmacy. To generate the sample population, the investigator spent
between eight and twelve consecutive work days at each pharmacy
identifying and interviewing patients as they appeared to have their
prescription drugs refilled. The pharmacist on duty informed the
investigator each time a patient was eligible for inclusion in the study.
This was accomplished by verbal signal or by the pharmacist's notation on
the patients' prescription bags. In some instances the investigator was
able to identify potential subjects by their prescription vials which
were brought into the pharmacy for refills. The specific criteria for
(1) the patient had to be at least 18 years old;
(2) the patient had to be taking at least one of the investigative
prescription drugs included in the study on a refill basis;2
(3) the patient had to have at least one month's experience in
taking the drug; and
(4) the patient had to pay for the drug, in whole or in part, out-
In those cases in which a patient was refilling more than one drug
product selected for the study, the first product which was identified
by the investigator or the pharmacist on duty was utilized in the
Patients who presented themselves in the pharmacies during the
investigation and who met the above criteria were approached and asked to
participate in the study. Patients were told that the investigator was a
pharmacist pursuing a graduate degree who would like to ask them some
questions about their medication therapy. Patients who agreed to
participate were asked for their signed consent in compliance with
University of Florida regulations which clearly stated the purpose of the
study. The consent form appears in Appendix C.
Prior to the start of the study, the investigator obtained three
samples of each drug product selected for the investigation. The various
tablets and capsules were affixed to cards of white cardboard (1 inches
square) and organized in an indexed expanding folder.
The first sample of each drug product was attached with the
trademark clearly visible. In Part II (the controlled experiment), these
products were to be shown to the patient alongside its mock generic
counterpart for visual comparison.
The second sample was attached with the trademark side down so as
not to be visible, therefore allowing only the drug product's size,
shape, and color appearance to be apparent. These products had a dual
purpose. In Part I, they were to be shown to patients for identifi-
cation; in Part II, they were to be utilized as the sample "look-alike"
The third sample was also attached with the trademark not visible.
However, these drug products were purposely altered to appear different,
solely in color, than the first and second sample drug products. They
were to be utilized as the sample "nonlook-alike" generic drugs in
Part II of the study.
The trade dress colors of the brand name drugs and the "nonlook-
alike" mock generic products utilized in the study appear in Appendix A.
To establish uniformity and control, the investigator recoated most of
the brand name products with a navy blue color. This shade of color is
frequently found in various drug products. In several instances, the
"nonlook-alike" mock generic products were created from existing
blue-colored drug products that had the same basic size and shape as the
brand name product. In those cases where the investigative brand name
products were already coated blue, the colors tan or yellow were utilized
to create the "nonlook-alike" generic product appearance.
Instrumentation and Pilot Study
During the initial phase of the study, a fixed form
interview/questionnaire instrument comprised of three parts was designed
to collect the data necessary to address the research objectives. The
instrument was developed so that each of its parts could be administered
separately. Part I incorporated the personal interview, Part II the
controlled experiment, and Part III the self-administered questionnaire.
The instrument was pretested in a two-day pilot study which began on
November 2, 1981. It was conducted at the Shands Teaching Hospital out-
patient pharmacy in Gainesville, Florida. A total of 12 patients who
For example, see the product identification section of a Physicians'
Desk Reference (PDR) for common drug product color coatings.
Physicians' Desk Reference (Oradell, New Jersey: Medical Economics
met the study criteria were surveyed. Each survey took between five and
ten minutes of the patient's time.
Following the pilot study, the interview/questionnaire instrument
was evaluated for question content and appropriateness of the response
categories. In addition, the instrument was corrected for potential bias
introduced by specific wording of the questions. Instructions to the
interviewer appearing in the instrument were rewritten for clarity to
assure consistent administration of the survey.
The pilot study was useful in that it familiarized the investigator
with the data collection process and provided him with experience in
conducting the survey with objectivity and consistency. The final form
of the interview/questionnaire instrument appears in Appendix D.
Variables of Interest and Data Collection Procedures
Part I-The Personal Interview
The specific variables to be investigated in this part of the study
were adopted from studies found in the marketing literature measuring
consumer awareness and knowledge of branded goods [38-45]. The variables
1. the patient's fundamental recognition of his drug product;
2. the patient's initial responses as to the identification of his
3. the patient's knowledge of the drug product's source after
4. the patient's knowledge of the drug product's source after
5. the patient's knowledge of the drug product's name after verbal
6. the patient's knowledge of the drug product's name after visual
7. the patient's knowledge of the function of his drug.
The investigator began each interview by showing the patient a
sample of his drug product which was attached to the sample card with its
trademark not visible. Next, the patient was handed the sample drug
product for close inspection. The investigator then asked the patient to
simply respond "yes" or "no" to the question of whether he recognized or
had seen the drug product before. The patient's answer was recorded by
checking the appropriate response category on the data collection
instrument. The drug product was then withdrawn from the patient's view.
A patient who responded positively to the first question was asked if he
could identify the drug product he was shown. One or more check marks
were entered in accordance with the patient's response and the categories
incorporated in the data collection instrument. If a patient had
initially given a negative response to the first question, he was told
the brand name of his medication. No further questions in Part I of the
survey were asked to that group; they were led to Part II of the study.
A patient who did not initially identify his drug product with the
manufacturer's name was asked by verbal probe if he knew what company
made the drug. If his response was negative or indefinite (e.g.,
incomprehensible), he was shown a prepared list of 22 drug manufacturers
which appeared in alphabetical order and asked if he could identify the
drug manufacturer. Appendix E illustrates the list of 22 drug
manufacturers incorporated in the visual prompt.
A patient who did not initially identify his drug product by its
trademark name was verbally probed as to whether he knew the name of his
drug product. If his response was negative or indefinite (e.g.,
incomprehensible), he was shown a prepared list of 48 trademark names of
drug products which appeared in alphabetical order and asked if he could
identify the name of his drug product. Appendix F illustrates the list
of 48 trademark names used as a visual prompt.
