Direct-to-Consumer Advertising of Drugs and Patients' Health Care Seeking Behaviors

Direct-to-Consumer Advertising of Drugs and

Patients' Health Care Seeking Behaviors

Chapter 1: Introduction to the Study

Background

The number of people taking prescription medications is on the rise.

In 1993, the average number of prescriptions per person annually was seven,

compared to 11 in 2000, and 12.1 in 2011, with West Virginia reaching a

high of 19.3 (National Center for Health Statistics, 2014). In the United

States alone, the total annual retail sales for prescription drugs filled in 2011

was $228 billion (National Center for Health Statistics, 2014). Over half of

all people in the United States take a minimum of one prescription

medication daily, on average (National Center for Health Statistics, 2014).

Additionally, Pharmaceutical Research and Manufacturers of America

(PhRMA; 2011) spent $28 million in 2009 lobbying members of Congress

(Blumenthal, 2010). Despite safety concerns for certain prescriptions and the

unknown effects of others, people rely on medications for numerous reasons.

There is also concern about drug quality (Tognoni, Toussaint, Herxheimer,

& Schaaber, 2014) and the association between research and advertising

(Koch, Brandenburger, Türpe, & Birringer, 2014; McCarthy, 2014; Sacks et

al., 2014; Sood, Kappe, & Stremersch, 2014). This worldwide drug

dependency, questions about physician reliance and ethical decisions (Graf,

Miller, & Nagel, 2014), and the involvement of medical financial resources

(Jofre, 2014; Kmietowicz, 2014) have all created concerns about how

pharmaceutical firms are portraying or marketing their products to

consumers.

These concerns, at least in part, are centered in whether the quality of

direct-toconsumer advertising (DTCA) creates a positive net benefit when

compared to ethical, social, and economic costs (Kesselheim, 2013; Lansing

& Vohra, 2013; Lichtenberg, 2011; Rusthoven, 2014). D. Lee and Emmett

(2012) found that physicians are concerned about denying patients’ requests

for advertised prescription medications. There is a need for further

comprehensive research of direct-to-consumer marketing (DTCM) and

DTCA to determine whether DTCA stimulates patients to seek prescription

drugs that they do not need (Jureidini, Mintzes, Raven, & Block, 2008;

Kulkarni, 2014; McKinlay, Trachtenberg, Marceau, Katz, & Fischer, 2014;

Moore, 2014a).

In this study, I focused on advertising (specifically DTCA) in the

overall marketing field. Hawthorne (2010) claimed that the practice of

advertising directly to consumers was an event that preceded the existence of

the U.S. Food and Drug Administration (FDA). DTCA can be defined as

using the lay media as a tool to promote prescription drug information to the

public (Ventola, 2011). Marketing directly to the consumer is not a new

practice. It was common during the 19th century to find phony medicines

advertised in newspapers. These advertisements often claimed that the

advertised medicine had healing abilities. Debates existed even then over the

advertising and the legal and ethical aspects of marketing campaigns aimed

at the general public. In the 1900s, the American Medical Association

(AMA) tried to end public advertising by contacting medical journals and

requesting an end to such practices (Hawthorne, 2010).

Critics during this time period accused the AMA of trying to control the

medication purchasing process for financial gain.

There has been a transition from the traditional approach wherein

pharmaceutical companies targeted physicians to introduce new products

and to increase the use of their products. In the mid-1990s, focus was

redirected on the end users of prescription drugs

(Hawthorne, 2010). With pressure from the pharmaceutical industry and

other parties, the FDA, which has been the U.S. regulatory body for

prescription medication since 1962, has changed its position on marketing

campaigns aimed directly at consumers

(Hawthorne, 2010). The first DTCA advertisement occurred in 1983, which

is when the FDA requested industry assistance in developing a formal

policy. A draft of this guidance was presented to the FDA in 1997, with the

final version issued to the industry in 1999. The regulations required the

inclusion of a brief summary of product in print advertisements. Possible

drug side effects, contraindications, and effectiveness were required

disclosures information for each drug summary.

The United States and New Zealand are the only two countries that

allow DTCA. Many countries have strict prescription requirements; in some

countries, physicians must provide patients with a password to access

information on the Internet to a drug company. DTCA presents its own set of

challenges for advertising practitioners due to the possible harmful impact

that medicines can have; however, as a whole, ethics in advertising is a

difficult balance to achieve. Despite these challenges, the practice of

DTCA is a sizeable marketing practice that continues to grow.

Pharmaceutical companies spend twice as much on advertising as they do on

research and development (Tadena, 2010). They also court physicians

(Moore, 2014b). The pharmaceutical industry has one of the highest profit

margins of any industry, with the top pharmaceutical companies, including

Pfizer, GlaxoSmithKline, and Bristol-Myers Squibb (Kornfield, Donohue,

Berndt, & Alexander, 2013). As of 2010, Pfizer media spending reached

$967.5 million, with Lipitor as the top advertised drug (Bulik, 2011).

Advertisement spending increased from $700 million in 1996 to $5.4 billion

in 2006 (Bulik, 2011). However, the spending declined by 20% from 2006 to

2010 (Bulik, 2011). The suggested causes for this decline include a decline

in the number of new drugs, a decline in consumer spending, and consumer

skepticism (Bulik, 2011).

Supporters of DTCA list several factors to support the practice, which

include the empowering of consumers with information, public health

awareness, and improved compliance with medication regimens. Those in

support of DTCA argue that, by giving individuals better information, more

informed health decisions can be made (Hawthorne, 2010). Providing this

information is viewed by some supporters as a right to which each person is

entitled. Also, with DTCA, people are becoming more in tune with their

health condition and there is an overall greater awareness of medical

conditions (Hawthorne, 2010). DTCA supporters believe that advertising has

resulted in the general public becoming more comfortable with discussing

medical conditions, as well as discussing such concerns with their physician.

Supporters also argue that the final decision to prescribe a medication still

remains with the doctor, but DTCA makes the patient aware that there may

be other possible treatments (Ahn, Park, & Haley, 2014; Liang & Mackey,

2011). These positions all assume that patients possess adequate

understanding and knowledge to interpret technical information.

There are possible risks associated with taking prescription

medications. Opponents of DTCA claim that DTCA fails to provide accurate

information, increases physician time to correct misconceptions, ignores

prevention and focuses on cure, and increases costs to the health care system

(Huh & Shin, 2014). Opponents argue that advertisements, especially via TV

commercials, cannot adequately cover these issues (Hawthorne, 2010).

Additionally, many drug advertisements are the same as advertisements for

basic and harmless other products. Although a physician’s prescription is

still required, DTCA can be a manipulative technique (Huh & Shin, 2014;

Moore, 2013). Hawthorne (2010) suggested that the decision to take

medications is not a simple process; rather, it is one that involves diagnostic

tests, family history analysis, possible interactions, and other important

elements. For these reasons the decision to take a prescription involves the

consideration of many factors and exposure to DTCA may or may not be

sufficient alone to make a final decision.

Studies have been conducted on exposure to DTCA for specific

diseases. Jureidini et al. (2008), using a prior study conducted by Block

(2007), examined the net social benefit of DTCA of antidepressants. Block

used a DTCA survey and other empirical research to analyze care-seeking

behavior when exposed to antidepressant advertising.

According to Block, the use of advertising of antidepressants leads to a large

net benefit.

However, Block argued that a net benefit would exist even if all people in

the United States were treated with an antidepressant. Jureidini et al. claimed

that Block’s model, which included four steps, was missing a step. In the

original study, the four steps included the following: the number of people

exposed to advertising that motivates treatment, those with clinical

depression, those who may or may not have depression but who receive

prescriptions, and those who benefit from drug treatment. The step missing

from Jureidini et al.’s study involved looking at the proportion of people

experiencing harm from the prescription for both those having and not

having depression.

The differences in the two DTCA research approaches in studying

depression includes arguments regarding the use of various estimates. Block

(2007), using data from the FDA, the Henry J. Kaiser Family Foundation,

and from the U.S. National Ambulatory Care Survey, estimated that 4.59

million (9.4%) of 48.9 million consultations would have been stimulated by

DTCA in 2000. Additionally, Block estimated that, for the same year, 5.85

million people experienced untreated depression. However, Jureidini et al.

(2008) argued that Block’s estimates were overly high. One example of

Block’s overestimate, according to Juriedini et al., was that the number of

untreated people in the United States was derived from a subset in which

there was minimal impairment, which goes against meeting the diagnostic

criteria. Although Jureidini et al. stated that Block’s work contributed to the

DTCA debate, there were some noted weaknesses. Jureidini et al. concluded

that advertising for antidepressants via television is presented in a seductive

manner; however, no bottom line can be drawn on the benefits of DTCA.

Mackert et al. (2013) stated that DTCA is an educational tool that has

benefited the patient by involving them in the decision making process. A

patient may research their condition and the various available drugs once

they are exposed to DTCA. Although physician approval is needed for a

prescription, the patient is no longer relying solely on the physician.

Extensive research on DTCA is limited. However, the FDA Office of

Prescription Drug Promotion (OPDP) has researched and continues to

research this topic. Most studies conducted by the OPDP have been

experimental in nature (HHS, 2004). The survey that was used for the

present study was taken from the only research conducted by the OPDP

using this type of research method. As such, the present study added to the

growing body of research in this field.

Statement of the Problem

There are potentially significant adverse public health care

consequences posed by DTCA, such as overuse, cost burden on patients,

exploited public policy, and wasted limited resources. DTCA, or prescription

drug advertising, is permitted by the FDA, the regulating government agency

in the United States (HHS, 2004). The FDA attempts to assure safe drugs for

both human and veterinary use. The office within the FDA that investigates

the applied/theoretical issues relating to the communication of risk and

benefit for DTCA and professional promotional prescription drug material is

the OPDP.

The pharmaceutical industry is a profitable business. Additionally,

DTCA has grown exponentially from less than $1 billion in 1996 (Bulik,

2011) to $4.2 billion in

2008 (Greene & Kesselheim, 2010; Macias, Lewis, & Baek, 2010), to $11

billion in 2010 (Ghosh & Ghosh, 2010). Pharmaceutical firms record more

than $300 billion in sales revenue annually (Spurling et al., 2010). Provider

targeting promotion has reached $36.1 billion (Kornfield et al., 2013) and

improper drug use costs U.S. employers more than

$276 billion (Conlon et al., 2012) in lost productivity, accidents, and health

care costs. Additionally, developing countries are now spending 20-60% of

their health budgets on prescription drugs, and governments tend to pay

above-market rates, straining already limited governmental resources

(Greene & Kesselheim, 2010). Finally, several pharmaceutical firms have

paid settlements of $2.3 and $3 billion when confronted with allegations of

inappropriate off-label marketing (Kesselheim, Mello, & Studdert, 2011;

Matthews, 2013; Outterson, 2012). The examples of the above settlements

and the overall possibility for large profits suggest the need for regulations

and supervision.

The 340B Drug Discount Program was created by Congress in 1992 to

provide significant discounts to organizations providing care for the

indigent. Despite good intentions, the program has resulted in abuses that

victimize the patients that the program was designed to help. Through an

unrealized contractual loophole, health care organizations were able to

increase their profit margins, while not passing on the expected discounts to

indigent patients (Bress, 2014; J. Lee, 2013; Pollak, 2013). From a public

policy perspective, this loophole provides an incentive to entice patients,

especially the indigent who tend to be less educated (Kaushal, 2014;

Robbins, Stillwell, Wilson, & Fitzgerald, 2012), to request prescriptions they

have seen in advertisements.

Marketing of prescription drugs is distinctive in that potential risks to

the patient in the marketing of these drugs are required to be disclosed by the

FDA. However, patient perceptions about these advertisements may be

distorted, given their often limited clinical understanding. In many cases,

patients are either not afforded adequate time and or lack the ability to fully

understand the risks and benefits of prescription drugs (Bishop & Salmon,

2013; Herbst, Hannah, & Allan, 2013). Not fully understanding the side

effects, complications, and the cumulative interactions of prescription drugs

presents a potential risk of injury or even death for patients. Additionally,

although physician assistance is required, exposure to these types of

advertisements may encourage medication-seeking behaviors (Bishop &

Salmon, 2013; Bradford & Kleit, 2011;

Callaghan, Laraway, Snycerski, & McGee, 2013; D. Lee & Emmett, 2012).

In this study, I used the social learning theory, information integration

theory, and prospect theory to examine consumer behavior as it relates to

DTCA of prescription drugs.

Purpose of the Study

The purpose of this quantitative study was to determine the possible

association between DTCA and health care-seeking behaviors and to explore

patient perspectives on DTCA relative to patients’ overall health care

experience. If there is a link between the advertising component of

marketing campaigns and health care-seeking behavior, an overuse of

unnecessary medications and a subsequent decline in health status, wasted

resources, exploited health policy, and cost burden could result for patients,

employers, and governments. FDA guidance and regulations, federal and

state health care policy, and patient interaction with health care professionals

are all impacted by DTCA.

Researchers remain unclear about the impact of DTCA on patient

behavior

(Callaghan et al., 2013; Lee-Wingate & Xie, 2013; Mukherjee, Limbu, &

Wanasika, 2013; Niederdeppe, Byrne, Avery, & Cantor, 2013). This study

will help to fill this gap and provide a deeper understanding of the possible

association between DTCA and patient behavior. Additionally, given the

concerns and risks associated with inappropriate prescription drug use

(Mackey & Liang, 2013), an additional purpose of this study is to provide

information for more socially responsible decision making for both patient

and doctor.

Research Questions and Hypotheses

The following research questions were addressed in detail and acted

as a catalyst for this study:

1. What is the relationship between direct-to-consumer advertising

(DTCA) and physician office visits?

2. What is the relationship between direct-to-consumer advertising

(DTCA) and patients asking for a prescription?

These questions were researched through two hypotheses. The

independent variable in this study was patient exposure. The dependent

variables were physician office visits and asking for a prescription. Each

hypothesis was tested to determine which independent variables were

significantly associated with patient health care-seeking behaviors.

Hypothesis 1 is related to physician office visits and exposure to DTCA:

H10: There is no relationship between patient exposure to DTCA and

subsequent physician office visits.

H1a: There is a direct association between patient exposure to DTCA

and subsequent physician office visits.

Hypothesis 2 is related to requests for a specific prescription medication and

exposure to DTCA:

H20: There is no relationship between patient exposure to DTCA and

a patient asking subsequently for a corresponding prescription drug.

H2a: There is a direct association between patient exposure to DTCA

and a patient asking subsequently for a corresponding prescription

drug.

Model components are depicted in Table 1. Chapter 3 provides a

detailed description of the research methods and techniques used to analyze

the hypotheses.

Table 1.

Components of the Model

Hypotheses Independent

variables

Dependent variables

1 Patient exposure Physician office visits

2 Patient exposure Asking for a

prescription

Theoretical Base

In this study, the theoretical framework consisted of social learning

theory, information integration theory, and prospect theory to examine

consumer behavior as it relates to DTCA of prescription drugs. Bandura’s

(1971) social learning theory includes elements of conditioning and learning

with cognitive aspects, while extending internal factors to explain behavior.

Social learning theorists attempt to clarify behavior by examining

environmental influences (Thorpe & Olson, 1990). Reciprocal determinism

is the term used to explain the interactional process that applies to the

shaping of behavior (Bandura, 1971). The interactions between the

environment, the person, and the behavior have an equal impact on or

influence of one another. Such interactions can include situations in which a

person makes self-adjustments or when a person talks to himself or herself.

Reciprocal determinism is important in explaining how a person learns by

observing or modeling.

There are various theories that attempt to explain how learning occurs.

Bandura (1971) argued that the majority of learning is gained through a

person’s direct experiences or secondhand observations. Thorpe and Olson

(1990) presented three effects of learning: observational learning effects,

inhibitory and disinhibitory effects, and response facilitation effects. When a

person uses previously learned responses in new situations, this behavior

demonstrates observational learning. In inhibitory and disinhibitory effects, a

person can repeat or not repeat an observed behavior of a model based upon

the consequences observed. If the observer is discouraged from engaging in

the behavior, an inhibitory effect occurs. With disinhibitory effect, the

observer is encouraged by the observed behavior. Response facilitation

effects can be described as observing a behavior that is defined as acceptable

and following four steps in the modeling process.

In the learning process the observer first observes the model and then

retains the information. The observer then compiles all of the observed

information together and then, in the final steps, the actual modeling of the

behavior occurs (Thorpe & Olsen, 1990). Modeling can take on many forms

in addition to an actual live model or observation. Modeling can also take

place from observing media, television, and so on. This type of modeling is

known as symbolic modeling (Thorpe & Olsen, 1990). Also, covert

modeling can take place in which the observer imagines observing a model

(Thorpe & Olsen, 1990). According to Bandura (1969), imitating an

observed behavior does not have to immediately occur following the

observation. A person may not be aware of the observed images or the

imitation of the observed behavior.

Studies have been conducted on DTCA and modeling. One example

of this is when Welch Cline and Young (2004) conducted a content analysis

based on Bandura’s (1971) social learning theory. The purpose of the Welch

Cline and Young study was to identify features of DTCA that may function

as modeling. Visual cues were examined as vicarious motivators. According

to Bandura’s social learning theory, when a behavior is observed, the

possible observed rewards can become motivators. Specifically in DTCA

cures, happy or healthy product users serve as motivators. The Welch Cline

and Young study consisted of reviewing advertisements in 18 popular

magazines for a 2-year period, from January 1998 to December 1999. Four

research questions were developed:

1. What percentage of direct-to-consumer print advertisements

depicted models?

2. What are the demographic characteristics of models depicted in

direct-toconsumer print advertisements?

3. What are the nature and frequency of identity rewards offered

in direct-toconsumer print advertisements via visual cues?

4. What are the natures and frequency of relational rewards

offered in directto-consumer print advertisements via visual

cues, respectively? Welch Cline and Young (2004) indicated

that more than 80% of the advertisements contained models and

35.7% included inanimate objects. Welch Cline and Young also

indicated the use of either male or female models in

advertisements; 33.3% of the ads used more female models

than men. Advertisements with only African Americans,

Hispanic Americans, or Asian Americans were few: 14.2%, 1.1%, and .5%,

respectively.

In 91.8% of the advertisements, people who appeared healthy were depicted.

Welch

Cline and Young concluded that changes in health care behavior might be

triggered by DTCA. Consumers are often exposed to visual models with

positive features, such as being active and friendly.

An example of another study focused on the nonmedical use of

prescription drugs. Mui, Sales, and Murphy (2014) conducted a study of

inappropriate prescription drug use by individuals between the ages of 18

and 25 in the San Francisco Bay area. Social learning theory was used as a

framework to understand the learned and imitated aspects of deviant

behavior. Mui et al. found that through differential association, imitation,

definitions, and differential reinforcement, social context can set the stage

for behavior. Nonmedical drug use is popular among young adults and social

learning theory provides support for the framework for understanding the

initiation of nonmedical drug use. Also, from the 120 total participants in

the Mui et al study, 73.3% were currently attending college.