In addition to collecting information on the variables listed above,
a decision was made following the initiation of the research project to
examine a variable related to the patient's understanding as to the
meaning of the trademark name of the drug product. A patient who
identified the trademark name of his drug product either initially, under
verbal probing, or under visual prompting was asked if the name of his
drug, "XYZ," was a trademark name identifying a specific company's drug
product, or if it was the name of the drug itself. Since this question
was added after 120 patients had already been interviewed, it could only
be posed to a potential sample of 120 patients remaining to be surveyed.
This question was added to page one of the data collection instrument as
The final question included in Part I of the study was whether the
patient knew the functional reasons) he was taking the drug. Following
the patient's response, the investigator continued to Part II of the
investigation, the controlled experiment.
Part II--The Controlled Experiment
The single dependent variable to be investigated in this phase of
the study was the patient's purchase choice between his brand name drug
and a less expensive generically equivalent drug. The specific product
attribute of interest and independent variable to be investigated was the
color appearance ("look-alike" or "nonlook-alike") of the less expensive
generically equivalent drug product. In addition to ascertaining the
patient's choice, two open-ended exploratory questions were posed to all
patients in order to gain some insight into (1) their decision-making
rationale; and (2) their attitude towards generic drug color appearance.
These questions were
1. Why did you choose the generic (or brand name) drug?
2. Would it matter to you if the color of the generic drug was the
same as (or different) than the brand name?
The responses to these questions were recorded and subsequently
categorized upon completion of the survey.
Prior to the start of the study, the investigator randomly assigned
240 patients to one of two experimental condition groups of 120 each. A
random numbers table was utilized for patient group assignment . The
appropriate assignment group for each patient was marked on the data
In the control group, 120 patients were shown a mock "look-alike"
generic drug product (sample card #2) alongside the brand name drug they
were taking (sample card #1). In the treatment group, 120 patients were
shown a mock "nonlook-alike" generic drug product (sample card #3)
alongside the brand name drug they were taking. Patients in both groups
were initially told that there was a generic drug that would soon be
available on the market. In addition, the patients were also told that
the generic drug would be the same medication in the same strength and
potency as their brand name drug, but that it would be less expensive.
The patients were then shown the mock generic drug samples alongside
their brand name drugs. The investigator explained that the Florida
generic drug substitution law required the pharmacist to offer the
patient a less expensive generic version of their brand name drug if the
physician would allow a substitution and if the pharmacist stocked the
equivalent generic drug for substitution purposes . In addition it
was explained that the dispensing of the generic drug would likely result
in approximately a one-third savings to the patient on his prescription.5
This savings was also translated into approximate dollar amounts. After
this explanation, each patient was asked if he would want the pharmacist
to continue to dispense the brand name drug, or if he would want the
A number of patients questioned why the generic drug was less expensive
if it was the same as the brand they were taking. They were told that
the company which made the generic product did not have the promotional
and advertising expenses of the company which made their brand name
drug, nor the research expenses, and thus could sell their products at a
The savings attributed to a generic drug substitution were based on
actual market data collected by the University of Florida. See Vuturo,
G.J., Krischer, J.P. and McCormick, W.C., "Drug Product Selection: The
Florida Experience," American Journal of Public Health, 70:479, May,
pharmacist to dispense the less expensive generic drug.6-7 The patient's
response was recorded by checking the appropriate indicator space
(look-alike choice or nonlook-alike choice) on the data collection
instrument. Next, the investigator asked the two open-ended questions
referred to above. The patient's responses were recorded as given on the
data collection instrument.
At this point in the interview, the investigator fully explained the
reason for the experiment:
"The generic drug product samples you have just been shown are
not available on the market. They are fictitious samples of
patented drugs which we have designed to see if their appearance
would affect your response to whether you would switch to the
generic drug product. However, sometime in the future a real lower
cost generic equivalent drug should be available. But for now, it
is impossible for us to know whether the real generic will be made
to look like or not look like the brand name drug you are taking.
Your response may help us gain input in resolving this question."
In a small number of cases the patient asked why the "nonlook-alike"
generic drug was different in color appearance. These patients were
told that the drug product was made by a different company who used
their own color coatings.
About one-fourth of the patients stated that they would accept whatever
their pharmacist recommended. They were told to assume that their
pharmacist was confident about the generic substitution and had
indicated it was their choice.
Part Ill--The Self-administered Questionnaire
After the explanation above was given, the patients were asked to
complete the short questionnaire (Part III) by checking the one
appropriate response for each question. The investigator helped motivate
the patients to complete the questionnaire by explaining that their
responses would aid in evaluating the results of the survey.
The initial data collected in the questionnaire were utilized to
describe the demographic characteristics of the patient sample and to
determine whether these characteristics were equally distributed among
the two experimental groups. The salient demographic variables of
interest were identified from previously conducted studies regarding
generic drug acceptance [53-613. These variables were
1. sex category;
2. age category;
3. annual family income category; and
4. education category.
The specific demographic brackets incorporated in the instrument were
formulated from standardized categories used in studies examining
utilization and expenditures for prescribed medications [81-82].
8In a small number of the interviews the patients indicated they could
not respond to the self-administered questionnaire because they could
not read it. Most often the patients stated they did not have reading
glasses on hand. In these cases, the investigator read the
questionnaire out loud to the patients.
In addition to the above information, a number of predispositional
variables identified in the literature which were thought to be
associated with patients' acceptance of generic drug products were
collected in order to determine whether they were equally distributed
between the experimental groups. These variables were
1. the total number of prescription drugs taken by the patient on
a regular basis;
2. the length of time the patient had been taking the brand name
drug being investigated;
3. the patient's source of payment for the prescription drug;
4. the patient's recollection of any conversations with his
physician or pharmacist concerning generic drugs; and
5. the patient's recollection as to having taken a generic drug in
Two exploratory questions were asked in the questionnaire in order
to gain information regarding where patients obtained knowledge of the
drug product's brand name identity, identification marks, or source of
manufacture. The two variables of interest were
1. the patient's knowledge of whether the brand name of the drug
was recorded on the prescription label; and
2. the patient's knowledge of any identifying names, numbers,
marks or symbols appearing on the drug product itself.