In comparison differences between social learning theory and

information integration theory. According to social learning theory

(Bandura, 1971), behavior is examined as related to environmental

influences, whereas information integration theorists (Anderson, 2014)

examine how various sources of information are integrated internally by a

person. Information integration theory was proposed by Anderson (2014)

and was developed around four concepts: stimulus integration, stimulus

valuation, cognitive algebra, and functional measurement. The fundamental

concept in integration theory is that the way in which a person thinks or

behaves depends upon multiple stimuli acting in cooperation with one

another. Anderson’s (2014) four concepts interlock when physical stimuli

impose upon a person. Integration function combines the transformed

psychological stimuli into an implicit response, which is then externalized

using the response function. A person uses simple algebraic rules on the

stimulus information before producing a response (Anderson, 2014).

Prospect theory is used to examine how a person makes decisions.

Developed by Kahneman and Tversky (1979), prospect theorists attempt to

describe decisions that are made in which a person makes a choice among

alternatives when risk is involved. There are two phases in the decision

process: editing and evaluation. During the editing phase, there is an initial

analysis of the possible alternatives and the outcomes for a decision. This

stage is followed by an evaluation of the alternatives, with the highest value

choice selected. During the narrowing of alternatives, distinguishing features

are evaluated in terms of gains or losses, while similar components are

disregarded, demonstrating the isolation effect (Nickerson, 2012).

Nature of the Study

In this study, I used a quantitative approach to determine the

association between DTCA and health care-seeking behaviors. Survey data

were collected using a virtual bulletin board accessible to students who

attended an online university and

SurveyMonkey. Various research studies are regularly available on the

virtual bulletin board. Students who show an interest can participate in a

research study after completing an online informed consent form prior to the

survey.

The collected data were analyzed and compared to data available in

the public domain from documents obtained by the U.S. Department of

Health and Human Services (HHS; 2004), of which the FDA is an agency.

Specific reviews of prescription drug advertisements and promotional

labeling are regularly performed by the FDA OPDP. The original research

design used surveys in which participants were asked approximately 65

questions and the variables were measured based upon the responses. The

results of the FDA study were published in 2004; however, the data were

collected in 1999 and 2002 from telephone surveys.

In this study, I used the same 2004 survey instrument used by the

HHS, but via online survey. The survey instrument is located on a public

domain. The data were analyzed using analysis of variance (ANOVA),

which were employed to test the hypotheses of this study. With each

analysis, statistical significance and predictive value were assessed as

applicable. Statistical Package for the Social Sciences (SPSS), Version 21,

was used for all data analyses, and statistical significance was assessed at the

p < 0.05 level. Finally, the predictive power of the two equations was

compared to determine the predictive value of DTCA. Further analytical

details are provided in Chapter 3, with results presented in Chapter 4.

Definitions of Terms

The following technical terms are used in this study:

Direct-to-consumer advertising (DTCA): A component of direct-to

consumer marketing that involves the practice of presenting advertising

material directly to the possible user rather than to the physician (Rollins &

Perri, 2014).

Direct-to-consumer marketing (DTCM): A method of marketing that

links consumers and suppliers together through comprehensive, systematic,

market-based planning, managing, promoting, and advertising of products

directly to consumers

(Rollins & Perri, 2014).

Foreign consumer culture positioning (FCCP): A marketing position

associated with a particular foreign culture (Alden, Steenkamp, & Batra,

1999).

Food and Drug Administration Amendments Act (FDAAA; 2007): A

law that added provisions to the Federal Food, Drug, and Cosmetic Act,

which gives the FDA authority to regulate drugs and other products.

Global consumer culture positioning (GCCP): An alternative

approach to brand positioning that is designed for international marketing.

GCCP involves sharing symbols that denote membership in a global

consumer segment (Alden et al., 1999). Local consumer culture

positioning (LCCP): An approach to marketing that is associated with a

particular local culture (Alden et al., 1999).

Patient exposure: A patient’s recall of seeing or hearing any

advertisement for prescription drugs (HHS, 2004).

Physician office visit: A face-to-face, care-driven interaction with a

physician, physician assistant, or nurse practitioner working for a physician

(HHS, 2004) Voluntary simplicity: The source of personal satisfaction

and happiness coming from nonmaterial aspects of life (Shaw & Newholm,

2002).

Assumptions

In conducting this study, I assumed that all participants responded

truthfully to the survey and that all participants had access to the Internet or

a computer. Additionally, I assumed that respondents to the HHS surveys,

which were used for comparative purposes, answered the questions honestly.

It was also assumed that the established statistical methods employed in this

study were reliable and representative of the national population.

Additionally, I assumed that the influence of mass media on society

(Bandura, 2001) is substantial and the pharmaceutical industry is aggressive

in nature (Angell, 2011).

Scope and Limitations

I collected data via an online survey posted on two survey websites: a

university participant pool and SurveyMonkey. The survey was adapted

from the research instrument used in 2004 by the FDA. Only college

students participated in the present study. However, the original data were

collected from three national surveys (two surveys involving patients and

one survey involving physicians) conducted by telephone. The original FDA

population consisted of individuals in the United States with a listed or

unlisted telephone number. No monetary incentive was offered in the FDA

study or in the present study.

One possible limitation to the present study was the sample size. Due

to the voluntary nature of this study, there was not a large sample. Sampling

error can occur when some persons in the population are omitted (Fowler,

2014). To address this concern, the minimal sample size was calculated. The

inclusion criteria for this study included adult students who had visited a

health care provider within the last year for a health condition of their own

(the same inclusion criteria used in the original FDA study).

The comparative data used in this study were collected in the Patient

and

Physician Attitudes and Behaviors Associated with DTC Promotion of

Prescription

Drugs study conducted by the FDA in 2004. Although the OPDP has

conducted other research relating to DTCA, the 2004 study conducted by the

FDA was the only survey type study at the time the present study was

conducted. Because the 2004 FDA study was one of the first studies in the

area, only general questions were used on the instrument. Also, no analysis

of the comprehension of the advertisement was included. Additionally, the

FDA survey did not distinguish the type of medium for the DTCA exposure,

which may differ from television, print, Internet, or radio. Although a

random sample of telephone numbers (including both listed and unlisted

numbers) were included in the

FDA study, individuals without such service were not represented in the

present study. Additionally, like the previous FDA study, the research

instrument used in the present study was administered only in the English

language.

Significance of the Study

The purpose of this study was to determine the association between

DTCA and health care-seeking behaviors in an academic community and to

explore patient perspectives on the issue related to the patients’ overall

health care experience. If there is an association between the advertising

component of DTCM and health care-seeking behavior, then there could be

an overuse of unnecessary medication. Additionally, patients who receive

and consume unnecessary medication may experience a subsequent

unnecessary decline to their health status and bear an additional cost burden.

The longrange outcome from these findings may indicate a need for stricter

FDA guidance and regulations, changes in federal and state health care

policy, and possible changes to patient interaction and health care

management.

This study will provide a greater understanding and insight into

opportunities to mitigate the potential adverse impacts of DTCA. Overuse,

inordinate cost burdens on patients; health care organizations exploiting

loopholes in public policy; patients, employers, and governments losing

limited resources through inappropriate prescription drug use; and the

diversion of innovation and research efforts to profitable drugs at the

expense of drugs that could have a clinical impact are all potential negative

outcomes resulting from inappropriate DTCA (Suh, 2012). Additionally,

physicians feeling pressured to switch patients from known, effective, less

expensive and older medications to new, more profitable medications, as

well as the potential increased physician workload that may prevent more ill

patients from obtaining limited appointments are other negative outcomes

that this study can help to prevent. DTCA has a potential that can be

leveraged for good, provided that responsibility is exercised.

As pharmaceutical companies are viewed as one of several

contributors to the sustainability of the health care industry, this study can

provide information on focusing DTCA efforts to help improve the

availability, dependability, capability, affordability, and marketability of

prescription drugs. As outlined by Mathaisel and Comm (2014), these

abilities can help the health care industry to minimize waste, create value,

and remain productive for the long term. According to Mathaisel and Comm,

the United

States is ranked 37 out of 191 on overall health system performance by the

World Health Organization. Further, of the seven industrialized nations, “the

U.S. ranked last on quality, efficiency, access, equity and ability for citizens

to lead long, healthy lives” (Mathaisel & Comm, 2014, p. 1046). In terms of

cost-benefit, the United States is viewed as a poor “value” for health care,

given the higher expenditures per capita that fail to yield the expected

benefits or outcomes. Along with patient overtreatment, medical errors,

failures in coordinating care, confusing bureaucracy, and fraud,

pharmaceutical firms are believed to contribute to this less-than-optimal

performance. This study will provide clearer information about

pharmaceutical company opportunities that could, through patients, help to

reverse these adverse national trends.

Increase in the use of DTCA by pharmaceutical companies suggests

that this practice is profitable (Sanky, Berger, & Weinberg, 2012; Yaqub,

2014). Given the potential opportunity for positive or negative impact,

Goldberg (2013) called for more quantitative assessments of DTCA to better

understand its impact. This study will help to address this gap in the

literature and offer opportunities for focusing further research in more

appropriate areas. By providing a clearer understanding of the impact of

DTCA, this study will provide managers and executives, as well as

governments and legislatures, with information that can help to guide policy

development, strategies, and health plan decisions. This guidance can help to

ensure that patients have adequate information to make appropriate

decisions, may drive education-adjusted DTCA, and can help to reduce risks

for patients. Also, guidance has the potential to reduce costs for governments

and employers, and can increase the awareness of inappropriate off-label

advertising. This information can also help to redirect limited economic

resources to provide a greater impact on society by supporting a healthier

workforce and economy (Huebner, 2014). The pharmaceutical industry has

the potential and ability to positively or negatively drive hundreds of billions

of dollars of direct costs and unknown billions in indirect or soft costs. A

deeper understanding of DTCA is necessary for better managerial, clinical,

and health policy decisions.

Summary

DTCA is a lucrative practice in the pharmaceutical industry. With the

overall financial success of this industry using DTCA and the associated

demand for significant returns by investors, discontinuing such practices is

not likely. Regulation by the FDA attempts to protect the public; however,

patient health-seeking behaviors are a concern. The purpose of this study

was to determine the association between DTCA and health care-seeking

behaviors, which were researched through two hypotheses. Additionally,

patient perspectives were explored as they related to the overall health care

experience. This study was quantitative and used an online university

participant pool and

SurveyMonkey to collect survey data. The survey instrument administered in

the present study was used in the original FDA (HHS) 2004 study. The

independent variables were patient exposure to advertising,

sociodemographic characteristics (i.e., age, gender, income, and ethnicity),

health status, and education. The dependent variable, health careseeking

behaviors, were composed of two variables: physician office visits and

asking for a prescription. Only college students were eligible for

participation in the present study.

Honesty by the respondents to the original FDA study and the present study,

the influence of mass media, and the aggressive practices of pharmaceutical

companies were assumptions of this study.

The remainder of the study is organized as follows. Chapter 2

includes a review of the literature on the subject. Chapter 3 contains an

explanation of how the study was designed. Chapter 4 includes a discussion

about the outcomes of the study. Chapter 5 includes a discussion about the

conclusions drawn from the study and an explanation of the

recommendations for further research.

Chapter 2: Literature Review

This review of the literature includes a discussion about the history of

the FDA, which monitors advertising of prescription drugs. Literature that

encompasses various positions regarding the DTCA debate is also included,

as well as literature about the impact of DTCA on pharmaceutical companies

and health policy. The theoretical framework section includes social

cognitive theory, choice, and agency theory as related to advertising. Global

consumer culture positioning is addressed in this review. The field of

research on this topic is limited; however, it continues to grow (Goldberg,

2013). Most studies conducted by the FDA have been experimental in

nature. Several researchers have used data from the original FDA (HHS,

2004) study to further this research base. However, patient attitudes may

have changed over the last 10 years. The present study adds to the current

field of research and serves as a comparison to other study findings.

Title Searches, Journal Articles, and Research Documents

I used refereed journal articles, scholarly books, and research documents

through

Internet search engines that included ProQuest Central, ProQuest Health &

Medical,

Academic Search Complete, Business Source Complete, Cumulative Index

to Nursing &

Allied Health Literature (CINAHL) Plus, Communication & Mass Media

Complete,

PsycARTICLES, PsycBOOKS, ProQuest Digital Dissertations, EBSCO

ebooks, and SAGE Premier. Similar research tools from several local

university libraries complemented the online research.

Background

The idea that certain expensive medications can only be obtained with

a prescription from a physician has stirred controversy. Because there are

physician incentives for the marketing of these medications to consumers

(Jofre, 2014; Kmietowicz, 2014), a better understanding of these interactions

is warranted. These potential conflicts of interest between physician and

patient may present significant concerns (Brill, 2013; Korn & Carlat, 2014;

Perry, Cox, & Cox, 2013). To clarify one aspect of this situation, I sought to

identify the relationships between DTCA and patient medication-seeking

behaviors.

Consumer behavior as it relates to DTCA was considered in this study

within a theoretical framework that included social learning theory,

information integration theory, prospect theory, and emerging theoretical

constructs from published works. This framework served as the basis for the

hypothesized relationships between DTCA, office visits, asking for a

prescription, and several other variables. I considered the potential impact of

DTCA on patients seeking access to prescription medications that can only

be obtained from a provider licensed to prescribe substances that are

regulated by the FDA.

History of the FDA and DTCA

The Pure Food and Drugs Act (1906) marked the beginning of efforts

by the federal government to ensure the safety and veracity of medications.

Standards and quality requirements for pharmaceuticals in Western nations

were relatively nonexistent prior to World War II. The mix of qualified and

science-based pharmacists and illprepared distributors created an increased

degree of confusion for consumers. This confusion was further exacerbated

by the limited scientific knowledge of the period and the tendency of

consumers to exercise several different options (Kastner, 2011), making

inferences beyond the anecdotal difficult. The increasing potency and danger

of formulations in the 1930s further transformed the environment by

introducing the concept of so-called “wonder drugs” with increasingly

positive outcomes within the advertisements. However, there were

significant negative side effects from these increasingly potent medications

provided support for increased governmental oversight and regulation

(Kastner, 2011).

Furthering the aim of public safety, the Federal Food, Drug, and

Cosmetic Act

(1938) gave the FDA oversight of medication production and dispensing to

ensure a more accurate accounting of the risks and benefits of formulations

(Kastner, 2011). Since 1951, federal law has required that drugs be

prescribed by a physician if they pose a high risk of harm if used incorrectly

or abused. Through the Kefauver-Harris amendments, the FDA was given

unprecedented and increasing enforcement authority in 1962 to demand drug

efficacy and safety from manufacturers and distributors (Kastner, 2011).

Wellington (2010) described DTCA in terms of a human right in which the

patient has access to the information needed to make health-related

decisions. This view of DTCA gives patients informed control over their

health related decisions.

With the ultimate goal of public safety and informed consent, the

FDA has considered the authenticity of claims and made appropriate

adjustments to prescribing guidance, limits, and restrictions. Achieving this

goal, based in clinical research and outcomes feedback, has become

increasingly complex with the evolving nature of formulations and

increasing opportunities for off-label usages that are discovered through

advancements (Sashegyi, Felli, & Noel, 2013; Sawyer, 2012). However, the

ratio of benefits to harm for patients taking newer medications varies with

the marketing of the drug (Brody & Light, 2011).

There are four steps that a pharmaceutical company must take to

obtain FDA approval. The first step is known as the preclinical trail

(Hawthorne, 2010). This process involves testing the drug on animals to

determine initial suitability for human testing. During this stage, the drug

company submits summaries of the animal test results and discusses the

manufacturing process. This stage also requires an outline of how the drug

will be tested on humans. The second step (Phase I clinical trials) involves

testing the drug on healthy volunteers to determine safety (Hawthorne,

2010). This trial consists of a small group of healthy participants. Stages 3

and 4, also called Phase II and Phase III clinical trials, consist of participants

with the disease (Hawthorne, 2010). The trials during Phases II and III

involve a larger participant pool than in Phase I. After successfully

completing Phase III, a company can submit an application to start selling

the drug (Hawthorne, 2010).

The FDA has established standardized guidelines on how long the

human trials should last to assist companies through this challenge. There

are also a general number of groups involved in the phases. Phase I is

typically made up of groups of 20 or 50 participants (Hawthorne, 2010). If

Phase I is determined to be safe, Phase II would then include several

hundred people with the disease that the drug is meant to target (Hawthorne,

2010). Phase II can last for 2 years before Phase III begins. Phase III consists

of several thousand patients, typically over the course of 3 years

(Hawthorne, 2010). Another challenge faced by drug companies is FDA

bureaucracy involving selecting the right- or best-fit division for the

proposed drug within the FDA.

Like other governmental agencies, the FDA continues to face many

challenges.

Political influence is one of many pressures on the FDA. Hawthorne (2010)

presented the Reye’s syndrome occurrence in the 1980s to illustrate the

extent of this pressure. Reye’s syndrome in children occurs after a viral

infection and leads to death in 20% of cases; it is also associated with mental

retardation. The Centers for Disease Control and

Prevention (CDC) concluded there was a link between Reye’s syndrome and

aspirin. Although this information was submitted to the FDA, protests from

the aspirin industry resulted in the FDA withdrawing its decision to require

warning labels. The warning label requirement was passed by Congress

more than 5 years after the CDC acknowledgments, after an interest group

sued the FDA. Another problem of the FDA is a history of having approved

products that are later found to be dangerous, such as the Bjork-Shiley

artificial heart valve, which resulted in approximately 500 deaths

(Hawthorne, 2010). Drug companies are not required to obtain FDA

approval prior to disseminating an ad. According to FDA regulations,

prescription drug ads cannot be misleading or omit material facts (Avery,

Eisenberg, & Simon, 2012). The ads must present a “fair balance” regarding

the presentation of risks and benefits. Additionally, if an ad is in print, the

format must include information in a “brief summary about side effects,

contraindications, and effectiveness” (Avery, Eisenberg, & Simon, 2012, p.

252). These guidelines attempt to present the necessary information for

patients to make informed decisions.

Pharmaceutical Industry and DTCA

There can be various reasons why DTCA is used in the industry.

Pharmaceutical companies often use DTCA in an attempt to improve their

brand name, as well as to influence patients to use their particular product.

Menon, Deshpande, Perri, and Zinkham (2003) conducted a study to

determine if consumers attend to the brief summary of the risk information

in the print DTCA. This summary is one of the requirements set by the FDA.

The secondary objective was to determine if consumers found this

information to be useful. Data collected from a 1999 national survey on

consumer attitudes toward DTCA were analyzed. A national sample of

telephone numbers was purchased by the magazine, with a sample size of

1,205. Ten combined hypotheses for the two objectives were developed. For

Objective 1, the hypotheses were as follows:

H1: General consumer characteristics (age, sec, race, and

educational level) influence whether consumers pay attention to

the brief summary. H2: Consumers’ interaction with DTCA

influence whether consumers pay attention to the summary.

H3: Consumers’ attitudes toward the concept of DTCA influences

whether consumers pay attention to the brief summary.

H4: Consumer characteristics specific to health care influence

whether consumers pay attention to the brief summary. (Menon et al.,

2003, p. 183)

For Objective 2, the following hypotheses were developed:

H5: General consumer characteristics influence usefulness of the

brief summary in discussions with physicians.

H6: Consumers’ interactions with DTCA influence usefulness of

summary discussions with physicians.

H7: Consumers’ attitudes toward the concept of DTCA influence

usefulness of brief summary in discussions with physicians.

H8: Consumers’ attention-related characteristics influence

usefulness of brief summary in discussions with physicians.

H9: Consumers’ perceptions of the clarity of the brief summary

influence its usefulness in discussions with physicians.

H10: Consumer characteristics specific to health care influence

usefulness of brief summary in discussions with physicians. (Menon et al.,

2003, p. 183) Menon et al. (2003) used a variety of measurement methods to

analyze the data, including using yes/no as the dependent variable and

demographic factors as the independent or predictor variable.

From the consumers surveyed with no missing responses, 438

subjects remained with an average age of 42. Additionally, the majority of

the participants were women

(62.5%) and Caucasian (87.58%). In summary, the results from the analyses

supported H1, H2, and H3. Under Objective 2, H8 and H9 were also

supported. However, the participants that consumed more prescriptions were

significantly less likely to attend to the summary, which did not support H4.

Additionally, under Objective 2, H5, H6, H7, and H10 were all not

supported. Menon et al. (2003) reached the conclusion that, when attending

to the summary of DTCA, the consumer’s age was not relevant and neither

was the number of prescriptions taken. Also, those who read the entire

summary were less likely to find it to be useful (Menon et al., 2003).

For the last 13 years, Prevention has conducted an annual DTC

survey. In 2010, the data were prepared by Princeton Survey Research

Associates International and consisted of a national sample of 1,501 adults.

Interviews were conducted via telephone and the margin of sampling error

was reported at +/- 3%. The results indicated that consumers believe that

pharmaceutical advertising is presented fairly and balanced in magazines

and television. This finding is related to the FDA regulated “fair and

balanced” mandate of presenting the risk and benefits equally. The study

also reported that, for 5 consecutive years, 79% of consumers have either

seen or heard the risk on TV ads, 73% had seen or heard the benefits on TV

ads, and 48% had seen or heard the benefits in ads in magazines. These

findings also reported 33% of consumers having conversations with their

doctor after seeing an ad, which has remained stable over the past 13 years

(Prevention, 2010)(“Prevention Magazine Releases,” 2010).

Because the knowledge and understanding required to make safe and

appropriate medication decisions is so advanced, the general public must

rely on the expertise of physicians to determine the appropriateness and

authority to obtain and consume some high-risk medications (Moore,

2014a). This agency relationship moderates the conditioning, learning,

integration of information, and prospect decisions of the average person

(Noor, Yap, Kok-Hong, & Rajah, 2014). Within this theoretical model, it is

suggested that patients, regardless of their sociodemographic profile, will

seek to engage their “agent” in their desire to be considered for or obtain

prescription medications.

Newton and Ford (2013) posed the question of whether business and

medicine are ethically incompatible. They described two types of ethics:

professional and market. When the professional acts in the interest of the

client, the behavior is known as professional or fiduciary ethics. In contrast,

market ethics involve each side working in his or her own interest. In the

physician-patient relationship, Newton and Ford (2013) described the

professional ethic as being reflected in the Hippocratic Oath, which states

that no harm or wrongdoing will come to the patient (p. 78).

The economic interests of physicians over patients remains a concern

today (Rusthoven, 2014). However, many changes have occurred in

medicine that affect this relationship. Medical facilities have expanded, and

this expansion has increased competition. There are also more investor-

owned health care facilities. The early view that medicine is a social good

has changed to one that considers medicine to be a commodity, where

patients select services that they can afford. These changes in medicine are a

few of the many factors that have caused providers to compete for patients

and make decisions that may be based on profits (Rusthoven, 2014).

Factors Having an Impact on Health Policy

The health policy aspirations of the federal government, the

capitalistic goals of big business, and the incentives for pharmaceutical firms

and investors, have suggested an additional area for consideration (Mackert

et al., 2013). Although there is evidence that DTCA reduces costs by

reducing preventable care caused by patient noncompliance (Bergner, Falk,

Heinrich, & Hölzing, 2013), significant potential downsides exist (Rodwin,

2013). The heightened degree of regulation by the federal government may

have fostered public confidence and reduced perceptions of risk. This

reduced perception of risk may have also created a secondary effect of

encouraging questionable utilization. Essentially, because the medications

are deemed safe, more consumers who otherwise would not seek or even

need prescription drugs may desire them. Three of the apparent downsides of

such questionable utilization are opioid abuse (Fischer, Keates, Buhringer,

Reimer, & Rehm, 2014), testosterone over-prescribing (Gan, Pattman,

Pearce, & Quinton, 2013), and the resistance to antibiotics that some

organisms have developed due to antibiotic overprescribing. This increase in

questionable utilization may also lead to an increase in patient complaints

(Lewin, 2013) and litigation.

An additional concern is one of underuse. On the heels of the Great

Depression, fear of a recurrence may have influenced health policy towards

ensuring that consumers did not go without needed medications. An

economic theory supporting this argument was that underuse would lead to

higher prices that exclude the less affluent. The combination of these

concerns, intents, agendas, and goals are believed to have framed the

development of advertising, price and patent rules, and guidelines of the

period. Prices, profits, patents, and advertising were increasingly viewed as

vital elements of the effective development and use of modern technology.

This statement is not meant to suggest malicious intent, but is important in

understanding the basis and rationale for legislative activity and decisions by

pharmaceutical firms and associated businesses. However, history does

suggest that regulatory legislation benefits have been heavily weighted

towards business and industry, and less so towards consumers (Blair, 2014;

Jofre, 2014; Kmietowicz, 2014).

Some of the world’s largest pharmaceutical companies are located in the

United

States and Great Britain. Among the largest in the United States are Pfizer,

Merck,

Johnson & Johnson, Bristol-Myers Squibb, and Wyeth. The British

companies include GlaxoSmithKline and AstraZeneca. As prescription drug

prices rise, so does the profit margin for these pharmaceutical companies.

Americans spend billions of dollars on prescription drugs each year. Many

pharmaceutical companies argue that high drug costs are due, in part, to

research and development (R&D). However, Angell (2011) argued that

R&D has little to do with these rising costs and pointed out three reasons to

support her position. First, the total amount spent on R&D makes up a

relatively small part of the drug company budgets (Jack, 2014). Second, the

number of completely new drugs developed is small. Many drugs are only

variations to other already existing drugs

(Angell, 2011; Boumil & Curfman, 2013). An example of this practice

involves Claritin (used for allergies) by Schering-Plough, which originally

made up approximately onethird of Schering-Plough revenues before

expiration of the patent. In an attempt to regain revenues, the company tried

to get users to switch to another patented drug, Clarinex, which was almost

identical to Claritin. The third argument against R&D costs is that

companies can charge whatever they want for the drugs with no or little

restrictions from the government (Peterson, 2014). In fact, the same drugs

are often priced higher in the United States than in other countries. Pricing is

driven, in part, by return on investment demands by shareholders.

Many pharmaceutical companies claim that R&D takes many years

and is extremely costly. The R&D process is divided into two stages:

preclinical and clinical. During the preclinical stage, companies try to find

promising drugs to treat a targeted property. This stage is often

computerized and involves testing various molecules to possible drug

candidates. It is the clinical testing phase that is often expensive (Sashegyi et

al., 2013).

Theoretical Framework

Social Cognitive Theory

Bandura’s (2001) social cognitive theory combines elements of

conditioning and learning with cognitive aspects while extending internal

factors to explain behavior. Thorpe and Olson (1990) explained that

Bandura’s theory attempts to clarify behavior by examining environmental

influences. Reciprocal determinism is the term used to explain the

interactional process in the shaping of behavior. The interaction between the

environment, the person, and the behavior have an equal impact on or

influence of one another. Such interactions can include situations in which a

person makes selfadjustments or when a person talks to himself or herself.

Reciprocal determinism is important in explaining how a person

learns by observing or modeling. Bandura (1971) argued that the majority of

learning is gained through a persons’ direct experiences or secondhand

observations. Thorpe and Olson

(1990) presented three effects of learning:

1. Observational learning effects: Learning that results from

combining previously learned response in new ways;

2. Inhibitory and disinhibitory effects: The consequences of the

model’s behavior has an impact on the observer’s behavior. An

inhibitory effect discourages the observer from engaging in the

model’s behavior. Conversely, a disinhibitory effect would

encourage learning to model the observed behavior.

3. Response facilitation effects: Generally, a person performs

acceptable behaviors after observing others. The consequences

observed influence the observer’s decision to perform the observed

behavior. This modeling process begins with a person showing

attention to the model, which is then followed by some form of

retention of the information. Next, the observer attempts to put the

observed performance together, which sometimes requires small

steps. The forth step involves modeling the behavior, if the

motivational factors are acceptable. (Thorpe & Olsen, 1990, p. 72)

When a person uses previously learned responses in new situations, he

or she is demonstrating observational learning. In inhibitory and

disinhibitory effects, a person can repeat or not repeat an observed behavior

of a model based upon the consequences observed. If the observer is

discouraged to engage in the behavior, the person is exhibiting an inhibitory

effect. But, in demonstrating the disinhibitory effect, the observer is

encouraged by the observed behavior. Response facilitation effects can be

described as observing a behavior that is defined as acceptable and following

four steps in the modeling process. First, the observer observes the model

and then retains the information. The observer then compiles all of the

observed information together and then, in the final step, the actual modeling

of the behavior occurs (Thorpe & Olsen, 1990).

Modeling can take on many forms, in addition to an actual live model

or observation. Modeling can also take place from observing media,

television, and so on.

This type of model is known as symbolic modeling. Also, covert modeling

can take place, in which the observer imagines observing a model.

According to Bandura (1969) imitating an observed behavior does not have

to immediately occur following the observation.

Bandura’s (1965) experiment on imitation of aggressive behaviors

was an attempt to explain learning by observation. Bandura used four-year-

olds in this experiment on imitation of aggressive behavior. The experiment

involved each participant individually observing an adult perform four

aggressive behaviors against a Bobo doll on film. The behaviors included

both physical and verbal aggression. The adult sat on the doll and punched

it, hit the doll with a mallet, kicked the doll, or threw rubber balls at the doll

(all physical behaviors were coupled with verbal aggressive statements). The

child participants then observed the adult being reinforced with snacks and

verbally praised; punished and scolded; or having no consequence

implemented. The children were then taken into a room with a Bobo doll and

other toys. The findings showed aggressive behavior by the children, often

with similar actions that resembled those of the model.

Social Cognitive Theory and Choice Theory

The role of reinforcement and other elements of social cognitive

theory in the learning process were compared and debated by Malone

(2002) in a compatibility analysis with Glasser’s (2010) choice theory.

Although both theories support individual responsibility, the two theories

differ in regard to views on reinforcement, punishment, and self-efficacy.

Choice theory states that all behavior is chosen and the most important need

that a person possesses is love and belongingness (Glasser, 2010). Behavior

is described as humanistic, and a person’s behavior is determined by

survival, love and belonging, power, freedom, and fun (Malone, 2002). The

term reinforcement, which is used in social cognitive theory, was not used

by Glasser due to the external nature that is implied by the term. Glasser’s

choice theory stresses internal control and views punishment as an unethical

and ineffective practice. Additionally, punishments and the administration of

rewards are seen as manipulating and coercive techniques. Internal

motivators can be eliminated if behavior is manipulated. Bandura’s (1965)

research with Bobo dolls illustrated the use of learning by watching others

and vicarious punishment.

One of several similarities between social learning theory and choice

theory is the decision to perform a learned behavior. A person can learn a

behavior in both theories but many decide when or if to perform. Malone

(2002) stated that, according to choice theory, a person creates images in his

or her mind. These images form a persons’ personal quality world. A person

can retain these images or go to the next step of actually acting out the

activity. A persons’ actions can be changed in both social cognitive theory

and choice theory. However, in cognitive theory, changing behaviors is

accomplished though a person changing the pictures within his or her quality

world and not the actual alteration of a behavior, as is the case according to

social cognitive theory. Human behavior is regulated by self-efficacy, and

the concept is an important aspect of both theories. Cognitive, motivational,

affective, and selection aspects are all involved in the self-efficacy process.

Social System

Bandura (2002) discussed how social cognitive theory adapts to

human development, adaptation, and change as part of the social cognitive

theory in cultural context. Culture is described as a social system that is

diverse and changing; therefore, human functioning within culture is unique.

According to Bandura, there are three modes of agency (an intentional

influence on a persons’ functioning, which includes a persons’ life

circumstances): direct personal agency, proxy agency, and collective agency.

Whereas personal agency involves a persons’ own direct control over

conditions, proxy agency relies on others to have a secure wellbeing.

Collective agency relates to group actions in accomplishing desired

outcomes. Bandura stated that people have limited direct influence or control

over social conditions or institutional practices, and people must combine

their skills and resources to accomplish many individual goals.

Cultural differences have an impact on how these three modes of

agency mix for successful functioning. All three types of agency are needed

each day, regardless of the specific culture. Despite the agentic blend,

personal efficacy plays a critical role in a persons’ actions. Personal efficacy

is a persons’ beliefs regarding his or her power to produce the desired

outcome. Human functioning, which includes cognitive, motivational,

affective, and decisional processes, are all regulated through a persons’

personal efficacy beliefs (Bandura, 2002).

The role of culture was debated by Bandura (2002) in terms of

whether a universal human nature exists or if there are several human

natures that combine in various cultural situations. Both direct and vicarious

experiences shape human nature. However, biological limits exist that

influence the shaping of human nature. These biological limits were

described by Bandura as permitting a broad range of possibilities within a

persons’ culture. People adapt to their environments by using various agentic

modes, while adjusting as needed to overcome physical limitations or

restrictions. Cultural differences exist both cross-culturally and

intraculturally; however,

“globalization and pluralization of societies” has ended the insular nature of

cultures (Bandura, 2002, p. 283) Culture is displayed in a persons’ style of

living. Culture has an impact on a person’s choices. Material and

nonmaterial are the two parts of culture. Culture can be described as habits

that encompass diverse factors to include morals and laws (Kahle &

Chiagouris, 2014). Additionally, Kolesnik (2013) using Hofestede,

Hofstede, and Minkov’s (2010) cultural dimensions theory asserted that

people from different places are distinguished by culture.

Advertising and Social Cognitive Theory

The influence of mass media on society is substantial, requiring an

explanation of the psychosocial role of communication and human behavior.

Bandura (2001) described psychosocial functioning using a triadic reciprocal

causation model that consists of personal, behavioral, and environmental

determinants. In general, people are proactive and self-reflecting beings.

Human nature is shaped or molded by direct and observational experiences

within biological limits. Behavior is impacted cognitively by external

influences rather than directly, and cognitive factors also determine which

environmental situations will be observed. A persons’ own knowledge and

understanding is derived from operating symbolically on personal and

vicarious experiences (Bandura, 2001). Bandura (2001) also described

people as self-reactors. Through internal standards and reactions to a

persons’ own behavior, one can self-regulate motivation, affect, and action.

This self-regulation is not limited to negative feedback of oneself, but also

on motivation. This motivation was described by Bandura as a person setting

challenging goals and mobilizing resources to accomplish these goals.

People have internal standards that assess the adequacy of established

achievements. Internal standards also regulate conduct, both socially and

morally. Internal standards can slowly change, but are usually stable.

Morality (right or wrong) was described by Bandura as inhibitive and

proactive. Inhibitive morality is a persons’ power to refrain from inhuman

behavior, and proactive morality is the power to behave humanely.

Bandura’s (2001) social cognitive theory also described people as

being selfreflective. This process involves a person generating ideas, acting

upon them, and then judging adequacy. Fours modes (enactive, vicarious,

social, and logical) exist in the verification of a person’s ideas or thoughts.

The fit between a persons’ thoughts and action results in what occurs with

enactive verification. This type of verification corroborates thoughts (a good

match) or refutes them (a mismatch). When a person observes another’s

actions, he or she can check the correctness of his or her own thinking

process, which is vicarious verification. Social verification is conducted

when one checks a persons’ views against the beliefs of others. Logical

verification involves checking fallacies by deducing from known

knowledge. Each of these four verification forms can produce faulty

thinking. Bandura (2001) presented examples of this faulty thinking in an

illustration in which distorted media versions of reality fostering shared

misconceptions.

In 21st-century society, consumers are presented with images through

media that would not otherwise be available in their lives. Bandura (2001)

summarized four subfunctions that govern observational learning, such as

through media. The subfunctions are attentional processes, cognitive

representational processes, behavioral production processes, and

motivational processes. Attentional processes are those in which a person

selectively observes modeling. Factors that determine selection include

cognitive skills, preconceptions, and held values. Bandura described

cognitive representational processes in terms of memory retention. Observed

events must be remembered to be influential; therefore, a process must occur

that will change or transform the modeled event into a useable form. The

third process, behavioral production, is the transformation of a concept into

action. The final process involves motivational aspects that determine what

encourages a person to perform the acquired behavior, because not all

observed behavior is performed.

Conflicts sometimes occur when there is a conflict between observed

events and internal moral standards. Additionally, vicarious motivators, such

as punishment or reward, can influence the performance of the observed

behavior by an observer. According to Bandura (2001), television and other

forms of media often glamorize representations that may cause internal

discord. This portrayal may work as a motivational element to influence a

person’s behavior. Bandura argued that the media can both create and alter

personal attributes. Technological advances have contributed to

communicating to large numbers of people in a single transmission.

DTCA persuades people to use medications that they may not need,

creating more harm than benefits (Rollins & Perri, 2014). There is also

evidence to suggest that DTCA has the potential to influence patients’

perceptions of treatment plans (Byrne, Niederdeppe, Avery, & Cantor,

2013). Physicians are also not immune to the social psychologically based

efforts of pharmaceutical firms (Ball & Mackert, 2013; Sah & Fugh-Berman,

2013; Sismondo, 2013). New drugs are profitable and, although some testing

has been done, long-term effects are not known. DTCA influences patients

to take these new drugs with sometimes unknown risks. The advertised drug

is often new and inferior to older methods, but is more expensive.

Established medications are often not advertised as heavily because of

generic competition (Ross & Kesselheim, 2013; Tenn & Wendling, 2014).

Pocock (2003) posited for and against non-inferiority trials, the aim of

which is the development of new treatments that are safe. A non-inferiority

trial refers to a “randomized clinical trial in which a new test treatment is

compared with a standard active treatment rather than a placebo or untreated

control group” (p. 483). During clinical trials, no patient is supposed to be

denied effective treatment. In some trials, there may be an active control

group and a placebo. Care during clinical trials is important to prevent the

use of ineffective and unsafe treatments. Also, Type I and Type II errors are

to be avoided. As Pocock (2003) explained, “a Type I error would be the

acceptance of a useless treatment into widespread use,” and “a Type II error

is the failure to use an effective active control treatment by adopting a

placebo control group instead” (p. 484). Most trials involve the patients

receiving the active drug. In some trials, the research participant may

already know that the new treatment is different from the active treatment.