The interview was terminated when the patient completed the
questionnaire. Patients were reminded that their responses and comments
would remain anonymous and they were thanked for their participation in
Data Processing and Preparation
The data collected on the interview/questionnaire instrument were
coded by the investigator, entered in blocks provided on the instrument,
key punched on a computer terminal, and transferred to computer disk
space. Responses to the open-ended questions on the instrument were
categorized and coded. The data coding sheets appear in Appendix G. The
investigator visually validated the coding, editing, and key punching
procedures. The data were electronically tabulated and analyzed in part
by utilizing the SAS package for data analysis .
Major Limitations of the Study
An extremely small number of patients who take brand name
prescription drugs were sampled. These patients were strictly defined by
the specific criteria required for inclusion in the investigation.
Patients who obtain refills for brand name prescription drugs in
Gainesville, Florida, may not be representative of all patients taking
various kinds of medications or representative of patients taking the
same medications in other cities and states. Consequently, caution must
be exercised in extending the findings beyond the sample patient
population examined in this study.
Given both the time constraints and the economic considerations,
every effort was made to design the study so that an unbiased,
representative number of appropriate patients could be surveyed.
However, the judgmental sampling procedures utilized to select patients
for the study do not guarantee a random or representative sample.
Therefore, the problems associated with selecting a nonrandom or biased
sample from the population of interest must be considered (e.g., the use
of statistics based on random samples).
The data collection instrument incorporated in this investigation
was original. While the instrument was pretested and evaluated for
content validity, it has no established norms or established reliability.
The controlled experiment part of the study was carried out in
hypothetical form. Since the knowledge levels, attitudes, intentions,
and behavior of people change over time, there is no guarantee that
patients would respond in the same way in an actual situation.
The personal interview is an interactive process in which the
background characteristics, perceptions, and reactions of both the
interviewer and interviewee are important determinants affecting the
subject's response. Every interview in this study was conducted by the
sole investigator in order to eliminate bias resulting from the use of
multiple interviewers. However, bias could still have been introduced by
the questions themselves, by the manner in which the questions were
asked, by probing the respondents, by motivating the respondents, or by
improperly recording the responses. Any bias which was unknowingly
introduced by the investigator should have remained consistent over the
RESULTS AND ANALYSES
The Findings of the Study
This chapter presents the results and the statistical analyses for
Part I, Part II, and Part III of the study in respect to the research
objectives and the specific research hypotheses outlined in Chapter I.
A total of 263 patients who met the study criteria were contacted by
the investigator. Twenty-three patients declined to be interviewed. The
interview/contact response rate was equal to 0.91 (91 percent). Most of
the patients who refused to be interviewed stated that they were rushed
or short of time. Only a few patients indicated that they did not want
to talk about their medications or their medication therapy. No hard
data, however, were collected in regard to patient refusals.
Part I--The Personal Interview
In regard to the first question, Table I shows that 232 (96.7
percent) of the 240 patient-respondents indicated that they recognized or
had previously seen the drug product which was shown to them. Eight
patients (3.3 percent) responded negatively to the recognition question
even though they were currently taking the drug. Although patients were
asked to respond simply "yes" or "no," approximately one-third of the
patients responded by saying "I believe so," or "it looks familiar," etc.
Hypothesis 1: More than 50 percent of the patients initially identify
their drug product by its trademark name or source of manufacture in a
Patients' Initial Recognition
of Their Prescription Drug Products
(yes or no)
"Do you recognize this drug? Have you seen it before?"
aThese patients were informed of the correct identification and
directed to Part II of the study.
In response to the second question, Table II indicates the manner in
which the patients initially identified their drug product. The "Z" test
statistic was utilized to test the first hypothesis for acceptance at a
confidence level of 95 percent. For a one-tailed test, this means that
a computed "Z" value greater than 1.65 would indicate that the research
hypothesis is accepted with 95 percent confidence.
Data from Table II show that 154 patients (64.2 percent) identified
their drug by giving only the trademark name. Accordingly, the computed
"Z" value was calculated to be 4.39. This means that Hypothesis 1 above
is accepted with 95 percent confidence. An important finding is that all
patients in the sample who identified their drug product in a trademark
sense did so by giving the trademark name of the medication. A total of
168 patients (70.0 percent) initially identified their drug by its
trademark name alone or in combination with function. No patients
initially identified their drug by giving the source of manufacture.
Twenty-eight patients (11.7 percent) were initially uncertain in their
response to the identification question. One patient (0.4 percent)
incorrectly identified the drug product.
Hypothesis 2: Less than 50 percent of the patients initially identify
their drug product by its colors, functions, or effects without regard to
The "Z" statistic is the appropriate statistical test assuming the
normal approximation to the binomial distribution. This assumption is
valid when n(p) and n(l-p) are greater than 10. For all hypothesis
testing in this study, the significance level associated with the
statistical test will be specified at 0.05. That is, the probability of
accepting the research hypothesis when it is false will be specified at
alpha equal to 5 percent. Accordingly, the confidence level of the test
(1-alpha), the probability of not accepting the research hypothesis when
it is false, is specified at a level of 0.95 or 95 percent. See
Churchill, G.A., Marketing Research Methodological Foundations, 2nd ed.
(Hinsdale, Illinois: The Dryden Press, 1979), pp.435-443.
Patients' Initial Manner of Identifying Their
Branded Prescription Drug Products
Manner of Initial
"Can you identify for me
the drug product you were
aTrademark identification was equated with patient's responses)
in terms of the trade name or name of manufacturer.
identification by function was equated with patients' responses)
in terms of color or function/effect, e.g., "my red heart pill."
cOne patient incorrectly identified NAPROSYN as PERCODAN. The
patient was informed of the correct identification and led to
Part II of the study.
The data show that of the 240 respondents, 35 (14.6 percent)
patients initially identified their drug product solely by function. A
computed "Z" value of 10.96 indicates that Hypothesis 2 is acceptable
with 95 percent confidence. An examination of Table II reveals that 14
(5.8 percent) patients initially identified their drug product both in a
trademark and functional sense. If these patients are added to the
number of patients who initially identify their drug product solely by
function, a total of 49 (20.4 percent) patients were found to initially
identify their drug in some functional manner. However, a calculated "Z"
value of 9.16 still indicates that Hypothesis 2 can be accepted with 95
Hypothesis 3: There is a significant difference between the percentage
of patients who initially identify their drug product in a trademark
sense and the percentage of patients who initially identify their drug
product by its colors, functions, or effects without regard to trademark.