Even with technological advances, diffusion of information differs

among individuals and social groups. Bandura (2001) discussed three

processes (innovative behaviors, adaptation of behaviors in practice, and

social networks) that govern social diffusion of new behavior patterns.

According to Bandura, the acceptance of new ideas or practices is a

challenge due to factors such as customs and social unfamiliarity.

Acceptance then takes on an accelerated course followed by a slowdown in

the rate of diffusion. The dispersion of innovative behaviors, which is one of

the three processes, relies on the method of diffusion (e.g., newspaper or

television). More complicated innovations are more difficult for others to

accept. In general, for modeling of a behavior to occur, many factors are

involved, including human competency and self-belief

(Bandura, 2001).

Another process presented by Bandura (2001) is adoption, which

relates to the many factors that determine if a person will act or engage in a

learned behavior. A person is more likely to adopt an innovation if there are

benefits. These benefits may include social recognition or status as

motivators. People also adopt behaviors that fit into their value system. The

third process that has an impact on diffusion is the social network. Whether

it is an occupational colleague group, friendships, or kinships, people are

often associated with networks. These networks contain various structural

levels and interconnectedness; however, no one social network can serve all

purposes. A person with many network ties is more likely to accept

innovations and have a greater likelihood of exposure to modeling.

Television and online transactions include a few media sources that cross

barriers (e.g., geographic and time) in the transmission of innovations. The

role of mass communication plays an important role in learning.

Social Cognitive Theory and Direct-to-Consumer Advertising

Rosenbaum (1999) used Sen’s (1977) work as an origin and

comparison for research. Sen’s contribution related to culture and

consumption. Rosenbaum divided his work into six sections. The first

section of Rosenbaum’s essay explored various views on consumption.

Although Rosenbaum defined and summarized consumer theory, beginning

as early as 1966, a general introduction to Sen’s background and

contributions were limited for a beginner reader. The second section of the

article explored the importance and impact of goods on culture. Rosenbaum

(1999) remarked, “In fact, goods receive meaning as a consequence of being

used as markers and they are used as markers because they carry meaning”

(p. 322).

Rosenbaum (1999) suggested that a persons’ identity is established by

commodities. The roles of preferences, cultural capital, and inequality are

followed in the fourth section. Rosenbaum noted that goods carry meaning

and should be taken seriously, unlike Sen’s (1977) suggestion. Rosenbaum’s

essay concluded by examining empirical evidence related to cultural capital.

Here, the author focused on social mobility. Erikson and Goldthorpe (1992)

debated the determinants of social mobility. Rosenbaum focused on how

three claims perform in empirical tests, concluding that the evidence

presented by Erikson and Goldthorpe left some doubt due to the difficulty in

representing cultural capital empirically in research. Rosenbaum concluded

that cultural capital has changed over time. He gave the example of the 20th-

century pension system that focuses on the nuclear family rather than the

traditional extended family. In conclusion, considering culture is an

important element. Individuals choose goods based on symbolic meanings

derived from their culture (Rosenbaum, 1999).

The extent of the influence of popular culture may vary. Cusic (2001)

attempted to measure the economic impact of popular culture by gathering

data from the U.S. Department of Commerce. The method by which popular

culture is transmitted is the media; therefore, advertising expenditures were

reviewed. Cusic began with a comparison of advertisement figures for 50

years, starting in 1948, which was the first full year of television

programming. The figures were obtained from the Television Bureau of

Advertising. In 1948, the most popular advertising medium was newspapers,

which received $1.745 billion in advertising revenue, compared to $562

million spent on radio advertising. In 1948, the total amount spent on

advertising was $2.784 billion, of which television received 0%, radio 20%,

magazines 17%, and newspapers 62.7%. In comparison, in 1994, the

advertising expenditures totaled $88.2 billion, of which 40% was spent on

television advertising, 39% on newspapers, 12% on radio, and 9% on

magazines. Therefore, over the 50-year period ending in 1994, advertising

expenditures continued to increase, with the most popular medium changing

from newspapers to television (Cusic, 2001, p. 1). With the popularity of the

internet a shift in the popular medium may have changed again.

Cusic (2001) discussed several possible reasons for the changes in

advertising channels. One theory suggests that newspaper and television

advertising are at-home activities, whereas radio penetration is primarily

outside of the home. These so-called inhome activities have a greater chance

of being used. The penetration of radio advertising may therefore be limited

by the reduced amount of time in which listeners engage in this type of

activity. A second assumption to explain medium changes explored the

number of radio stations compared to the number of daily newspapers in a

market. In a given market, there is often one newspaper, but several radio

stations. It is more economical to advertise in one medium than on several

radio stations. Cusic reported that the number of

U.S. households with cable television increased to 62.4% in 1995, up from

6.7% in 1970. The application of these economic findings has an impact on

the production of movies, music, products, and prescription medications

(Cusic, 2001). The practices of marketing prescription drugs directly to

consumers began in the 1980s, when the FDA removed restrictions. The

possible effects of such practices on the patient/physician relationship,

prescription writing, and consumer education are all factors that may

experience the impact of DTCA.

There have been other studies that have used social cognitive related

theories and

DTCA. Welch Cline and Young (2004) conducted a content analysis based

on Bandura’s

(2001) social cognitive theory. The purpose of the study was to identify

features of DTCA that may function as modeling. Visual cues were

examined as vicarious motivators. Social cognitive theory describes

behavior change in terms of rewards associate with observed behaviors that,

in turn, becomes motivators. Motivators may be direct or vicarious. In

DTCA, financial incentives or claims regarding the potential of a product to

cure an illness may be direct motivators, whereas images of healthy, happy,

or socially engaged product users may be vicarious motivators (Welch Cline

& Young, 2004, p. 136). The present study used Bandura’s social cognitive

theory as part of the theoretical framework.

Many industries, including alcohol and tobacco, engage in social

modeling through the use of billboard advertisements. Social rewards

associated with product consumption in advertisements are known as

relational motivators, whereas the use of attractive cues are identity

motivators. For observational learning to be effective, the reader must be

attentive to the advertisement and associate a positive image.

Additionally, the ad must exhibit the same realities of the reader.

Magazine advertising continues to be a popular medium used by

pharmaceutical companies for DTCA. The study conducted by Welch Cline

and Young (2004) consisted of reviewing advertisements in 18 popular

magazines for a two-year period (January

1998 to December 1999). The researchers coded the articles based on factors

such as whether models were present and the number of people in the ad.

Illustrations featuring rewards were also identified and coded, such as if they

were healthy, active, and friendly models. The first of the four research

questions was, What percentage of direct-toconsumer print advertisements

depicted models? The results indicate that over 80% of the advertisements

contained models and 35.7% included inanimate objects. Welch Cline and

Young further analyzed the advertisements with models by medical

condition. Advertisements for musculoskeletal conditions and respiratory

conditions used people in the advertisements 100% of the time.

Advertisements for allergies, urologic condition, and dermatologic condition

depicted people, 94.3%, 92.3%, and 91.7%, respectively. Infectious

conditions (other than the human immunodeficiency virus [HIV]) and

tobaccoaddiction advertisements were the least likely to depict people, both

62.5%, from the sampled advertisements.

The second research question Welch Cline and Young (2004) asked

was, What are the demographic characteristics of models depicted in direct-

to-consumer print advertisements? The findings showed that either male or

female models were commonly found in advertisements 33.3% of the time.

The use of female-only models (38.3%) in advertisements was more popular

than advertisements containing only male models (26.8%). The authors also

found trends in the advertisements based upon the type of medical condition.

Advertisements for cancer disproportionately showed females (75%). Other

types of medical conditions that depicted women more often than men

included obstetric-gynecologic (89.5%), psychiatric-neurological (71.4%),

and tobacco addiction (60%). The use of men only by medical condition was

popular for such conditions as cardiovascular (66.7%) and gastrointestinal-

nutritional advertisements (66.7%). Advertisements for the drug Viagra

(which, for advertising purposes, the type of medical condition is presented

as undisclosed), depicted both a male and a female model in 100% of the

reviewed advertisements.

With regards to ethnic groups, Welch Cline and Young (2004) found

nearly three fourths of the advertisements (71.6%) depicted Whites only.

Advertisements with only African Americans, Hispanics, or Asians were

few (14.2%, 1.1%, and .5%, respectively).

Additionally, differences were also found for the type of medical condition

and ethnicity. The depiction of Whites was dominant in advertisements for

cancer, cardiovascular, psychiatric-neurological, respiratory, and tobacco-

addiction conditions. African

Americans were most commonly reflected in advertisements for HIV/AIDS

and diabetes, 48.4% and 33.3%, respectively. Hispanics were also more

likely to be depicted in advertisements for HIV/AIDS, even though this

group was less likely to appear in any type of advertisement.

Restating RQ3 and RQ4 of the Welch Cline and Young (2004) study

(What are the nature and frequency of identity rewards offered in direct-to-

consumer print advertisements via visual cues, and what is the nature and

frequency of relational rewards offered in direct-to-consumer print

advertisements via visual cues, respectively). The findings of Welch Cline

and Young’s (2004) study indicated that in 91.8% of the advertisements,

people that appeared healthy were depicted. Additionally, the majority of the

advertisements showed people in some form of activity, social activity

(17%), and physical activity (43.4%). Most advertisements also showed

either smiling (72%) or friendly (64.1%) behaviors. The depiction of social

contexts was found in many of the advertisements (40.7%). This social

context was often family- or romantically orientated,

31.1% and 29.8%, respectively.

Based on these findings, Welch Cline and Young (2004) concluded

that change in health care behavior might be triggered by DTCA. Consumers

are often exposed to visual models with positive features, such as active and

friendly models. Additionally, over 90% of the advertisements showed only

healthy-looking people. Exposure to DTCA may be misleading and promote

stereotypes. Age (Abernathy, Adams-Price, & Henley, 2013), ethnicity

(Ceccoli & Klotz, 2013), where consumers live (Spake, Joseph, & Megehee,

2014), and the mode of DTCA delivery (Bhutada, Deshpande, Menon, &

Perri, 2013; Huh & Shin, 2014; Koch et al., 2014; Vats, 2013) have also

been found to have potential impacts on consumers’ perceptions. The

present study may add to the debate regarding ethical questions surrounding

the use of DTCA. This study presents a comprehensive discussion regarding

the power of DTCA and states that analysis supports observational learning,

which can influence to relationship between physician and patient.

Given these theoretical perspectives, patients are believed to

synthesize information that encourages them to seek medications that reduce

the disparity between their current perceived health status and wellness, and

their desired health status and wellness. DTCA could provide consumers

with information to develop the perception or belief that the advertised

medication could provide better options and more favorable health status

alternatives. This consumer information processing typology suggests that

DTCA can make an impact on consumer behavior by developing the

cognitive processes that synthesize information, develop a need for a change

in health status or standing, identify a solution to that need, and encourage a

desire to act on that need by seeking medications (Mukherjee et al., 2013).

Moderating and Other Theoretical Constructs

Most consumers do not possess the educational background,

pharmaceutical knowledge, or experience to appropriately understand

medication risks. Additionally, they do not have the authority or ability to

obtain prescription drugs without physician approval. As a result, social

learning, information integration, and prospect theoretical postulates must be

mediated or at least moderated within the context of agency theory (Epstein

& Ketcham, 2014).

Agency theory suggests that situations involving information

asymmetry or lack of decision-making authority require a qualified “agent”

to act on behalf of the consumer (Epstein & Ketcham, 2014; Wang, Dou, Li,

& Zhou, 2013). Principle-agent relationships are found in many industries

and products, from real estate to accounting to prescription drugs.

Associations involving these principle-agent relationships must involve

sound ethical and moral standards, given the redistribution of value that

creates the relationship (Epstein & Ketcham, 2014). In the case of the

present study, the consumer-physician relationship required for a

prescription is at least partially facilitated by DTCA. This facilitation,

coupled with consumers’ limited knowledge of behavior-shaping learning,

information processing, and risk factors, completes the theoretical

framework required to better understand the impact of DTCA on

prescription-seeking behaviors (Arney & Lewin, 2013; Spence, 2013).

In keeping with the tenets with agency theory, consumer search

behavior, at least in part, is motivated by perceived risk and consumer ability

to acquire pertinent information to determine whether a purchase is

necessary. In the case of prescription drugs, this risk can be defined as a

diminished health status, illness, injury, or even death (Fountain & Reith,

2014; Ross & Kravitz, 2013). Anvari and Amin (2010) provided further

support for the mediating effects of perceived risk and the increased

consumer search activity. This phenomenon is known as surrogate

consumption theory. The theory focuses on those who must or believe that

they must rely on other persons for the acquisition- and consumption-related

activities that they desire (Lantos, 2010). Given disenfranchisement, a lack

of opportunity, or an inability to act on their own behalf, consumers often

engage in surrogate consumption activities. Galbraith’s controversial thesis

suggested that large corporations seek to manage the demand for their

products (Goldberg, 2013). Implying that the direct manipulation of

consumers’ fuels product demand, this thesis provides a potential financial

explanation for why DTCA may increase demand (Goldberg, 2013).

The FDA OPDP investigates the applied/theoretical issues relating to

the communication of risk and benefit for DTCA and professional

promotional prescription drug material (HHS, 2013). In 2011, several

experimental research studies were completed by the OPDP. In its 2012

study, the FDA investigated alternative formats for presenting the brief

summary statement and examined how people read through and understand

the summary. The study findings showed that adding a serious risk to the

statement being read did not increase or decrease the overall amount of time

taken by the participants. The readers’ intention of asking for the

prescription also did not change under these circumstances. Additionally,

individual characteristics had a greater influence over reading time

compared to characteristics found in the ad statement.

The FDA (2012) found that the amount of time spent reading the

display page and summary were significantly related to the reading speed,

age, and health of the reader. Finally, there was evidence to support that

serious risk information scares possible consumers away. Focus groups are

also used by the OPDP to gather information as related to DTC advertising.

The Food and Drug Administration Amendments Act

(FDAAA) of 2007 has assisted in providing resources for ongoing research

and reviews. Alden et al. (1999) offered global consumer culture

positioning (GCCP) as an alternative approach to brand positioning design

for international positioning. The researchers examined and contrasted

GCCP with local consumer culture positioning (LCCP) and foreign

consumer positioning (FCCP). GCCP involves sharing symbols that denote

membership in a global consumer segment. These shared symbols can take

on several forms, such as language. Alden et al. used the example

advertising with the English language, which is viewed as modernism.

English is therefore used for a brand to convey this associated meaning.

Alden et al. (1999) formulated eight hypotheses. First, all three

positioning approaches are meaningful in television advertising. Second, the

most commonly used strategy in television advertising is LCCP. Third, in

the United States, GCCP is used less frequently in television advertisements.

Fourth, in the United States, LCCP is used more frequently in television

advertisements. Fifth, in television advertisements, more indirect, image-

oriented approaches are used when GCCP is employed. Sixth, if GCCP is

used, food products are less often represented, whereas durable goods often

use this approach. Seventh, in television advertising for food, LCCP is used

most, but LCCP is used least often for durable goods. Eighth, for services,

LCCP is more often used.

A broad array of Asian and Western cultures (e.g., United States,

Germany, Korea, India, Thailand, France, and the Netherlands) were

selected by Alden et al. (1999) to participate. Alden et al. selected random

samples of television advertisements from each country. A random sample

of 20% to 25% of ads was then taken from all collected advertisements. A

total unduplicated sample of 1,267 remained. Graduate students conducted

in-depth coding for each of their native countries.

The results supported Alden et al.’s (1999) primary hypothesis, with

85% of advertisements displaying one of the three positioning options,

GCCP, LCCP, or FCCP. The researchers’ second hypothesis was also

supported, with 59% of advertisements using LCCP, compared to 22.4%

that used GCCP and less than 4% that used FCCP.

Additionally, Chi-square analysis resulted in support for Alden et al.’s third

and fourth hypotheses, both of which relating to positioning in the United

States. Significance was found for types of products that used GCCP, with

56.3% of advertisements using soft-sell approaches, which supported Alden

et al.’s fifth hypothesis. The other findings also support the sixth, seventh,

and eighth hypotheses. GCCP was found to be highest (33.3%) for high-tech

durables, but was only 18.6% of food advertising. In service advertising,

LCCP was most frequently used often positioning option.

Alden et al. (1999) concluded that the then relatively new approach of

GCCP may be beneficial. For example in economically developing

countries, GCCP may work more effectively than might LCCP. This study

contributed to the existing body of knowledge by adding an alternative

approach to advertising. The researchers acknowledged several limitations

in their study, including coding and interpretation of advertisements. The

study appeared to be generalizable, with the use of seven diverse countries

(Alden et al., 1999).

Voluntary simplicity (VS), which has evolved in meaning over the

years, can be defined as a commitment to the nonmaterial aspects of life or

living a simpler life. In practice, a person minimizes consumption of

material goods and may even carry over to working less (Ekstrom & Glans,

2012).

Shaw and Newholm (2002) examined the differences between ethical

consumption behaviors and voluntary simplified behaviors, and presented

findings from two qualitative studies that explored consumer behaviors.

Briefly defined, VS is a consumer’s choice to simplify individual

consumption behavior. In comparison, ethical consumers restrain from

consumption because of some ethical concern, such as taking public

transportation rather than driving a fuel-inefficient car. Shaw and Newholm

(2002) stated that “a fine distinction cannot be made between the extent and

nature of consumption in affluent consumer societies” (p. 169). Therefore,

consumers engage in the following behavior approaches: downshifting,

voluntary simplicity, and sustainable levels of consumption. Shaw and

Newholm described downshifting as a type of VS; however, downshifters

seek to maximum quality time over other motivations, such as income.

Shaw and Newholm’s (2002) premise was that VS may be

demonstrated among consumers whose behavior includes some ethical

consideration of the environmental and social impact of their consumption

choices (p. 180). The two qualitative studies examined by Shaw and

Newholm were conducted in the United Kingdom between 1996 and 1999.

One study consisted of 15 ethical consumers forming two focus groups. The

other study involved 16 case studies of ethical consumers. The respondents’

attitudes in relationship to three consumption areas (diet, car or non-car

travel, and use of secondhand products) were discussed. Shaw and Newholm

reported that, in both studies, the majority of respondents had modified their

diets for various ethical reasons, including treatment of animals and

environmental concerns. Dietary changes included becoming a vegetarian,

reducing meat purchases, only buying free-range animal products, or

increasing purchases of organic products.

Environmental concerns were one of the issues associated with

choices made relative to transportation. Although some respondents owned

cars, they indicated a lack of or problems with public transportation. Shaw

and Newholm (2002) cited that some respondents decided to live in areas in

which employment and amenities were in walking distance to address the

transportation concern. The use of secondhand items was another choice

made by respondents. One respondent reported that it was more economical

and efficient to purchase used products or items, including homes, clothes,

and appliances. The authors concluded that the findings “suggest[ed] that

consumers who start from the premise that ethical issues are applicable to

their consumption also consider the extent of that consumption” (Shaw &

Newholm, 2002, p. 180).

Finally, Ahn et al. (2014) and Park, Ju, and Kim (2014) provided

empirical evidence of the positive association between DTCA and consumer

perceptions of the prevalence and risks of depression. By playing a role in

creating social reality of diseases and medicine, the social cognitive effects

of DTCA are considered far-reaching and influencing to pharmaceutical firm

marketing strategy. This effect further raises complex ethical concerns.

Although the study was limited to print advertisements for antidepressants, it

provided support for the broader exploration of this phenomenon in general

DTCA. The present study sought to contribute to the overall body of

knowledge regarding DTCA by providing empirical support for the

association between DTCA and consumer behaviors.

Review of Research Methods and Differing Methodologies

The approach to research designs can be qualitative, quantitative, or a

combination of both. There are many approaches to each design; some of the

more frequently used include field research, experimental, evaluation, and

survey. Singleton, Singleton, and Straits (2010) defined field research as

directly observing others in a natural setting. This form of observation may

extend over a period of time and may include interaction. Case studies are

one of two approaches of field research. The other type of study is known as

ethnographies. According to Singleton et al., most field research involves a

case study in that a single unit analysis is examined. In comparison,

ethnographies usually describe a culture based upon long field investigation.

Singleton et al. stated that researchers select this type of research to obtain

an insider’s viewpoint. In addition to observing the unit of analysis or social

phenomenon, the researcher can better understand substance of views within

a setting.

Like other research approaches, there are advantages and

disadvantages to using field research. Field research can be less costly in

comparison to other methods because it can be conducted nearby; however,

this type of research tends to be laborintensive. Other disadvantages to using

field research include replication difficulty and issues involving

generalizability. In situations where the researcher has limited knowledge

about the subject or ethical challenges restrict other research approaches,

field research may be used (Singleton et al., 2010).

Experimental research is another approach that is described as an

empirical investigation that attempts to describe a causal explanation. In a

true experimental research design, the participants are assigned randomly to

either an experimental group or a control group (Marczyk, Dematteo, &

Festinger, 2010). Additionally, according to Singleton et al. (2010), this type

of research entails the manipulation of an independent variable while

controlling exposure or contact to other events. Control of extraneous

variables is crucial due to the threat to internal validity.

Singleton et al. (2010) also defined evaluation research as a type of

social science research that focuses on analyzing social programs and

policies. These social programs are primarily instituted by government

entities (federal, state, or local). Evaluation research uses the same types of

methods and addresses many of the same issues as other research methods;

therefore, the designs and validity (internal and external) issues are similar to

those discussed under other methods. The primary distinction is that

evaluation research applies research to social context. Monette, Sullivan, and

DeJong (2013) described three reasons for conducting evaluation research:

to test hypotheses, to support evidence-based practice, and for administrative

purposes.

This present study used survey research. Like the other approaches,

there are many advantages and disadvantages to this approach. Marczyk et

al. (2010) described survey studies as a method in which the researcher asks

a large number of people questions. These questions can address attitudes,

opinions, or specific behaviors. The findings from surveys may be limited to

describing people’s responses, but an attempt may be made to find a

relationship between reported behaviors/opinions and the respondents’

characteristics. Similarly Singleton et al. (2010) explained that the researcher

examines the relationships among the measures once information is gathered

from the survey. Additionally, survey research designs are divided into two

categories: cross-sectional and longitudinal.

The cross-sectional design is one in which data are collected at one

point in time from a sample of respondents selected to represent the target

population. According to Singleton et al. (2010), this one point in time

means that the data are collected in the shortest feasible amount of time. The

cross-sectional design has two variations:

contextual design and social network design. Both types of designs can be

used to study individuals within the same social context; however,

contextual designs involve sample cases within a particular group to

describe characteristics. In comparison, social network designs, which often

require the researcher to interview every person in the study, are used to

examine the relationship among people or other target performers.

Longitudinal designs are studies in which data are collected at more than one

point in time. Researchers using this type of design may ask the same

questions to every individual or independently select samples from the same

population. When repeated surveys are administered to independent samples

of the same population, this type of longitudinal design is known as a trend

study. The other type of study in which the same individuals are surveyed

more than once is known as a panel study. The cross-sectional design is the

more popular of the two designs (Singleton et al., 2010).

Using the survey approach, researchers often strive to make inferences

about a whole (known as the population) from observations taken from a

sample of the whole (Singleton et al., 2010). Because it is often impossible

to observe all actions or events, sampling may be a solution; a sample that is

representative (characteristics are close to those of the target group) of the

target population is used. Typically, sampling reduces time and cost of a

study. Once a researcher has selected the unit of analysis, the number of

units and the method upon which the units will be selected must be decided.

According to Singleton et al. (2010), the first step in sampling is defining the

population of interest, which is a two-step process. The target population,

which is the population to which results are to be generalized, must be

identified. This process can be performed by the researcher by establishing

criteria to determine which cases to include and exclude in the population.

Geographic boundary and time frame are two characteristics identified by

Singleton et al. that can assist in defining the researcher’s target population.

The second step is making the target population operational, which

requires constructing a sampling frame. A sampling frame for a survey

approach often entails obtaining a listing of the population or subgroups of

the population. Next, a sampling design is developed. This design

establishes how cases will be selected for observation and falls into two

categories: probability sampling and nonprobability sampling.

In probability sampling, all cases have a chance of being randomly

selected from a population. In contrast, nonprobability sampling does not

have this known probability because random selection does not occur.

Singleton et al. (2010) discussed five different types of probability sampling:

random selection, simple random sampling, stratified random sampling,

cluster sampling, and systematic sampling. In random selection, each case

within a population has an equal chance for selection. According to

Singleton et al. (2010), mechanical or electronic aids should be used in this

process. Similarly, in simple random sampling, each case has an equal

chance of selection, but this procedure refers to combinations of cases. In the

third type of probability sampling, stratified random sampling, a subdivision

is made to the population first and then simple random samples are selected

from each segment. These subdivisions are mutually exclusive and this

procedure can increase sample precision.

The fourth type of sampling, cluster, also involves dividing the

population; however this breakdown into clusters is according to natural

areas or groupings. Random sampling then occurs from the clusters.

Singleton et al. (2010) offered several examples of clusters to include

colleges, churches, states, and cities. Clustering is performed most often to

reduce the expenses associated with data collection. Another type of

probability sampling is systematic sampling in which a researcher chooses a

number, then selects a case with that number from a complete list of the

population, such as every 10th case (Singleton et al., 2010).

There are several modes of survey instrumentation: face-to-face

interviews, telephone interviews, self-administered questionnaires, or a

combination of these approaches. The overall interviewing technique has the

advantage of allowing the interviewer to clarify or restate questions. Probing

by the interviewer can also be used to clarify answers given by the

respondents. One possible disadvantage to interviewing is the required

training that is often needed for the interviewers. Interviewing can be done

in person (face to face) or over the telephone. Face-to-face interviews

typically have a high response rate, allows for the use of visual aids and the

making of unobtrusive observations. However, one disadvantage to face-to-

face interviewing is the cost. Costs may be reduced by using telephone

interviews; however, establishing rapport with respondents is often more

difficult. Additionally, complex questions may not be suitable for telephone

interviews. Another mode is self-administered questionnaires via the mail or

electronic means, such as the Internet. This approach is often the least costly

of all approaches and has the advantage of allowing the respondents to

complete the item at their convenience (Singleton et al., 2010). The present

study used an online survey approach.

Summary

In this chapter, the theoretical framework was discussed. This

theoretical framework included social cognitive theory, choice, and agency

theory as related to advertising. Global consumer culture positioning was

also discussed in this review.

Research does exist in the overall advertising field of study; however,

research specific to DTCA in comparison is limited. DTCA is a particular

type of advertising; its possible risks may be associated with the use of

prescription medications. Several studies have been conducted by the FDA,

with most being experimental in nature. A 2004 study by the FDA was one

of the first studies conducted in this field, and it has been referenced

frequently in the review of literature. Like the 2004 FDA research, the

present study used the survey method. Other research discussed in this

literature review included a content analysis based on Bandura’s (2001)

social cognitive theory. The purpose of the Welch Cline and Young (2004)

study was to identify features of DTCA that could function as modeling.

Relevant literature covering the history of governmental regulation of

the production and sale of prescription medications, the potential supply and

demand side incentives, as well as the behavioral and agent aspects of

consumer drug seeking were reviewed in developing the theoretical

framework. As a comprehensive consideration of the phenomena associated

with DTCA, this literature review and theoretical framework provided a

foundation and balanced understanding of the factors and issues pertinent to

understanding the complexities of suggested associations. This framework

established the premise for the present study and served as the basis for the

two specific hypotheses that were tested. Chapter 3 contains an explanation

of how the study was designed. Chapter 4 includes a discussion about the

outcomes of the study. Chapter 5 includes discussion, conclusions, and

recommendations for future research.

Chapter 3: Research Method

Introduction

The purpose of this study was to determine the association between

DTCA and health care-seeking behaviors and to explore patient perspectives

on their overall health care experience. Singleton et al. (2010) defined

research, specifically social research, as a process of first formulating

questions and then seeking answers to the questions about a person’s social

environment or surroundings. Both future producers of research and

consumers of research findings benefit from understanding research

methods. From the consumer perspective, research findings are displayed on

a daily basis in everyday life. One example of such use of research is in

advertising, in which results from studies and other forms of research are

introduced to encourage the viewer to purchase a specific product or service.

At times, it can be difficult to decipher which advertisements are presenting

misleading information. Therefore, the consumer must listen and understand

the methods that are used in the overall research process.

Generally, the approaches to research are either qualitative or

quantitative. Research designs can also use a combination of both

(qualitative and quantitative) approaches. Additionally, there can be a

variety of methods or strategies used with either approach (Singleton et al.,

2010). This section provides details on the research design, research

questions and hypotheses, data collection, data analysis, and theoretical

framework.

Research Variables

This research study included two types of variables. The independent

variables were patient exposure to advertising as measured by eight

questions (see Table 2), sociodemographic characteristics (i.e., age, gender,

income, and ethnicity), health status, and education. The dependent variable,

health care-seeking behaviors, were comprised of two variables: physician

office visits as measured by four questions (see Table 2) and asking for a

prescription as measured by four questions (see Table 2).

Research Hypotheses

The two hypotheses of this study posited that patient exposure to

DTCA is associated with physician office visits. Specifically, Hypothesis 1

related to physician office visits and exposure to DTCA:

H10: There is no relationship between patient exposure to DTCA and

subsequent physician office visits.

H1a: There is a direct association between patient exposure to DTCA

and subsequent physician office visits.

Hypothesis 2 related to requests for a specific prescription medication and

exposure to

DTCA:

H20: There is no relationship between patient exposure to DTCA and

a patient asking subsequently for a corresponding prescription

drug.

H2a: There is a direct association between patient exposure to DTCA

and a patient asking subsequently for a corresponding

prescription drug.

Research Design and Approach

In this study, I used a quantitative approach. There are pros and cons

to either (qualitative or quantitative) research design. Denzin and Lincoln

(2011) defined qualitative research as an approach in which the researcher

studies factors or events in their natural settings. The data or material can be

collected through a variety of methods, including case studies, interviews,

observations, historical account, or personal experience. Creswell (2012)

presented a list of the characteristics of qualitative research:

•Natural setting as source of data (the environment in which the

observed event occurs);

•Researcher as key instrument of data collection

•Data collected as words or pictures

•Outcome as process rather than product

•Analysis of data inductively, attention to particulars

•Focus on participants’ perspectives, their meaning;

•Use of expressive language

•Persuasion by reason. (Creswell, 2012, p. 44)

Quantitative research was described by Marczyk et al. (2010) as

involving studies that use statistical analyses in the process of obtaining

findings. Quantitative methods can also be defined as focusing on strict

quantifiable data. In quantitative research, large-scale sampling procedures

are most often used, as well as statistical tests to study averages and

variances of a group (Willig, 2013). The history of quantitative research can

be traced back to the natural sciences, whereas the qualitative research

approach is more related historically to the human or social sciences.

Quantitative research is often represented by the following characteristics:

deductive in nature, is theory-driven, attempts to understand laws and

causes, and is outcome-oriented. In contrast, qualitative research is

inductive, theory-generating, seeks meanings, and is discovery-oriented.

A quantitative survey was used to collect data from participants in the

present study. Specifically, I used the 2004 FDA survey instrument in this

study. Marczyk et al. (2005) claimed that researchers use survey studies to

ask questions to a large number of people. These questions can be used to

investigate attitudes, opinions, or specific behaviors. The findings from

surveys may be limited to describing responses, but an attempt may be made

to find a relationship between reported behaviors/opinions and the

respondents’ characteristics. Similarly, Singleton (2010) stated that

relationships among the measure are examined once people answer

questions, which is the basic idea of a survey. Reasons for using this

approach vary; however, Fowler (2014) stated that surveys may be the only

means of getting information that is easy, quick, and inexpensive. The

budget of researchers vary and surveys may allow for quick collection of

data that is also less expensive.

Surveys are unique from other means of collecting data. Singleton et

al. (2010) described three features of survey research. First, many surveys

consist of a large number of respondents. To represent the target population,

the respondents are selected through probability sampling. Second,

systematic questionnaires or interviews are used as a part of the process of

asking prearranged questions. These interviews may be structured,

unstructured, or semistructured. The optimal structured interview contains

specific objectives with prearranged questions. Comparatively, in an

unstructured interview, the interviewer may make adaptations. The third

feature of surveys consists of the numerically coding and analyzing of

respondents’ answers.

Survey research methods have numerous advantages and

disadvantages. One advantage is that surveys can be an efficient means of

gathering data. This approach can address multiple research questions within

one survey. Another advantage is the lower costs of obtaining data, which

can be even less when using secondary analysis of survey data collected

from professional or other resources. There are also several limitations to

survey research. Surveys are less adaptable and systematic measurement

error may occur (Singleton et al., 2010).

Setting and Sample

The survey for this study was available to participants via the

university website and SurveyMonkey. Inclusion criteria for this study

consisted of adult college students. A diverse population of people

worldwide was represented by this online university and SurveyMonkey.

According to the U.S. Department of Education (2013), the actual number of

college students in 2010 was 21 million and, for the fall of 2013, it was

projected to be 21.8 million. Published sample size tables, online sample

calculators, and formulas are some of the tools that can be used to determine

sample size of a study.

Additionally, the population size, sampling error, and the overall purpose of

the study are all factors that have an impact on the needed sample size for a

study (Cottrell & McKenzie, 2010).

In this study, I used Cochran’s (1963) formula for calculating sample

size:

n = Z2(pq)/e2

where the sample size is n, Z is the appropriate Z score for a

confidence interval, p represents an estimate of the proportion of the

attribute in the population, q is 1 - p, and e is the margin of error or level of

precision. A conservative estimate was used, whereas the value of p was 0.5,

given the unknown variability of the attribute (health care-seeking

behaviors), and the margin of error, e, was 10%. The confidence interval of

95% was used for this study. Using this equation and the projected number

of college students (21.8 million), the suggested minimum sample was 96

participants.

Given the small sample size suggested by Cochran’s (1963) formula, a

power analysis was conducted. Because the alpha level is the probability of

incorrectly rejecting the null hypothesis (Type I error), the alpha level is the

chance of incorrectly inferring a difference where none exists. Beta is the

probability of incorrectly rejecting the null hypothesis (Type II error), or

incorrectly inferring no difference where one actually exists. The power of a

test is measured by 1 – beta and therefore relates to Type II errors.

Decreasing the alpha level increases the probability of a Type II error by

decreasing power but increases the confidence in the results by decreasing

the probability of a Type I error. Conversely, increasing the alpha level

increases the potential for a Type 1 error and decreases the confidence in the

results but decreases the potential for a Type

II error by increasing power. Because the standard deviation and mean for

the population are unknown, a conservative alpha level of 0.05, medium

effect size of 0.30, and power of 0.95 were used to ensure adequate power

and confidence in the results. Using this information, a power analysis and

sample size-determining statistical program called G*Power version 3.1.9.2

(Faul, Erdfelder, Buchner &, Lang, 2009; Jones & Lentz, 2013) determined

a minimum sample size of 177 to provide statistical power of 0.95. This

more conservative required sample size was used for this study.

Instrumentation and Materials

The survey instrument used in the present study was adapted from a

2004 study by the FDA. Report findings from the FDA study, the survey,

and the dataset were all available to the public on the FDA.gov website. The

FDA survey was conducted nationally via telephone in 1999 and 2002.

These previous surveys were basically identical, with only slight revisions

made for clarification purposes and the inclusion of health insurance

questions. The present study used only 24 questions (relevant to study

variables) from the patient survey, the original of which contained 65

questions. The survey was divided into the following sections:

1. Survey inclusion: Participants must be 18 or older and have visited a

doctor, nurse practitioner, or a physician’s assistant for a health

condition or concern of his or her own within the last year. This visit

was for a concern of the patient’s own, not for a child or parent or

someone else.

2. Awareness of prescription drug advertising: Exposure to

advertisements.

3. Interaction with doctor: Patient type of visit and conversation.

4. Attitude/questions about prescription drug advertising: Overall attitudes

about

DTCA.

5. Demographic Information: Education, ethnic group, health insurance,

and so

on.

The survey instrument was adapted for the present study. The original

study was conducted via telephone; therefore, adaptations were made to

conduct the survey online. This change allowed participants to select from a

list of answer choices online compared, to responding verbally to questions

posed over the telephone. The majority of the survey questions were

measured using Likert scales. Table 2 provides a “crosswalk” of the

variables and the selected questions that were derived from the operational

definitions.

Table 2.

Study Variables and Corresponding Survey Questions

Study variables Survey question*

Patient exposure Q3. In the last year, do you recall seeing or hearing

any advertisements for prescription drugs? (Recall

DTCA) Yes No Q4. Have you seen or

heard any ads for prescription drugs in any of the

following ways: (Forms of DTCA)

a. On television

b. On the radio

c. In a magazine

d. In a newspaper

e. On the Internet

f. In a letter, flyer, or announcement you got in the

mail

g. On an outdoor billboard

h. In a grocery store or pharmacy

i. Anywhere else? (please specify)

Table continues

Study variables Survey question

Patient exposure Q5. In the last year, how many different prescription

drugs do you

recall seeing advertised in any form? (Number of

DTCA)

None One Two Three

Four Five Six Seven

Eight Nine Ten More than ten

Q6. Thinking about the ads you have seen both in print

and on television, has an advertisement for a

prescription drug ever caused you to look for more

information, for example, about the drug or about your

health? (Look for more DTCA information) Yes

Q7. What information did you look for? (Type of

DTCA information)

Side effects

Interactions with other drugs/medicines

Dangers of the drug

Cost of the drug

Other__________

Table continues

Study variables Survey question

Patient exposure

Q8. Has an advertisement for a prescription drug

ever caused you to ask a doctor about a medical

condition or illness of your own that you had not

talked to a doctor about before? (Ask doctor about

medical condition) Yes No

Q15. I like seeing advertisements for prescription

drugs. (Like seeing DTCA)

Agree strongly

Agree somewhat

Neither agree nor disagree

Disagree somewhat

Disagree strongly

Q16. Advertisements for prescription drugs help me

make better decisions about my health. (DTCA help

in decision making)

Agree strongly

Agree somewhat

Neither agree nor disagree

Disagree somewhat

Disagree strongly

Table continues

Study variables Survey question

Physician office

visits

Q1. How long has it been since the last time you

saw a doctor, a nurse practitioner, or a physician’s

assistant where you talked about a health condition or

concern of your own, not for a child or parent or

someone else? (Last Visit)

Within the last week

1 to 4 weeks ago

5 weeks to 3 months ago

4 to 6 months ago

7 to 11 months ago

1 year ago

More than 1 year ago

Never

Q2. Was this a routine visit, such as a checkup or

physical?