Table II also contains 95 percent confidence intervals constructed
for each of the percentage estimates of patient responses to the second
interview question. Statistically, a 95 percent confidence interval
indicates that in a repeated process of drawing samples from a population
and computing percentage estimates and confidence intervals for a
particular parameter, 95 percent of the intervals so formed contain the
"true" percentage in the total population. Information from Table II
reveals that the true percentage of patients who initially identify their
drug product in a trademark sense would lie in the interval 58.1 to 70.3
percent, and the true percentage of patients who initially identify their
For a complete discussion of calculating and using confidence intervals,
see Ott, L., An Introduction to Statistical Methods and Data Analysis
(Belmont, California: Duxbury Press, 1977) pp.74-79, 278-80.
drug product soley in a functional sense would lie in the interval 10.1
to 19.1 percent. The percentage estimate measuring the difference
between the number of patients who identify solely by function and the
number of patients who identify solely by trademark is 49.6 percent. A
95 percent confidence interval indicates that the true percentage
difference in the population would lie in the interval 40.3 to 58.9
percent. Since this interval does not contain "zero" percent, Hypo-
thesis 3 can be accepted with 95 percent confidence.
As mentioned above, a total of 49 (20.4 percent) patients are found
to make identification of their medication solely by function or by
function and trademark. Therefore, a 95 percent confidence interval
constructed to measure the true percentage difference between the number
of patients who identify in some functional manner and the number of
patients who identify solely by trademark is equal to 36.6 54.0 per-
cent. Since this interval does not contain "zero" percent, Hypothesis 3
can still be accepted with 95 percent confidence.
Hypothesis 4: More than 50 percent of the patients who do not initially
identify their drug product by its trademark name are knowledgeable of
its trademark name after verbal probing and visual prompting.
Table III shows the patients' knowledge of their drug product's name
initially, after verbal probing, and after assistance under visual promp-
ting. Table IV shows the patients' knowledge of their drug product's
name only after probing and prompting. Of the 232 patients who recog-
nized their drug product, 168 (72.4 percent) initially identified the
drug by its trademark name. One patient (0.4 percent) misidentified the
product. The remaining 63 (27.2 percent) patients were either verbally
probed or visually prompted to give the name of their drug product.
Patients' Knowledge of Their Drug Product's Name
Initially, After Verbal Probing, and After Visual Prompting
(n = 232)a
Initially gave correct name
Gave correct name under verbal probing
Gave correct name under visual prompting
Could not identify name
Initially incorrectly identified named
aEight patients out of 240 total did not initially recognize their
drug product and were led to Part II of the study.
b"Can you tell me what is the name of the drug product?"
cThe visual prompt was a listing of 48 drug products in alphabetical
order grouped in a series of 4 trademark names (see Appendix F).
dThis patient was given the correct name of the drug product and
led to Part II of the study.
Patients' Knowledge of Their Drug Product's Name
After Verbal Probing and Visual Prompting
(n = 63)a
Response Number Percent
Gave correct name under verbal probing
Gave correct name under visual prompting
Could not identify name
aThis number represents those patients who recognized their drug
product but who did not initially identify the drug by its trademark
The data in Table IV indicate that 46 (73.0 percent) of the patients
who were verbally probed and visually prompted gave the correct name of
their drug product. A total of 17 (27.0 percent) patients could not
identify their drug product's name. A computed "Z" value of 3.65
indicates that Hypothesis 4 is acceptable with 95 percent confidence. In
this study, a total of 214 (92.9 percent) patients who initially
recognized their drug product were knowledgeable of their drug product's
Hypothesis 5: More than 50 percent of the patients who do not initially
identify their drug product by its source of manufacture are
knowledgeable of its source of manufacture after verbal probing and
Table V shows the patients' knowledge of their drug product's source
after verbal probing and after assistance under visual prompting. No
patient who initially recognized his drug product identified it by the
source of manufacture. Thus, all of these patients were given verbal
probing or visual prompting as to source.
The results indicate that only 17 (7.4 percent) patients were
knowledgeable of their drug product's source after verbal probing and/or
visual prompting. Moreover, 198 (85.7 percent) patients indicated that
they could not identify the drug product's source. Hypothesis 5 was not
accepted at a 95 percent confidence level. The hypothesis was tested for
rejection at the 95 percent confidence level, and a calculated "Z" value
of 12.94 indicates that the hypothesis is rejectable with 95 percent
confidence. A total of 16 (6.9 percent) patients incorrectly identified
their drug product's source.
Patients' Knowledge of Their Drug Product's Source
After Verbal Probing and Visual Prompting
(n = 231)a
Knew the source of manufacture after
Knew the source of manufacture after
Could not identify source
Incorrectly identified source
aEight patients out of 240 total did not initially recognize their
drug product and one patient initially incorrectly identified their
drug. These patients were led to Part II of the study.
b"Can you tell me what company makes the drug?"
cThe visual prompt was a listing of 22 drug manufacturers in alpha-
betical order (see Appendix E).
Hypothesis 6: Less than 50 percent of the patients who are knowledgeable
of their drug product's trademark name associate that name with the
trademark of a specific company's product.
Hypothesis 7: More than 50 percent of the patients who are knowledgeable
of their drug product's trademark name associate that name with the drug
entity itself without regard to the trademark of a specific company's
Since the interview question relative to Hypotheses 6 and 7 was
incorporated into the study after half the patient sample had been
surveyed, the maximum number of subjects that could be questioned was
120. Thus, all patients in the second half of the patient sample who
were knowledgeable of their drug product's trademark name initially,
after verbal probing, or after visual prompting were asked the interview
Table VI indicates that only 25 (22.9 percent) of the patients who
were knowledgeable of their drug product's name thought the name
identified a trademark of a specific company's product. A computed "Z"
value of 5.66 indicates that Hypothesis 6 is acceptable with 95 percent
Table VI also shows that 62 (56.9 percent) patients thought the
trademark name identified the name of the drug itself. Twenty-two (20.2
percent) patients responded that they "did not know." A computed "Z"
value of 1.44 indicates that there is insufficient evidence on which to
accept Hypothesis 7 at a confidence level of 95 percent. However, the
data indicate that a significant majority or 77.1 percent of patients
(69.2 85.0 percent with 95 percent confidence) did not know or think
that the name of their medication was a trademark name of a specific
company's drug product.