(Routine visit) Yes No Don’t know

Q9. At any of the visits to your doctor, did you talk

about a prescription drug? (Talk about prescription)

Yes No

Table continues

Study variables Survey question

Physician office

visits

Q14. Overall, how would you rate your interaction

with your doctor at this visit? (Interaction with doctor)

Excellent

Good

Only fair

Poor

Asking for a

prescription

Q10. Did you go to this visit expecting your doctor to

prescribe a drug for you? (Expect Prescription) Yes

Q11. At that visit, did you ask whether there might be

a prescription drug to treat you? (Ask about

prescription)

Yes No

Q12. Did you mention an advertisement you saw or

heard for a drug or bring information about the

advertised drug with you? (Mention

DTCA)

Yes, I mentioned an ad I saw or heard

Yes, I brought something about the drug with me

Yes, both

Table continues

Study variables Survey question

Asking for a

prescription

Q13. Did your doctor do one or more of the

following: (Doctor recommendation) [Select all that

apply]

Give you the prescription drug you asked about

Not give you the prescription drug you asked about

Recommend a different prescription drug

Recommend an over-the-counter drug

Recommend no drug

Recommend you make changes in behavior or

lifestyle

Something else (specify)

Sociodemographics Q17. Overall, would you say your health is: (Health

Status)

Excellent Very good Good

Fair Poor

Q18. How many hours in a typical week do you use

the Internet or

World Wide Web at home and at work? (Use of

Internet or World

Wide Web).

___________________ Do not have a computer.

Q19. Gender (Gender): Male Female

Table continues

Study variables Survey question

Sociodemographics Q20. What is your marital status? (Marital Status)

Married Single Widowed

Divorced Separated

Q21. What is the last grade of school that you

completed?

(Education)

Grade school or less Some high school

Completed high school Some college

Completed college Graduate school or more

Other beyond high school (business,

technical, etc.) Q22. Are you of Hispanic, Latino,

or Spanish origin? (Hispanic origin)

No, not of Hispanic, Latino, or Spanish origin

Yes, Mexican, Mexican American, Chicano

Yes, Puerto Rican

Yes, Cuban

Yes, another Hispanic, Latino, or Spanish origin

– Please state origin, for example Argentinean,

Colombian, Dominican,

Nicaraguan, Salvadoran, and so on

____________________.

Table continues

Study variables Survey question

Sociodemographics

Q23. What is your race? (Race) Select one or more

White Black, African American, or Negro

American Indian or Alaska Native Asian Indian

Chinese Filipino Japanese

Korean Vietnamese

Other Asian – (For example Laotian, Thai,

Pakistani, Cambodian, and so on)

_______________________.

Native Hawaiian Guamanian or Chamorro

Samoan Other race __________________.

Q24. What year were you born? (Age) _________

Note. An abbreviation for each question is presented in parentheses.

Reliability and Validity

Validity refers to the extent to which an instrument measures what it

is purported to measure (Leedy & Ormrod, 2013). There are no statistical

tests for validity, but an instrument is considered valid when the researcher

reaches the opinion that the instrument is measuring what it was designed to

measure. There are several types of validity (criterion, content, and

construct). Criterion validity checks the performance of an instrument to

outcomes that are already held to be valid. Content validity considers where

the instruments items are logically associated with the phenomenon to be

measured. According to Cronbach and Meehl (1955), “construct validity

must be investigated whenever no criterion or universe of content is

accepted as entirely adequate to define the quality to be measured” (p. 282).

In addition to FDA assurances of survey validity from their use of the

instrument for other studies, the questions or variables were compared to

other drug advertising research in peer-reviewed journals as an assessment

of content and construct validity (Bhutada et al., 2013). Because the

questions were direct and were not intended to measure complex

psychometric concepts such as personality, trust, mental capacity, or quality

of life, face validity was determined by comparing the question to the

applicable operational definition to ensure congruency.

Leedy and Ormrod (2013) described reliability as referring to the

extent to which an instrument produces consistent results on repeated tests.

Reliability of an instrument is closely associated with its validity; however,

an instrument cannot be valid unless it is reliable (Tavakol & Dennick,

2011).

The present study used Cronbach’s alpha to measure reliability with

the following formula:

α = Nρ / [1 + ρ(N - 1)]

where N is equal to the number of items and ρ is equal to the mean inter-item

correlation. This calculation is expressed between 0 and 1 and is a measure

of the internal consistency of a test/scale. No pilot testing was performed;

however, researchers who developed, validated, and used the survey

instrument in the 2004 FDA study reported a reliability coefficient of α = .71

when using three items for doctor-patient interaction.

Data Collection

Approval from the institution review board (IRB) was obtained prior

to uploading the study to the participant pool and SurveyMonkey. Once

uploaded and approved, a mass e-mail informing the college community of

the study, as well as a link to the survey, was distributed by the college and

SurveyMonkey. Participants were able to access the survey anonymously.

Based upon the 15-minute collection time reported by the FDA (HHS,

2004), it was anticipated that participants would take approximately 5–8

minutes to complete this online survey. The survey was made available for

several weeks to reach the target sample. A total of 235 participants

completed the survey.

Data Analysis

The present study was quantitative and the data were analyzed using

descriptive statistics and analyses of variance (ANOVAs), which were

employed to test the hypotheses of this study. Descriptive statistics are used

in research to summarize data and numerically describe variables. Inferential

statistics, in contrast, are used to make an assumption about a population

based upon the sample (McNabb, 2013).

With each analysis, statistical significance and predictive value were

assessed, as applicable. A two-step technique was used, as applicable. First,

the independent variables and the dependent variable were loaded into an

ANOVA to determine the predictive power of the equation. Next, the

applicable sociodemographic variable(s) were loaded into an equation with

the dependent variable to assess the impact of the variable(s). Finally, the

predictive power of the two equations were compared to determine the

predictive value of DTCA. Statistical Package for the Social Sciences

(SPSS), Version

21, was used for all data analyses. Given that multiple tests would be run on

the data, the Bonferroni correction was used to control the family-wise error

rate. Although considered conservative, this correction seeks to maintain an

acceptable probability of false positives and false negatives by adjusting the

p value to a level more stringent than 0.05. An

ANOVA can be used to determine if the considered means are different and

the

Bonferroni correction helps to identify specifically which means are

different. Using SPSS Version 21, statistical significance was assessed at the

p < 0.025 level, given the more stringent level required by the Bonferroni

correction. This value was derived by dividing the traditional alpha level of

0.05 by the number of hypotheses being tested (0.05/2=0.025; Jiang,

Barmada, Cooper, & Becich, 2011; Patel, Chen, Kodama, Ioannidis & Butte,

2012; Pollak, Jones, Castillo, Bosse, & MacKenzie, 2010).

The 2004 FDA study sought to assess patient awareness of and

opinions about DTCA efforts, and patients’ processes for obtaining more

information and asking questions. As a largely descriptive study, few

inferential assertions were included in the original study. The present study

used the raw data obtained from the FDA survey for comparative purposes.

A sociodemographic profile of the survey respondents for the 1999 and 2002

surveys is presented in Table 3. The highest proportion of respondents were

between 35 and 54 years of age (43%/40%), female (65%/65%),

White/Caucasian (77%/79%), with incomes of less than $50,000

(53%/53%), completed college or more

(40%/40%), were married (56%/58%), and reported believing that their

health status was excellent/very good (56%/51%). The stark

sociodemographic similarities between the two surveys and society at large

support the reliability of the employed sampling techniques. The present

study had 235 respondents, with characteristics shown in Table 4. Chapter 4

provides more details regarding study respondents.

Theoretical Model

This research used a theoretical framework consisting of social

learning theory, information integration theory, and prospect theory to

examine consumer behavior as it relates to DTCA of prescription drugs. The

2004 study by Welch Cline & Young conducted a content analysis based on

Bandura’s social learning theory. The purpose of the Welch Cline & Young

study was to identify features of DTCA that may function as modeling.

Visual cues were examined as vicarious motivators. Bandura’s social

learning theory describes that when a behavior is observed the possible

observed rewards can become motivators. Specifically, in DTCA cures,

happy or healthy product users are motivators. Welch Cline & Young

concluded that change in health care behavior might be triggered by DTCA.

Consumers are often exposed to visual models with positive features, such

being active and friendly.

The fundamental concept of integration theory is that the way a

person thinks or behaves depends on multiple stimuli acting in cooperation

with one another. Integration function combines the transformed

psychological stimuli into an implicit response, which is then externalized

using the response function. A person uses simple algebraic rules on the

stimulus information before producing a response (Anderson, 2014).

Table 3.

Sociodemographics of FDA (HHS, 2004) Survey Respondents

18–24 7 69 7 65

25–34 17 155 15 140

35–44 23 218 18 171

45–54 20 196 22 208

55–64 14 131 17 164

65+ 20 191 21 196

Respondents

1999 (N2002 ( = 960) N = 944)

% n % n

Age

Gender

Male 35 334 35 327

Female 65 626 65 616

Table continues

Respondents

1999 (N = 960) 2002 (N =

944)

% n % n

Ethnicity (multiple responses

permitted)

American Indian/Alaska Native 4 34 4 34

Asian/Pacific Islander 3 25 2 23

Black/African American 12 116 10 99

Hispanic/Latino 4 43 4 36

White (Caucasian) 77 742 79 747

Income

Less than $20,000 20 189 19 185

$20,000–less than $34,999 17 162 17 161

$35,000–less than $49,999 16 153 17 158

$50,000–less than $74,999 14 132 18 166

$75,000+ 16 155 20 191

Education

Completed high school or less 36 341 39 366

Some college 24 226 21 201

Completed college or more 40 388 40 375

Marital status

Married 56 534 58 550

Table continues

Respondents

1999 (N = 960) 2002 (N =

944)

% n % n

Single 22 212 21 197

Widowed 10 100 9 83

Divorced 9 84 10 95

Separated 3 27 2 16

Health status

Excellent/very good 56 536 51 481

Good/fair/poor 44 421 49 463

Prospect theory similarly examines how one makes decisions.

Prospect theory attempts to describe decisions that are made among

alternatives where risk is involved (Nickerson, 2012). The knowledge and

understanding required to make safe and appropriate medication decisions

is so advanced that the general public must rely on the expertise of

physicians to determine the appropriateness and authority to obtain and

consume some high-risk medications (Pardun, 2014). In agency theory,

this agency relationship moderates the conditioning, learning, integration

of information, and prospect decisions of the average person. Within this

theoretical model, it is suggested that patients, regardless of their

sociodemographic profile, seek to engage their agent in their desire to be

considered for or obtain prescription medications. To summarize the

theoretical model, social learning theory considers environmental

influences on behavior. In the present study, DTCA is considered an

environmental influence on patient behavior. Information integration

theory considers the interaction of multiple stimuli on behavior. In the

present study, DTCA is considered a stimulus affecting patient behavior.

Prospect theory suggests that individuals engage in beneficial decision

making, ensuring that gains are greater than losses. In this study, DTCA

helped individuals draw conclusions about gains and losses in terms of

their health. All these theories about individual behavior must be

considered within the context of agency theory because only doctors can

prescribe prescription medications and patients are generally less educated

about health care than their doctors. Therefore, doctors act in an agent role

on behalf of their patient.

Protection of Participants

The university IRB oversees all proposals to maintain participant

rights and protections. The IRB ensured that the methods of data collection

for this study presented minimal risk to participants, complied with ethical

principles, and met confidentiality requirements. Participation in this study

was voluntary. Participants had the right to withdraw from the study at any

time during the study process. Additionally, participants had the right to ask

questions during the survey process or afterwards. No incentives were given

for participation in this study. Approval from the IRB was obtained before

data collection began.

Summary

The purpose of this study was to determine the association between

DTCA and health care-seeking behaviors and to explore patient perspectives

on the issue relative to patients’ overall health care experience. This research

used a theoretical framework consisting of social learning theory,

information integration theory, and prospect theory to examine consumer

behavior as it related to DTCA of prescription drugs. This research study

included two types of variables. The independent variables were patient

exposure to advertising, sociodemographic characteristics (i.e., age, gender,

income, and ethnicity), health status, and education. The dependent variable,

health care-seeking behaviors, was the summary of two variables: physician

office visits and asking for a prescription. The two research hypotheses of

this study posited that patient exposure to DTCA was associated with

physician office visits.

This study uses a quantitative approach however; there are pros and

cons to either design method. The survey instrument that was used was

taken from the 2004 study conducted by the FDA. This instrument was

made available to participants in the present study via the university website

and SurveyMonkey. Approximately 5,000 individuals were included in the

overall university participant pool. Inclusion criteria for this study consisted

of adult students who had seen a doctor, a nurse practitioner, or a

physician’s assistant for a condition or concern of his or her own in the past

year. A diverse population of people worldwide was represented in this

online university. Data were analyzed using descriptive statistics and

ANOVAs, which were employed to test the hypotheses of this study. The

Bonferroni correction was used to control the family-wise Type I error rate.

With each analysis, statistical significance and predictive value were

assessed, as applicable. A two-step technique was used, as applicable.

Approval from the university IRB was obtained prior to the data collection

process. The results of the analyses of the collected data are presented in

Chapter 4.

Chapter 4: Results

Introduction

The purpose of this study was to determine the association between

DTCA and health care-seeking behaviors and to explore patient perspectives

on patients’ overall health care experience. Two research questions acted as

a catalyst for this study:

1. What is the relationship between direct-to-consumer advertising

(DTCA) and physician office visits?

2. What is the relationship between direct-to-consumer advertising

(DTCA) and patients asking for a prescription?

These questions were researched through two hypotheses. Hypothesis 1

related to physician office visits and exposure to DTCA:

H10: There is no relationship between patient exposure to DTCA and

subsequent physician office visits.

H1a: There is a direct association between patient exposure to DTCA

and subsequent physician office visits.

Hypothesis 2 related to requests for a specific prescription medication and

exposure to

DTCA:

H20: There is no relationship between patient exposure to DTCA and

a patient asking subsequently for a corresponding prescription

drug.

H2a: There is a direct association between patient exposure to DTCA

and a patient asking subsequently for a corresponding

prescription drug.

The data were collected via an online survey and were analyzed with

the statistical program SPSS, Version 21. This chapter includes a

presentation of the collected data, an analysis of the findings, and

summarized results.

Data Collection

Data were collected via an online survey posted on two survey

websites, a university website participant pool and on SurveyMonkey. The

collected data were analyzed with SPSS using ANOVAs, which were

employed to test the hypotheses of this study. The statistical significance and

predictive value were assessed, as applicable.

After receiving approval from the IRB, the survey was uploaded to the

university website for access by an online participant pool. A mass e-mail

list of newly posted studies for that month was sent to inform the college

community. Interested participants were then able to register (if needed) to

use the site and then access the anonymous survey instrument via this online

participant pool. The survey was made available for 4 weeks with few

participants (five). To reach the target sample of 96 participants, the survey

was made available for an additional 3 months, for a total of 132 days. Still,

only 30 participants had accessed and taken the survey via the online

participant pool. To obtain additional participants, I engaged the services of

SurveyMonkey to assist in obtaining additional participants who met the

criteria of at least 18 years of age and a current student.

This study remained a focus of an academic community; therefore, the

criteria remained unchanged. According to the U.S. Department of

Education (2013), the actual number of college students in 2010 was 21

million and, for the fall of 2013, was projected to be 21.8 million.

Maintaining a confidence level of 95% and a confidence interval of 10, the

sample size remained at 96 participants. Prior approval was obtained from

the IRB to make this procedural change of posting on SurveyMonkey.

SurveyMonkey participants received the same invitation/consent forms and

were offered the survey to complete online. The survey consisted of 24

questions (eight questions used to assess patient exposure, four questions for

physician office visits, four questions for asking for a prescription, and eight

sociodemographic questions). There were 205 respondents via

SurveyMonkey. Data from the participant pool and SurveyMonkey were

combined and analyzed for a total of 235 participants. The

sociodemographic profile of participants in this survey was also compared to

the profile of the participants in the FDA (HHS, 2004) 2002 survey. To

minimize the number of deleted cases in each analysis, pairwise deletion

was used to address missing completely at random data (Baraldi & Enders,

2010).

Descriptive Statistics

The survey was opened for volunteers for approximately 8 months. A

total of 89 male students (38%) and 144 female students (62%) participated.

The 2002 FDA survey, although having a larger sample size (944), had a

similar composition of 35% male participants and 65% female participants,

as shown in Table 4. All participants were over 18-years-old, with 90

(38.3%) between 18- and 24-years-old, 59 (25.1%) between the ages of 25

and 34, 27 (11.5%) between the ages of 35 and 44, 16 (6.8%) between the

ages of 45 and 54, and 13 (5.6%) over 55 years of age. By comparison, there

were a larger number of older respondents in the original FDA (HHS, 2004)

survey, with 58% being over the age of 45. The ethnicities of respondents in

both the present study and the FDA survey were similar: 79% were

White/Caucasian American for the FDA survey and 80.5% for this survey.

However, respondents to the original FDA survey showed a higher

percentage of being married at 58%, compared to 28% in this study. Table 3

represents the sociodemographic characteristics of participants in the

original FDA survey. Table 4 shows a summary of the respondent’s

sociodemographic data for this study.

Table 4.

Respondent Sociodemographics (N = 235)

Sociodemographic % n

Age 18-

24 38.3 90

25-34 25.1 59

35-44 11.5 27

45-54 6.8 16

55-64 4.3 10

65+ 1.3 3

Gender

Male 38 89

Female 62 144

Ethnicity (multiple responses permitted)

American Indian/Alaska Native 3 7

Asian/Pacific Islander 1.3 3

Black/African American 9.5 22

Hispanic/Latino 6 14

White (Caucasian) 80.5 186

Table continues

Sociodemographic % n

Income

$0–$24,999 24.9 51

$25,000–$49,999 18 37

$50,000–$99,999 24.9 51

$100,000–$149,999 12.2 25

$150,000+ 20 41

Education

Completed high school or less 7.4 15

Some college 46 94

Completed college or more 46.6 95

Marital status

Married 28 66

Single 60 140

Widowed 1.3 3

Divorced 9.4 22

Separated 0.9 2

Health status

Excellent/very good 58 136

Good/fair/poor 42 99

Explanation of Tables

SPSS Version 21 was the statistical program used to perform

ANOVAs. ANOVA determines if there is a significant difference between

the means of at least two independent variables or groups. The results were

grouped by the applicable hypothesis.