I I I 4J
4-I I. 0 I. >r-4-
in a) H3 u'
4-i a4 0 03
a 0. bo
00 0 0 $4
t-. 'o o-
w c0 a *
co 14 C 4 Ln N b0&
>U w r-4 A 0 caB
"0 04 0-l =
r. >0 II. 60
Ca) c4- a)
0U 4- 4-
H 0 C%4 0H 0.
>u o. C4O
> 0 U B a
0 0 z to 0
(o al M o a )
*. 4 0 0W 4
CO 4O 0.
pp co "a 00
U a *X41 $4 s 01C
**8 0 4- 4- N *H
COM cd iO O i- ( 00
H C- 4 .J 0. 41
3- tf 4-4 .o m54 |45-s
0 -H o4 *0 4 O
U C 0 a) 0 a
Ca a o ,e 4- C 5-
C 01 -0 0 w 00 01-
5- P4C ^1 -U 4-JO)-
Pcs C~0 a) $ .C .4.5
444 3 a0 C0 0-
P4HWJ 44 c
U w4 l Ue 4a
(3- c3 0 441 N a)
u ~ CO 00 cr >
P "o co XU
SP4 co~ AjO~a0 1-
4 OO W QO o
4M ,C U C 0 *C *r T3
Ua -W R > H
4a -14 L 0 *
54 C a-i *- -H a) 4 0
CLo tw co wu Q) 44 o4J c!i-
H 3 4- w-14oI 4.-H 4J M
t8 4J- 4J U1H 1C
-Y ll.4- co m) M)4i4 )*- -
.7-4 4J 0 C 0 v-0
ca w : ca41 z-4
(U Q) 10i- l) 0 Cd "ar>
)-1 d 41 w ^ wO 4d C w 0 U w
aa CH 4I 3 H w^u
B f LW 41 4J A- 10 J -i
T)0 43T -H 44 r. J (00
H w 441 4W 4-4 J~ 4J *0
M 3 C 3 0 W e
wU 0 w 0 03 wU 3 4-J"
Ja! 10 4! w) i-H o -1 :1 >1
41 l-H -W a y ca 0 f-i
W u 4c c a ca) r
-Hi Q 3 *H^ 1<)
4J 4J 4J 4J 4-J 0 t-I G
ca m) co wdi-
P0i 44| P4 co
Hypothesis 8: There is a significant difference between the percentage
of patients who associate the trademark name with the trademark of a
specific company's product and the percentage of patients who associate
the trademark name with the name of the drug entity itself.
Table VI gives the 95 percent confidence intervals constructed for
the corresponding percentage estimates of the patient responses. The
estimated percentage difference between patients who associate the
trademark name with the drug entity and patients who associate the
trademark name with a specific company's drug product is 34.0 percent. A
95 percent confidence interval indicates that the true difference in the
population lies in the interval 18.6 to 49.4 percent. Since this
interval does not contain "zero" percent, Hypothesis 8 is acceptable with
95 percent confidence.
Table VII shows the responses to the question of whether patients
were knowledgeable of the reason why they were taking the drug. This
question was incorporated in order to gain information about patient's
knowledge of the function of their drugs. Although a specific hypothesis
was not formulated, the results indicate that 223 or 96.5 percent of the
patients (94.2--98.8 percent with 95 percent confidence) did know the
functional reason why they were taking their drug.
PART II--The Controlled Experiment
Table VIII depicts the choices made by patients between their brand
name drugs and the samples of less expensive generically equivalent
look-alike and nonlook-alike drug substitutes. Information from the
table was utilized to test Hypothesis 9.
Patients' Knowledge of the Reason(s)
Why They Take Their Drugs
(n = 231)
Patient knows 2
Patient does not know
"Can you tell me the reason why you take this drug?
l 0 0 0
4-J C'.1 Ce|
000 & o
-4 0 Q)
a) w 4 Cd0
Uw 4J C 0$0 0
E .4 c i C
04 0 -1
A 1 U ) 0i)
4. Qi) 41 r i3
r.ca I 0 $.4 tz
4 .0 0 0 0
,C *). 0
(U 4-1'0 -i
U) (3 Co
4 0 E!*
HO ( 1 -' 0.) C1
(L 4J 00 00 0 0.
H Q) 0J4 1 0
E-4 4. L4pa 4i- 1u
u 0 o
>-,4- & o ^3 r~ w o m
.Q CO J -| c -
-4W W) -i 3- Ca i C) 0
z co 9) 0 0 m 4
^M ,C -M O
w Q 03 ( w-
0 C) 0
5 ~ ~ ~ 4 C)c s a 4i ~ ^ ^ s'
S4 4) 4 -
0 u 0
^~O TI ( (U4 S O O C .*
W 0 .-4 0
go 0 5
Q ft C 1).4$-.
a cn~c a o o o :
44 4J Q4 W1Q
0-A- ,-4 -14 0
0U) .,A 0u
z cu .1
rn -14 0 03
pH 0 00
u -4 4
uS4 -H Ma 0-
Cd .4 1 0
) w r-4 Q
l0 0 0
a0) 0 0 E-4U
Hypothesis 9: There is a significant difference between the percentage
of patients who choose "look-alike" generically equivalent drug products
and the percentage of patients who choose "nonlook-alike" generically
equivalent drug products when patients are offered a less expensive
generic drug substitute in place of their brand name drug.