Results are presented in the following ANOVA tables with abbreviated

question titles.

The corresponding full-text survey questions are presented in Table 2.

Hypothesis 1

Hypothesis 1 related to physician office visits and exposure to DTCA:

H10: There is no relationship between patient exposure to DTCA and

subsequent physician office visits.

H1A: There is a direct association between patient exposure to DTCA

and subsequent physician office visits.

Patient exposure to DTCA was associated with several aspects of

physician office visits. As shown in Table 5, seeing an advertisement for

prescription drugs was associated with rating the interaction with a doctor as

positive [F = (1, 229) = 15.94, p = 0.00]. There were no statistically

significant differences (p < 0 .025) between seeing an advertisement for

prescription drugs and more recent visits to a doctor, a nonroutine visit, and

talking to doctor about a prescription drug.

The number of different prescription drug advertisements that were

seen over the past year was associated with rating the interaction with a

doctor as positive [F = (11, 218) = 2.09, p = 0.02], as shown in Table 6.

There were no statistically significant differences (p < 0.025) between the

number of different prescription drug advertisements that were seen over the

past year and talking to a doctor about a prescription drug, the timing of the

last office visit, or the type of visit (routine or nonroutine).

Analytical results in Table 7 demonstrate that viewing a prescription

drug advertisement that caused a search for more information was associated

with a more recent visit to a doctor, nurse practitioner, or physician’s

assistant [F = (1,228) = 8.05, p

= 0.01], and talking to a doctor about a prescription drug [F = (1,231) =

34.70, p = 0.00]. There were no statistically significant differences (p <

0 .025) between viewing a prescription drug advertisement that caused a

search for more information and a nonroutine visit or rating the interaction

with a doctor.

As shown in Table 8, there were no statistically significant differences

(p < 0 .025) between looking for additional information (side effects,

interactions with other drugs/medicines, dangers, and/or costs) and the

timing of the last visit, type of visit, talking to doctor about prescription, or

rating the interaction with the doctor.

Table 5.

One-Way ANOVA for Physician Office Visits: Question 3, Recall DTCAs

Visit/purpose Factor

Sum of

squares df

Mean

square F Sig.

Last visit Between groups

Within groups

Total

.138

891.086

891.224

230

231

.138

3.874

.036 .850

Routine visit Between groups

Within groups

Total

.053

61.566

61.619

229

230

.053

.269

.199 .656

Talk about Rx Between groups

Within groups

Total

.335

57.945

58.281

233

234

.335

.249

1.349 .247

Rate Between groups 8.086 1 8.086 . 15.942 .000*

interaction Within groups

Total

116.148

124.234

229

230

507

Note. Significance was assessed at the *p < 0.025 level. Last visit, routine

visit, talk about Rx, and rate interaction correspond to survey questions 1, 2,

9 and 14 respectively, as all survey questions are displayed in Table 2.

Table 6.

One-Way ANOVA for Physician Office Visits: Question 5, Number of

DTCAs

Visit/purpose Factor

Sum of

squares df

Mean

square F Sig.

Last visit Between groups

Within groups

Total

45.415

845.809

891.224

220

231

4.129

3.845

1.074 .384

Routine visit Between groups

Within groups

Total

1.685

59.934

61.619

219

230

.153

.274

.560 .860

Talk about Rx Between groups

Within groups

Total

5.013

52.970

57.983

222

233

.456

.239

1.910 .039*

Rate

interaction

Between groups 11.683 11 1.062 .5

2.093 .022*

Within groups

Total

110.648

122.330

218

229

Note. Last visit, routine visit, talk about Rx, and rate interaction correspond

to survey questions 1, 2, 9 and 14 respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

Table 7.

One-Way ANOVA for Physician Office Visits: Question 6, Look for More

DTCA Information

Visit/purpose Factor

Sum of

squares df

Mean

square F Sig.

Last visit Between groups

Within groups

Total

29.913

847.569

877.483

228

229

29.913

3.717

8.047 .005*

Routine visit Between groups

Within groups

Total

1.119

60.042

61.162

227

228

1.119 .

265

4.232 .041*

Talk about Rx Between groups

Within groups

Total

7.546

50.231

57.777

231

232

7.546 .

217

34.702 .000*

Rate

interaction

Between groups .143 1 .143 .263 .608

Within groups

Total

123.307

123.450

227

228

.543

Note. Last visit, routine visit, talk about Rx, and rate interaction correspond

to survey questions 1, 2, 9 and 14 respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

Table 8.

One-Way ANOVA for Physician Office Visits: Question 7, Type of DTCA

Information

Visit/purpose Factor

Sum of

squares df

Mean

square F Sig.

Last visit Between groups

Within groups

Total

3.515

887.709

891.224

229

231

1.758

3.876

.453 .636

Routine visit Between groups

Within groups

Total

1.410

60.209

61.619

228

230

.705

.264

2.670 .071

Talk about

Between groups

Within groups

Total

.718

57.563

58.281

232

234

.359

.248

1.447 .237

Rate

interaction

Between groups 1.543 2 .772 1.434 .241

Within groups

Total

122.691

124.234

228

230

.538

Note. Last visit, routine visit, talk about Rx, and rate interaction correspond

to survey questions 1, 2, 9 and 14 respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

Viewing a prescription drug advertisement that caused a search for

more information was associated with a more recent visit to a doctor, nurse

practitioner, or physician’s assistant [F = (1,228) = 8.05, p = 0.01], and

talking to a doctor about a prescription drug [F = (1,231) = 34.70, p = 0.00],

as shown in Table 9. There were no statistically significant differences (p <

0 .025) between viewing a prescription drug advertisement that caused a

search for more information and a nonroutine visit, the type of visit or how

the interaction with the doctor was rated.

As shown in Table 10, agreeing that advertisements for prescription

drugs help make better decisions about health was associated with talking to

a doctor about a prescription drug [F = (4,230) = 2.94, p = 0.02]. However,

there were no statistically significant differences (p < 0 .025) between

agreeing that advertisements for prescription drugs help make better

decisions about health and the timing of the last office visit, the type of visit,

or the rating for interaction with a doctor.

Several sociodemographic variables were associated with more

physician office visits. Better self-reported health status was associated with

rating the interaction with a doctor as positive [F = (4,226) = 4.20, p = 0.00],

as shown in Table 11. There were no statistically significant differences (p <

0 .025) between the self-reported health status and not talking to a doctor

about a prescription drug, timing of the last visit or the type of visit.

In Table 12, being a male participant was associated with a more recent visit

[F =

(1,228) = 9.31, p = 0.00] and talking to a doctor about a prescription drug [F

= (1,231) =

7.69, p = 0.01]. There were no statistically significant differences (p <

0 .025) between being male participant and the type of visit or rating the

interaction with the doctor. As shown in Table 13, being married was

associated with talking to a doctor about a prescription drug [F = (4,228) =

3.23, p = 0.01]. There were no statistically significant differences (p <

0 .025) between marital status and the timing of the last visit, the type of

visit, or rating the interaction with the doctor.

As shown in Table 14, a lower education level was associated with

talking to a doctor about a prescription drug [F = (6, 227) = 3.34, p = 0.00].

There were no statistically significant differences (p < 0 .025) between

education and the timing of the last visit, the type of visit, or rating the

interaction with the doctor.

Table 9.

One-Way ANOVA for Physician Office Visits: Question 8, Ask Doctor About

Medical Condition

Visit/purpose Factor

Sum of

squares df

Mean

square F Sig.

Last visit Between groups

Within groups

Total

19.065

857.404

876.469

226

227

19.065

3.794

5.025 .026*

Routine visit Between groups

Within groups

Total

.154

60.542

60.696

225

226

.154

.269

.574 .450

Talk about Rx Between groups

Within groups

3.606

53667

229

3.606 .

234

15.387 .000*

Total 57.273 230

Rate

interaction

Between groups

Within groups

Total

.376

122.540

122.916

225

226

.376

.545

.691 .407

Note. Last visit, routine visit, talk about Rx, and rate interaction correspond

to survey questions 1, 2, 9 and 14 respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

Table 10.

One-Way ANOVA for Physician Office Visits: Question 16, DTCA Help in

Discussion Making

Visit/purpose Factor

Sum of

squares df

Mean

square F Sig.

Last visit Between groups

Within groups

Total

27.098

864.126

891.224

227

231

6.775

3.807

1.780 .134

Routine visit Between groups

Within groups

Total

.355

61.264

61.619

226

230

.089

.271

.327 .860

Talk about Rx

Between groups 2.836 4

.709

.241

2.941 .021*

Within groups

Total

55.445

58.281

230

234

Rate

interaction

Between groups

Within groups

Total

1.003

123.231

124.234

226

230

.251

.545

.460 .765

Note. Last visit, routine visit, talk about Rx, and rate interaction correspond

to survey questions 1, 2, 9 and 14 respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

Table 11.

One-Way ANOVA for Physician Office Visits: Question 17, Health Status

Visit/purpose Factor

Sum of

squares df

Mean

square F Sig.

Last visit Between groups

Within groups

Total

29.197

862.027

891.224

227

231

7.299

3.797

1.922 .108

Routine visit Between groups

Within groups

Total

1.059

60.560

61.619

226

230

.265

.268

.988 .415

Talk about Rx Between groups 2.425 4 .606 2.497 .044*

Within groups

Total

55.856

58.281

230

234

.243

Rate

interaction

Between groups

Within groups

Total

8.598

115.636

124.234

226

230

2.149 .5

4.201 .003*

Note. Last visit, routine visit, talk about Rx, and rate interaction correspond

to survey questions 1, 2, 9 and 14 respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

Table 12.

One-Way ANOVA for Physician Office Visits: Question 19, Gender

Visit/purpose Factor

Sum of

squares df

Mean

square F Sig.

Last visit Between groups

Within groups

Total

34.571

846.250

880.822

228

229

34.571

3.712

9.314 .003*

Routine visit Between groups

Within groups

Total

.066

61.052

61.118

227

228

.066

.269

.244 .622

Talk about Rx Between groups 1.865 1 1.865 .2

7.694 .006*

Within groups

Total

55.998

57.863

231

232

Rate

interaction

Between groups

Within groups

Total

.161

123.542

123.703

227

228

.161

.544

.296 .587

Note. Last visit, routine visit, talk about Rx, and rate interaction correspond

to survey questions 1, 2, 9 and 14 respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

Table 13.

One-Way ANOVA for Physician Office Visits: Question 20, Marital Status

Visit/purpose Factor

Sum of

squares df

Mean

square F Sig.

Last visit Between groups

Within groups

Total

18.280

871.911

890.191

225

229

4.570

3.875

1.179 .321

Routine visit Between groups

Within groups

Total

.478

60.913

61.391

225

229

.119

.271

.441 .779

Talk about Rx Between groups 3.107 4 .777 3.234 .013*

Within groups

Total

54.756

57.863

228

232

.240

Rate

interaction

Between groups

Within groups

Total

2.579

121.125

1243.703

224

228

.645

.541

1.192 .315

Note. Last visit, routine visit, talk about Rx, and rate interaction correspond

to survey questions 1, 2, 9 and 14 respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

Table 14.

One-Way ANOVA for Physician Office Visits: Question 21, Education

Visit/purpose Factor

Sum of

squares df

Mean

square F Sig.

Last visit Between groups

Within groups

Total

20.631

870.079

890.710

225

230

4.126

3.867

1.067 .379

Routine visit Between groups

Within groups

Total

1.524

59.867

61.391

224

229

.305

.267

1.140 .340

Talk about Rx Between groups

Within groups

Total

4.705

53.367

58.073

227

233

.784

.235

3.336 .004*

Rate

interaction

Between groups

Within groups

Total

3.011

120.833

123.843

223

229

.502

.542

.926 .477

Note. Last visit, routine visit, talk about Rx, and rate interaction correspond

to survey questions 1, 2, 9 and 14 respectively, as all survey questions are

displayed in Table 2.

Significance was assessed at the *p < 0.025 level.

Hypothesis 2

Hypothesis 2 related to requests for a specific prescription medication

and exposure to DTCA:

H20: There is no relationship between patient exposure to DTCA and

a patient asking subsequently for a corresponding prescription

drug.

H2A: There is a direct association between patient exposure to DTCA

and a patient asking subsequently for a corresponding

prescription drug.

Patient exposure to DTCA was associated with patients asking for a

prescription drug. As shown in Table 15, there were no statistically

significant differences (p < 0 .025) for seeing or hearing advertisements and

asking whether there might be a prescription drug to treat the patient/survey

participant, the expectation for a prescription, the mentioning of an

advertisement, or to the doctor’s outcome.

The number of different prescription drug advertisements that were

seen was associated with expecting a doctor to prescribe a drug [F =

(11,219) = 3.31, p = 0.00] and a doctor giving a prescription, recommending

a prescription, recommending an over-thecounter drug, or recommending a

behavior or lifestyle change [F = (11,222) = 2.23, p =

0.01], as shown in Table 16. However, there were no statistically significant

differences (p < 0 .025) between the number of different prescription drug

advertisements and asking whether there might be a prescription drug to

treat the patient/survey participant or mentioning a prescription drug

advertisement that was seen or heard.

As shown in Table 17, seeing an advertisement for a prescription drug

in print or on television that caused a patient/survey participant to look for

more information was associated with expecting a doctor to prescribe a drug

[F = (1,228) = 16.25, p = 0.00], asking whether there might be a prescription

drug to treat the patient/study participant [F = (1,226) = 43.23, p = 0.00],

mentioning a prescription drug advertisement that was seen or heard [F =

(1,229) = 43.33, p = 0.00], and a doctor giving a prescription, recommending

a prescription, recommending an over-the-counter drug, or recommending a

behavior or lifestyle change [F = (1,231) = 37.51, p = 0.00].

As shown in Table 18, looking for side effects, interactions, dangers,

or cost information about a prescription drug was associated with

mentioning a prescription drug advertisement that was seen or heard [F =

(2,230) = 7.76, p = 0.00] and a doctor giving a prescription, recommending a

prescription, recommending an over-the-counter drug, or recommending a

behavior or lifestyle change [F = (2,232) = 14.90, p = 0.00]. There were no

statistically significant differences (p < 0 .025) between looking for side

effects, interactions, dangers, or cost information about a prescription drug

and expecting a doctor to prescribe a drug or asking whether there might be

a prescription drug to treat the patient/survey participant.

Table 15.

One-Way ANOVA for Prescription Request: Question 4, Forms of DTCA

Behavior Factor

Sum of

squares df

Mean

square F Sig.

Expect Rx Between groups

Within groups

Total

2.240

54.204

56.444

223

231

.280

.243

1.152 .330

Might Rx Between groups

Within groups

Total

3.608

50.953

54.561

221

229

.451

.231

1.956 .053*

Mention ad Between groups

Within groups

Total

.930

24.461

25.391

224

232

.116

.109

1.064 .389

Did Dr.

give Rx

Between groups

Within groups

Total

5,445

169.746

175.191

226

234

.681

.751

.906 .512

Note. Expect Rx, Might Rx, Meaningful, and Did Dr. give Rx correspond to

survey questions 10, 11, 12, and 13, respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

Table 16.

One-Way ANOVA for Prescription Request: Question 5, Number of DTCAs

Behavior Factor

Sum of

squares df

Mean

square F Sig.

Expect Rx Between groups

Within groups

Total

8.031

48.238

56.268

219

230

.730

.220

3.314 .000*

Might Rx Between groups

Within groups

Total

3.723

50.687

54.410

217

228

.338

.234

1.449 .153

Mention ad Between groups

Within groups

Total

1.195

23.425

24.621

220

231

.109

.106

1.021 .429

Did Dr.

give Rx

Between groups

Within groups

Total

17.448

157.714

175.162

222

233

1.586 .7

2.233 .014*

Note. Expect Rx, Might Rx, Meaningful, and Did Dr. give Rx correspond to

survey questions 10, 11, 12, and 13, respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

Table 17.

One-Way ANOVA for Prescription Request: Question 6, Look for More

DTCA Information?

Behavior Factor

Sum of

squares df

Mean

square F Sig.

Expect Rx Between groups

Within groups

Total

3.731

52.361

56.091

228

229

3.731 .

230

16.24 .000*

Might Rx Between groups

Within groups

Total

8.677

45.358

54.035

226

227

8.677 .

201

43.233 .000*

Mention ad Between groups

Within groups

Total

4.035

21.324

25.359

229

230

4.035 .

093

43.332 .000*

Did Dr.

give Rx

Between groups

Within groups

Total

24.000

147.802

171.803

231

232

24.000 .

640

37.510 .000*

Note. Expect Rx, Might Rx, Meaningful, and Did Dr. give Rx correspond to

survey questions 10, 11, 12, and 13, respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

Table 18.

One-Way ANOVA for Prescription Request: Question 7, Type of DTCA

Information?

Behavior Factor

Sum of

squares df

Mean

square F Sig.

Expect Rx Between groups

Within groups

Total

1.728

54.716

56.444

229

231

.864

.239

3.615 .028*

Might Rx Between groups

Within groups

Total

1.474

53.087

54.561

227

229

.737

.234

3.152 .045*

Mention ad Between groups

Within groups

Total

1.604

23.786

25.391

230

232

.802

.103

7.757 .001*

Did Dr.

give Rx

Between groups

Within groups

Total

19.938

155.254

175.191

232

234

9.969 .

669

14.897 .000*

Note. Expect Rx, Might Rx, Meaningful, and Did Dr. give Rx correspond to

survey questions 10, 11, 12, and 13, respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

ANOVA results shown in Table 19 indicate that a prescription drug

advertisement that caused a patient to ask a doctor about a new medical

condition or illness was associated asking whether there might be a

prescription drug to treat the patient/study participant [F = (1,224 ) = 31.98,

p = 0.00], mentioning a prescription drug advertisement that was seen or

heard [F = (1,227) = 64.90, p = 0.00], and a doctor giving a prescription,

recommending a prescription, recommending an over-the-counter drug, or

recommending a behavior or lifestyle change [F = (1,229) = 23.60, p =

0.00]. However, there were no statistically significant differences (p <

0 .025) for expecting a doctor to prescribe a drug. Liking to see

prescription drug advertisements was associated with asking whether there

might be a prescription drug to treat the patient/study participant [F =

(4,223) = 4.53, p = 0.00] and mentioning a prescription drug advertisement

that was seen or heard [F = (4,226) = 9.19, p = 0.00], as shown in Table 20.

However, there were no statistically significant differences (p < 0 .025)

between liking to see prescription drug advertisements and expecting a

doctor to prescribe a drug or for the doctor giving a prescription, or

recommending a behavior or lifestyle change.

Table 19.

One-Way ANOVA for Prescription Request: Question 8, Ask Doctor About

Medical Condition

Behavior Factor

Sum of

squares df

Mean

square F Sig.

Expect Rx Between groups

Within groups

Total

.856

54.877

55.732

226

227

.856

.243

3.524 .062

Might Rx Between groups

Within groups

Total

6.684

46.824

53.509

224

225

6.684 .

209

31.977 .000*

Mention ad Between groups

Within groups

Total

5.631

19.696

25.328

227

228

5.631 .

087

64.899 .000*

Did Dr.

give Rx

Between groups

Within groups

Total

16.043

155.706

171.749

229

230

16.043 .

680

23.595 .000*

Note. Expect Rx, Might Rx, Meaningful, and Did Dr. give Rx correspond to

survey questions 10, 11, 12, and 13, respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

Table 20.

One-Way ANOVA for Prescription Request: Question 15, Like Seeing DTCA

Behavior Factor

Sum of

squares df

Mean

square F Sig.

Expect Rx Between groups

Within groups

Total

1.734

54.358

56.091

225

229

.433

.242

1.794 .131

Might Rx Between groups

Within groups

Total

4.077

50.182

54.259

223

227

1.019 .2

4.530 .002*

Mention ad Between groups

Within groups

Total

3.549

21.810

25.359

226

230

.887

.097

9.193 .000*

Did Dr.

give Rx

Between groups

Within groups

Total

5.775

167.359

173.133

228

232

1.444 .7

1.967 .100

Note. Expect Rx, Might Rx, Meaningful, and Did Dr. give Rx correspond to

survey questions 10, 11, 12, and 13, respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

Agreeing that advertisements for prescription drugs help make better

decisions about health was associated with asking whether there might be a

prescription drug to treat the patient/study participant [F = (4,225) = 4.08, p

= 0.00] and mentioning a prescription drug advertisement that was seen or

heard [F = (4,228) = 6.68, p = 0.00], as shown in Table 21. However, no

statistically significant differences (p < 0 .025) were found between agreeing

that advertisements for prescription drugs help make better decisions about

health and expecting a doctor to prescribe a drug or for the doctor giving a

prescription, or recommending a behavior or lifestyle change.

Several sociodemographic variables were associated with patients

asking for prescription drugs. As shown in Table 22, better self-reported

health status was associated with not mentioning a prescription drug

advertisement that was seen or heard [F = (4,228) = 4.19, p = 0.00] and a

doctor not giving a prescription, recommending a prescription,

recommending an over-the-counter drug, or recommending a behavior or

lifestyle change [F = (4,230) = 3.27, p = 0.01]. But self-reported health

status was not associated with expecting a prescription or asking if there

might be a prescription drug to treat the patient/study participant (p <

0 .025).

Table 21.

One-Way ANOVA for Prescription Request: Question 16, DTCA Help in

Decision Making

Behavior Factor

Sum of

squares df

Mean

square F Sig.

Expect RX Between groups

Within groups

Total

1.188

55.255

56.444

227

231

.297

.243

1.221 .303

Might RX Between groups

Within groups

Total

3.686

50.875

54.561

225

229

.922

.226

4.075 .003*

Mention ad Between groups

Within groups

Total

2.662

22.728

25.391

228

232

.666

.100

6.677 .000*

Did Dr.

give RX

Between groups

Within groups

Total

3.125

172.067

175.191

230

234

.781

.748

1.044 .385

Note. Expect Rx, Might Rx, Meaningful, and Did Dr. give Rx correspond to

survey questions 10, 11, 12, and 13, respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

Table 22.

One-Way ANOVA for Prescription Request: Question 17, Health Status

Behavior Factor

Sum of

squares df

Mean

square F Sig.

Expect Rx Between groups

Within groups

Total

1.332

55.112

56.444

227

231

.333

.243

1.372 .245

Might Rx Between groups

Within groups

Total

.660

53.901

54.561

225

229

.165

.240

.689 .600

Mention ad Between groups

Within groups

Total

1.740

23.651

25.391

228

232

.435

.104

4.193 .003*

Did Dr.

give Rx

Between groups

Within groups

Total

9.436

165.756

175.191

230

234

2.359 .7

3.273 .012*

Note. Expect Rx, Might Rx, Meaningful, and Did Dr. give Rx correspond to

survey questions 10, 11, 12, and 13, respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

As shown in Table 23, the number of hours per week spent using the

Internet or World Wide Web was associated with mentioning a prescription

drug advertisement that was seen or heard [F = (36, 191) = 1.6, p = 0.02].

Even so, no statistically significant differences (p < 0 .025) were found

between the number of hours per week using the Internet or World Wide

Web and the expectation for a prescription, asking whether there might be a

prescription drug to treat the patient/study participant, or for a doctor giving

or recommending a prescription.

Being a male patient/study participant was associated with going to a

visit and not expecting a doctor to prescribe a drug [F = (1,228) = 5.65, p =

0.02], as shown in Table 24. No statistically significant differences (p <

0 .025) were found for being male patient/study participant and the

expectation for a prescription, mentioning of an advertisement, or to the

doctor’s prescribing outcome.

Being married was associated with not mentioning a prescription drug

advertisement that was seen or heard [F = (4,226) = 4.11, p = 0.00], as

shown in Table 25. Still, no statistically significant differences (p < 0 .025)

were indicated between marital status and a doctor giving a prescription,

recommending a prescription, recommending an over-the-counter drug, or

recommending a behavior or lifestyle change, the expectation for a

prescription, or asking whether there might be a prescription

drug to treat the patient/study participant.

Table 23.

One-Way ANOVA for Prescription Request: Question 18, Use of Internet or

World Wide Web

Behavior Factor

Sum of

squares df

Mean

square F Sig.

Expect Rx Between groups

Within groups

Total

11.318

44.083

55.401

190

226

.314

.232

1.355 .101

Might Rx Between groups

Within groups

Total

7.393

46.189

53.582

188

224

.205

.246

.836 .733

Mention ad Between groups 5.690 36 .158 1.600 .024

Within groups

Total

18.872

24.561

191

227

.099

Did Dr.

give Rx

Between groups

Within groups

Total

33.271

139.773

173.043

192

229

.899

.728

1.235 .182

Note. Expect Rx, Might Rx, Meaningful, and Did Dr. give Rx correspond to

survey questions 10, 11, 12, and 13, respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

Table 24.

One-Way ANOVA for Prescription Request: Question 19, Gender

Behavior Factor

Sum of

squares df

Mean

square F Sig.

Expect Rx Between groups

Within groups

Total

1.349

54.412

55.761

228

229

1.349 .2

5.651 .018*

Might Rx Between groups

Within groups

Total

.327

53.475

53.803

226

227

.327

.237

1.384 .241

Mention ad Between groups

Within groups

Total

.039

24.567

24.606

229

230

.039

.107

3.64 .547

Did Dr.

give Rx

Between groups

Within groups

Total

1.134

173.338

174.472

231

232

1.134 .7

1.511 .220

Note. Expect Rx, Might Rx, Meaningful, and Did Dr. give Rx correspond to

survey questions 10, 11, 12, and 13, respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

Table 25.

One-Way ANOVA for Prescription Request: Question 20, Marital Status

Behavior Factor

Sum of

squares df

Mean

square F Sig.

Expect Rx Between groups

Within groups

Total

.995

54.936

55.930

225

229

.249

.244

1.018 .399

Might Rx Between groups

Within groups

Total

.605

53.431

54.035

223

227

.151

.240

.631 .641

Mention ad Between groups

Within groups

Total

1.721

23.638

25.359

226

230

.430

.105

4.113 .003*

Did Dr.

give Rx

Between groups

Within groups

Total

6.945

167.527

174.472

228

232

1.736 .7

2.363 .054*

Note. Expect Rx, Might Rx, Meaningful, and Did Dr. give Rx correspond to

survey questions 10, 11, 12, and 13, respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

As shown in Table 26, no statistically significant differences (p <

0 .025) were indicated between higher levels of education and asking

whether there might be a prescription drug to treat the patient/study

participant, mentioning a prescription drug advertisement that was seen or

heard, or the expectation for a prescription or to the doctor’s prescribing

outcome. Likewise, as shown in Table 27, there were no statistically

significant differences (p < 0 .025) between race and the expectation for a

prescription, asking if there was a prescription drug for treatment,

mentioning an advertisement that was seen or heard, or the doctor’s

response.

Summary

The purpose of this study was to determine if an association exists

between DTCA and health care seeking behaviors and to explore patient

perspectives on the issue as they relate to the overall health care experience.

This study was quantitative and the data were analyzed using descriptive

statistics and ANOVA, which were employed to test the hypotheses of this

study. A two-step technique was used, as applicable, using SPSS, Version

21. First, the independent variables and the dependent variable were loaded

into an ANOVA to determine the predictive power of the equation. Next, the

applicable sociodemographic variable(s) were loaded into an equation with

the dependent variable to assess their impact. Finally, the predictive power

of the two equations was compared to determine the predictive value of

DTCA. All data analyses and statistical significance were assessed at the p <

0.025 level, given the Bonferroni correction.

Table 26.

One-Way ANOVA for Prescription Request: Question 21, Education

Behavior Factor

Sum of

squares df

Mean

square F Sig.

Expect Rx Between groups

Within groups

Total

2.527

53.577

56.104

224

230

.421

.239

1.761 .108

Might Rx Between groups

Within groups

Total

3.221

50.963

54.183

222

228

.537

.230

2.338 .033*

Mention ad Between groups

Within groups

Total

1.544

23.831

25.375

225

231

.257

.106

2.430 .027*

Did Dr.

do Rx

Between groups

Within groups

Total

5.337

169.825

175.162

227

233

.890

.748

1.189 .313

Note. Expect Rx, Might Rx, Meaningful, and Did Dr. do Rx correspond to

survey questions 10, 11, 12, and 13, respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

Table 27.

One-Way ANOVA for Prescription Request: Question 23, Race

Behavior Factor

Sum of

squares df

Mean

square F Sig.

Expect Rx Between groups

Within groups

Total

.515

55.929

56.444

229

231

.258

.244

1.055 .350

Might Rx Between groups

Within groups

Total

.662

53.899

54.561

227

229

.331

.237

1.393 .250

Mention ad Between groups

Within groups

Total

.537

24.853

25.391

230

232

.269

.108

2.485 .086

Did Dr.

do Rx

Between groups

Within groups

Total

.810

174.381

175.191

232

234

.405

.752

.539 .584

Note. Expect Rx, Might Rx, Meaningful, and Did Dr. do Rx correspond to

survey questions 10, 11, 12, and 13, respectively, as all survey questions are

displayed in Table 2. Significance was assessed at the *p < 0.025 level.

For Hypothesis 1, there was limited evidence that college-affiliated

adult patients/study participants who saw advertisements for prescription

drugs were more likely to (a) rate their doctor interactions higher; (b) talk to

their doctor about a prescription drug; and (c) have a recent visit to their

doctor, nurse practitioner, or physician’s assistant. Lower health status,

being a man, being married, and lower education levels were all associated

with more office visits.

For Hypothesis 2, there was limited evidence that college-affiliated

adult patients/study participants who saw advertisements for prescription

drugs or searched for additional information about a prescription drug were

more likely to (a) ask their doctor if there was a prescription drug to treat

them; (b) expect a doctor to prescribe a drug; (c) have a doctor give them a

prescription, recommend a prescription, recommend an overthe-counter

drug, or recommend a behavior or lifestyle change; and (d) mention a

prescription drug to their doctor. Lower health status, the number of hours

spent on the Internet or World Wide Web per week, being a woman, being

single, higher education levels, and being an ethnic minority were all

associated with requesting prescription drugs. Chapter 5 includes an

interpretation of the findings, discussion of the findings, the limitations of

the study, implications for social change, recommendations for future study,

and conclusions.

Chapter 5: Discussion, Conclusions, and Recommendations

Summary

The purpose of this quantitative study was to determine if an

association exists between DTCA and health care-seeking behaviors. The

theoretical framework consisted of social learning theory, information

integration theory, and prospect theory. The research questions addressed in

this study included identifying if exposure to DTCA (a) is associated with

physician office visits, (b) influences a patient/physician conversation

regarding a prescription, (c) influences requesting a prescription, and (d) has

an impact on patients’ ratings of the overall interaction with the physician.

Data were derived from an online survey adapted from an FDA study (HHS,

2004). Participants included 235 college-affiliated adults. Data were

analyzed using descriptive statistics and ANOVAs. The Bonferroni

correction was used to control the family-wise Type I error rate. According

to study results, seeing advertisements for prescription drugs was associated

with a recent doctor visit, asking whether a prescription drug was available

to treat a condition, expecting to receive a prescription, receiving a

prescription, and mentioning a prescription drug to a doctor. Future

researchers should consider a non-college-affiliated sample and the post-

implementation impact of the Affordable Care Act. Social change

implications of the study include better consumer education and protection,

more responsible health care policy and corporate decision making, and the

potential prevention of unnecessary drug- and health care-seeking behavior.

Significance of the Study

The United States is ranked low on quality, efficiency, access, equity,

and ability for citizens to lead longer and healthier lives (McCarthy, 2014;

Mathaisel & Comm, 2014). Pharmaceutical companies are viewed as one of

several contributors to the sustainability of the health care industry.

Therefore, this study can provide information to focus DTCA efforts to help

improve availability, dependability, capability, affordability, and

marketability for prescription drugs. The increase in the use of DTCA by

pharmaceutical companies suggests that this practice is profitable. Given the

potential opportunity for positive or negative results, more assessments of

DTCA are needed to understand its impact. This study could help to address

this gap in the literature and offer opportunities for focusing further research

in more appropriate areas.

This study could also provide greater insight into opportunities to

mitigate the potential impacts of DTCA. Overuse and inordinate cost

burdens on patients are all potential negative outcomes resulting from

inappropriate DTCA. Additionally, physicians feeling pressured to switch to

new, more profitable medications may be another negative outcome that this

study can help to better understand. Understanding the impact of DTCA will

provide managers and executives, as well as governments and legislatures,

with information that can help to guide policy development, strategies, and

health plan decisions. Additional guidance can help to ensure that patients

have adequate information to make appropriate health care decisions.

Assumptions

Several assumptions were made in this study. Among them are the

following. It was assumed that all participants responded truthfully to the

survey and all participants had access to the Internet or a computer.

Additionally, I assumed that respondents to the HHS (2004) survey, which

was used for comparative purposes, answered the questions honestly. It was

also assumed that the established statistical methods employed in this study

were reliable and representative of the national population. Additionally, I

assumed that the influence of mass media on society (Bandura, 2001) is

substantial and the pharmaceutical industry is aggressive in nature (Angell,

2011).

Limitations

Several limitations must be acknowledged relative to this study.

Generalizability is limited because this study was open to only college-

affiliated adults. Additionally, due to the reliance upon the online university

participant pool and SurveyMonkey, participants had to have online access.

In this sample of 235 participants, there was an underrepresentation of many

racial ethnic groups, with 80.5% of participants selfreporting as members of

the White/Caucasian American category. Given the statistically significant

differences identified across sociodemographic characteristics, this disparity

could have affected the outcomes. Additionally, there may be differences

relative to ethnic minorities who were not discovered, given the small

representation of this population in this study. Overall, due to the nature of

the survey—personal health— respondents may not have been comfortable

answering survey questions. Finally, college students are traditionally

younger, more educated, and in other ways different from members of

society in general.

Interpretation of Findings

The association between DTCA and health care-seeking behaviors

and patient perspectives on the issue relative to the patient’s overall health

care experience were studied. SPSS was used to analyze data. All data

analyses and statistical significance were assessed at the p < 0.025 level. The

following research questions were addressed in detail and were the catalyst

for this study:

1. What is the relationship between direct-to-consumer advertising

(DTCA) and physician office visits?

2. What is the relationship between direct-to-consumer advertising

(DTCA) and patients asking for a prescription?

These questions were researched through two hypotheses. Hypothesis

1 related to physician office visits and exposure to DTCA:

H10: There is no relationship between patient exposure to DTCA and

subsequent physician office visits.

H1a: There is a direct association between patient exposure to DTCA

and subsequent physician office visits.

Hypothesis 2 related to request for a specific prescription medication and

exposure to

DTCA:

H20: There is no relationship between patient exposure to DTCA and

a patient asking subsequently for a corresponding prescription drug.

H2a: There is a direct association between patient exposure to DTCA

and a patient asking subsequently for a corresponding prescription

drug.

For Hypothesis 1, there was limited evidence that college-affiliated

adult study participants who saw advertisements for prescription drugs were

more likely to (a) rate their doctor interactions higher; (b) talk to their doctor

about a prescription drug; and (c) have a recent visit to their doctor, nurse

practitioner, or physician’s assistant. Lower health status, being a man, being

married, and lower education levels were all associated with more office

visits.

For Hypothesis 2, there was limited evidence that college-affiliated

adult study participants who saw advertisements for prescription drugs or

searched for additional information about a prescription drug were more

likely to (a) ask their doctor if there was a prescription drug to treat them;

(b) expect a doctor to prescribe a drug; (c) have a doctor give them a

prescription, recommend a prescription, recommend an over-the-counter

drug, or recommend a behavior or lifestyle change; and (d) mention a

prescription drug to their doctor. Lower health status, the number of hours

spent on the Internet or World Wide Web per week, being a woman, being

single, higher education levels, and being an ethnic minority were all

associated with requesting prescription drugs.

Implications for Social Change

The United States and New Zealand are the only two industrialized

nations that permit DTCA, and pharmaceutical firms in these countries have

the highest profit margins of any industries in these two countries

(Hawthorne, 2010). With increasingly greater numbers—more than 50%—

of all people in the United States taking prescription drugs, the total annual

retail sales of prescription drugs exceeding $300 billion a year, and more

than $28 million being spent by PhRMA to lobby members of Congress, it is

difficult for the FDA to ensure consumer protection (PhRMA, 2011). The

significant societal dependency on prescription drugs provides a potentially

endless supply of individuals who are vulnerable to abuse. Deadly side

effects and contraindications are but a few of the many dangers associated

with prescription drugs. Additionally, the principalagent relationship

between consumers and physicians provides an example of the potential

patient vulnerabilities.

The most significant findings of this study are that DTCA is

associated with patients asking more questions, having more office visits and

patients having a lower overall health status. As DTCA is associated with

patients asking their physicians more questions, this increased

communication could help patients make better decisions about the potential

risks, benefits and costs of prescription drugs. As patients, especially those

who perceive themselves as less healthy, attempt to understand the very

complex issues around their health care, asking questions can help.

Given that DTCA is associated with more office visit, physicians can

help patients to better understand this tendency as well as the positive or

negative impacts.

Physicians are trained to help patients make informed decisions about their

health care. Being aware of the DTCA and office visits association can help

physicians assist patients in taking better care of themselves, minimizing

unnecessary office visits, and reducing their out of pocket costs.

Evidence from this study also suggests a strong association between

DTCA and patients with a lower health status. Knowledge of this association

can help patients and