Data from Table VIII reveals that 103 (85.8 percent) patients of a
total of 120 patients assigned to the "look-alike" group accepted the
generic drug substitute, while 101 (84.2 percent) patients of a total of
120 patients assigned to the "nonlook-alike" group accepted the generic
drug substitute. The data indicate that 1.6 percent more patients chose
the look-alike generic drug substitute. The "Z" statistic was utilized
to test for a significant difference between the percentages of patients
in each experimental condition who chose the generic drug product. For
a two-tailed test, a calculated "Z" value greater than 1.96 indicates
acceptability of Hypothesis 9 with 95 percent confidence. The "Z" value
was calculated to be 0.35. Thus, the statistical test indicates there is
insufficient evidence upon which to accept the research hypothesis that
the difference between the percentage of patients who chose look-alike
generic drugs and the percentage of patients who chose nonlook-alike
generic drugs is statistically significant.
Tables IX and X show the reasons given by the patients for choosing
the generic drug substitute or for remaining with their brand name drug.
In regard to the 204 (85.0 percent) patients who chose the generic drug,
200 (98.0 percent) patients indicated the reason they chose the generic
The "Z" statistic is the appropriate test under the assumption of
(a) the normal approximation to the binomial distribution and
(b) independent, random samples. The Z- test incorporates a pooled
sample variance. See McClave, J.T. and Dietrich, F.H., STATISTICS
(San Francisco: Dellen Publishing Company, 1979), pp.310-14.
Patients' Reasons for Choosing the Generic Drug
(n = 204)
Cheaper or because of savings
Previous (good) experiences with
"Why did you choose the generic drug?"
Patients' Reasons for Choosing the Brand Name Drug
(n = 36)
Reason Number Percent
Want only what their physician
ordered 21 58.3%
Prefer the brand name drug 12 33.3%
Previous (bad) experiences with
generic drugs 3 8.4%
Total 36 100.0%
"Why did you choose the brand name drug?"
drug was because of the savings involved in the substitution. The
remaining four (2.0 percent) patients indicated they chose the generic
drug because of previous good experiences with generic drug substitutes.
In regard to the 36 (15.0 percent) patients who chose to stay with their
brand name drug, 21 (58.3 percent) patients indicated that they wanted
only what their physicians had originally ordered. Twelve (33.3 percent)
patients indicated that they preferred the brand name drug, and 3 (8.4
percent) patients indicated they previously had bad experiences with
generic drug products. No patients who chose to stay with their brand
name drug indicated that the color or "looks" of the generic drug
affected their choice.
Table XI shows the responses pertaining to the open-ended question
of whether the trade dress colors of generic drugs mattered to the
patients. Examination of the table reveals that 136 (56.7 percent)
patients indicated that the colors of the generic drug did not matter, 20
(8.3 percent) patients indicated that the colors did matter or were of
concern, and 84 (35.0 percent) patients indicated that the colors did not
matter as long as they were informed, in some manner, that the generic
product being dispensed was the same drug as their brand name product.
Of the 20 patients who indicated that color was of concern, 11 patients
chose the brand name drug. Six of the 11 patients were shown a
look-alike generic drug and five were shown a nonlook-alike generic drug.
In regard to the remaining nine patients who chose the generic drug and
indicated their concern about color, six patients were shown a look-alike
generic product and three patients were shown a nonlook-alike generic
product. The additional data appears to support the finding that trade
dress colors of generic drug products do not significantly impact on
4-1 I ~ &S f-
3 l- 0 0
w 0 Ln 0(
0 -4 Q-
4-1 [0 1
0 o C 0
C0 0 n '.0 -T~a p4
iU4J 0 On
020 a) 10 f..4L(
4 -40 0 0 C8
0 a0) w w S '
o- U)$4$4 (V.4~ 4 ) $
0 rOU O 4 r U H
u 0 a)
olo w >
z 00 -H
QM S l
( 0 w 4J 4
En r.- r0!?-.
Q) 4) ca w -'
0 ca na
40 0 r-4 N0
ca 4 3 0 Y 1
E 4J 41 C z :j :3- 4
a) 02 0 40
P4 o w )04 p p
-: 4- 0 04J 4C 4J 0
F-4 ( u 0 0
-Ha 4 41 U U U 4 W0
cc 0Q 4 -H u 4)i
tt1 ol Cl) a) $4 -l6 a
02 0 P., ,-4
'H r0 0a
a) 0$ u P4 0$4
m --H Q 0 0
p 40 0 0
S4( U a4 a
00 m4 Q3 0) :3 '4-1 00 0
0 0 .
4JC1 0. -4 -4 -4 ^
-'4 W 00 0 0
p^O O pc bO 0
4-- p 1
-J 0 c
401 41 41T3M
c2- a) 0 0
0 w 4.JM 0 0
-, W 0& r1 M 4 e
9 .u o 0o O & Uf J O )
4 4.) u S a c 0 -4
4- CO 4 C00 0) 44 C OA
a0 CO 4J CO 0 O M
-^C' (U 3U ( J t U ~ 0'
41. 41 4- 1 i- H W) Ca a) a) C
44 4- 4 4J E 0 0
C:i p0 9z 02 o q () 0 1z
0 a) 4 WM J a) t3 E a &"
o- $4 $4- 8 4 a n e
4 3 H g 0 M a
0) 4J0 4 4-1 a 4 o c
C! Si C 0 (3 O COl T 0)0 34J 3
r. 0H 0d 40. O 3 4 02 0.
03 4 U l 4- J '0 4J 1X 3C y S
0 "o 0. 1~0 C a, m4 .)-)
o u 00 bo 0 u
0 O a) Os-4 a) O H
0-4 01a0) a) a)4 w 01 0 w 0
-4 0 -4 4-1 r-4 4-J 1-' 4J "0 U w 0)
00 0440 14-C '! C
U. U.4- 4- 0.I $40 $40(
0 0 0 0 0
En o O$
patients' choice among their brand name drugs and less expensive
generically equivalent drug substitutes.
PART Ill--The Self-administered Questionnaire
The self-administered questionnaire was utilized to collect descrip-
tive information concerning the patient population. Data collected in
the instrument were used to compare the equivalence of the experimental
groups in respect to demographic characteristics and other specific
variables related to patients' knowledge of their prescription drugs.
Table XII shows the demographic categories of sex, age, income, and
educational levels of patients in the sample population. Of 240 patients
interviewed, 92 (38.3 percent) were male and 148 (61.7 percent) were
female. Eighty-two (34.2 percent) patients indicated they were 64 years
of age or older, and a total of 191 (79.6 percent) patients in the sample
indicated they were at least 45 years of age.
Fifty-three (22.6 percent) patients indicated that annual family
income was less than $10,000. Slightly less than half of the patients
(110 or 46.8 percent) indicated their family income fell into the
$10,000-$24,999 per year range. Thus, approximately 70 percent of the
patients indicated that their family income was less than $25,000 per
year. Nineteen (8.0 percent) patients indicated their annual family
income was $50,000 or higher. Five patients did not respond to the
In regard to the patients' highest level of education, 118 (49.2
percent) patients responded that they had received a high school diploma.
However, 41 (17.7 percent) patients indicated that their highest level of
*** .n CM
CM N -s-
CM i-< o ys-
Cen tn CM4
CM 0 NM 00
Le 1-1 Ln r-I
4 *4J CDCD
S0 CD a'
H 0r- 40 V
0 C a' I 0
r-i 4 o o
O 1i-1 .cn
r- CM Jo
Lf0 < 0
r- n C
C) Ln i-I
CM %0 CM P-4
r-4 M i CM4
r- C% CM r-i o
r-i It CM4rq 4
r1 0 1-
4 0 p
w- 0. 0 w
CU O1H ~O)
W 0-0 I
w) 0 0
H0 0 0
.4J *X -3 .-1 4J1
Of (Od 00ilt
a 'n II
* *j j
00 iln o -
r-i in IT
r-q m rl o
CMina CM r-.
r- n o o
CM4 in CVn r-4
0 r>M C
1') 1a0 -T
C14 o n1
en .D oT
-1 Ln (7%
r-1 Cn o
,I n -'T
b* b"* b*, I*
V) M14 M
education was grade school. A total of 81 (33.7 percent) patients
indicated they were college graduates, and 27 (11.2 percent) of these
patients indicated they had received graduate degrees.
In summary, it appears that the sample patient population was skewed
towards middle aged and elderly females, with low to middle family
incomes, and low levels of education. However, previous studies
conducted on prescription drug utilization indicate that the subjects
were not atypical of patients who take prescription medication [81-82].
Table XII also presents a breakdown of the respondents' demographic
characteristics in each of the look-alike and nonlook-alike experimental
groups. The chi square test was utilized to determine whether any
particular patient characteristic was dependent on the experimental
groups. In this investigation, a chi square value with a corresponding
"P" value of 0.05 or less is considered statistically significant and
would indicate with 95 percent confidence that the two experimental
groups are not equal in respect to that particular patient character-
istic. An examination of Table XII reveals that the calculated chi
square values were not sufficiently significant to conclude dependence
existed among the patients' demographic characteristics and the
The "p" value associated with the chi square test indicates the
probability of exceeding the magnitude of the chi square statistic for a
given number of degrees of freedom. If this probability is small, the
likelihood that the disparities between the observed frequencies and the
actual frequencies are due to chance is small. In this study, a
particular cross-classification will be deemed dependent if the
probability of exceeding the chi square statistic is less than or equal
to 0.05. See Akrin, R. and Caltow, R.R., STATISTICAL METHODS (New
York: Harper and Row, 1970), pp.131-35.
Table XIII presents the patients' drug products which were
investigated and their distribution within the experimental groups. An
examination of the table reveals that the most frequently investigated
drug product was INDERAL (35 or 14.6 percent), followed by DYAZIDE (25 or
10.4 percent), VALIUM (22 or 9.2 percent), and TAGAMET (20 or 8.3
percent). The drugs investigated the least number of times were NALFON,
NORPACE, and TOLECTIN (each having only two observations). Two drugs
originally selected for the survey, ATROMID-S and ZAROXOLYN, were not
investigated during the study period since no patients requested refills
Table XIV shows the distribution of drug products within
pharmacologic class and by experimental groups. A total of 134 (55.8
percent) drug products were in the cardiovascular category. Mild
tranquilizers and nonsteroidal anti-inflammatory drugs constituted 18.8
and 17.1 percent, respectively. Gastrointestinal drugs accounted for
only 8.3 percent. A chi square value of 2.322 (p = 0.5064) indicates
that the pharmacologic class of the drug products investigated are not
dependent on the experimental groups.
Data from Table XV show the total number of prescription drugs which
the patients indicated they were taking on a regular basis. A chi square
value of 2.846 (p = 0.4159) reveals that the number of prescription drugs
the patients took on a regular basis is not dependent on the experimental
groups. A total of 103 patients, or approximately 43 percent, indicated
they were taking three or more prescription drugs on a regular basis.
Seventy-two patients (30.0 percent) indicated they were regularly taking
only one drug.
V) r-4 -1
* No 8N* o
It 1.0 -4 cli 1-4
N* 0 O
0 Cu 0LrC4
mn m m "M r- '-IT m s-I't --IT "M r-.4 r~ r-4 T r-4 3-T n cn r-4 cn cM4
r-I r-4 r- i
T- (n r-i r--i ,o r-i T i -i -T cn m T q -i NM -i cM Ln r- -i mo
r-I r-4 CM4 r 4
'-4 l -4
cc 00 q0 1
H H a -
0>> < 1- c
iUl i-4 1- 4.1
>-I > 4 0m 0)
>~ ~ ~ O 0 *w w 4-H
> > U Z > > > 0 H 0
SU ...U U U Z E-4 0 i H
EMMAOM i-Hl HOO I O O ^-Z 2 "' S ^'1- 3.I
E-4 P4] 14 r Za z 1:4 rano! / >4i wd' E-4 1- Z 0g 0 i-H c
QQM1 -lM4QQQcna ZE 0f4iSP4OZ E-4
H tH-4 >4 Z Z > < E- -
<< u u =l-4w H 04 04 1-4 z. z ZZ 6 H H H > u Z z 0
S ITmMneMn nWonoi nr~- C.0oc"s0OOoc
r-l i- C4 N- cM i C4
oo on T
S00 c 0
S o9 **
0 00 N- 00 0
U L -4 ,-i 0
0n p 0o e
O 04 44 0
0 U r1i
E 0 :1 r *
o0 "0 0
U 4 0 p
C-4 >,Q C'4 U
p Mo r C1 411
0 C 0 .0 II
T4 -4 0
0 9 *0
"a*1 1 -40
02 1-1 4
0 3 cc m E-
M (0 QI
m2 3 0 &
0 -H 44 00
00 10 1w 4
0 C 4 4 ih
Q n rr.* M m 0
Cu 0l 1- 000
ca -H 0 Cu
u z 0
C)O V 01 N cc
u 0 r'*- 0 oW 0
14 CI CM4 CM4 1- 0
0 N e4
Sr- M 0
ci .-d4, r-0 0Q 0 -
00~0 00 e-
C i-4 (z
C fi CO O- BS t ~ M fl C
3 *r 0is
U.s 0 0
4 .4 0
i- CO 0 0
z -H *r P- Co 0
M 1 1 *fC
0f P6 03- a
^C O C
w C o
O ;4 A
a) Q) 0)
0 0 0
r 4 0
0 a E-4 17
Table XVI shows the length of time patients had been taking their
prescription drug. Two-thirds of the patients (160 or 66.7 percent)
indicated they had been taking the drug for more than one year. Of the
remaining 80 patients, 45 (18.7 percent) indicated they had been taking
the drug less than six months, and 35 (14.6 percent) indicated they had
been taking the drug between six months and one year. A calculated chi
square value of 1.822 (p = 0.4021) indicates no dependence on the length
of time the patients had been taking their drug and the experimental
Data from Table XVII show the manner in which patients in the study
normally paid for their prescription drugs. Approximately 79 percent of
the patients (189 or 78.8 percent) indicated they paid for their drug
products entirely out of pocket. The remaining patients indicated that
insurance paid a portion of their prescription expenses. A chi square
value of 0.025 (p = 0.8746) indicates there is no dependence regarding
the manner in which patients pay for their prescriptions and the
Table XVIII is a summary of answers to the question investigating
whether patients had previously talked with their pharmacist or physician
about taking a generic drug. Less than half of the patients (112 or 46.7
percent) indicated they had discussed generic drugs with their pharmacist
or physician at any time. A chi square value of 1.674 (p = 0.1957)
reveals no dependence on whether patients had previous conversations with
their pharmacists or physicians about generic drugs and the experimental
Patients were also asked if they had knowledge of taking a generic
drug in the past. The results are depicted in Table XIX. Slightly more
C rN -0 r-
S-I -4 '.
30 L J D
0 C14 -1t 0
C14 C14 r
Z0 'n -4 '0
c -4 cr
Q) Ca r-I C/
f a 00O f
(X SO C
S 1.0 z
r5 < >
-* 4 0 C '
to in r-i
\D > 1*
(~l ~ r)
than 50 percent of the patients (121 or 50.4 percent) indicated they had
taken a generic drug in the past. Ninety (37.5 percent) patients
indicated they had not taken a generic drug, and 29 (12.1 percent)
patients reported "Don't know." A chi square value of 1.069 (p = 0.5861)
indicates that there is no dependence on whether patients had taken a
generic drug and the experimental groups.
Data from Table XX show the responses to the first exploratory
question pertaining to whether the patients had knowledge of any identi-
fying names, numbers, symbols, or marks imprinted on the drug product
itself. Forty-seven (19.6 percent) patients indicated on the
questionnaire the presence, in part or in whole, of some identification
imprinted on their drug product. Eighty-four (35.0 percent) patients
indicated that they were aware of some identification on the product, but
could not recall the specific imprint. Even though all of the drug
products included in the study had some kind of identifying imprint, 38
(15.8 percent) patients indicated "no" to the question. A total of 109
(45.5 percent) patients responded "no" or "don't know." A chi square
value of 3.725 (p = 0.2927) indicated there is no dependence regarding
patients' knowledge of any identification imprinted on their drug product
and the experimental groups.
Table XXI illustrates the results to the second exploratory question
investigating whether the patients knew if the brand name of their drug
product appeared on the prescription container label. In all of the 240
refill prescriptions studied, the pharmacist had typed the brand name of
the patient's drug product on the prescription label. The table shows
that 197 (82.4 percent) patients answered "yes" and 30 (12.6 percent)
patients answered "don't know" to the interview question. Twelve (5.0
fl, -4- 00 -11
-T 00 cn r-
ON -T m c
(NI -3 C m
co o m H
-N -4 M
4.J 6^ 6- -b B
C -:1- 0 <0 0
U eML L C 0
p 00 C O
0S 0T r 41 cn r
CO & 0
g o cd co wr r
3 e-- r l
.Z 0 0 ,
4-4 Wo 0
Q) P4 Wn
"0 Q) 0
3 0 0 0
0 wr to 14
41 U 00
cdH O r
41 co. a4 oyD o a
u. f mn 014 M
i-4 O\C'-i- *i
U ~* I3
00 fc 0 1
0 0 r
4- 0 r '> 0 E-4 0- 0
a) 1-1rt 3 -i a)
i-i ni z to
41 0i 0 0
rt Z- o
percent) patients responded "no" even though their drug product
containers were labeled with the brand name. A chi square value of 3.942
(p = 0.1392) indicates there is no dependence regarding the patients'
knowledge of whether the brand name appeared on the prescription label
and the experimental groups. The patient responses to this question,
however, must be evaluated with caution. First, patients were told
earlier in the interview that the name of their medication was, in fact,
a trademark (brand) name. Second, several patients who had answered "no"
to this question later explained to the investigator that they thought
the wording "brand name" referred to the manufacturer's name and not the
trademark name of their drug. Consequently, the validity of this
question as worded in the questionnaire is in doubt.
These findings complete the results chapter. The next chapter,
Chapter VI, presents a summary and a discussion of the results of the
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID EARDEN95X_YFWEGS INGEST_TIME 2012-03-02T22:27:07Z PACKAGE AA00009115_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC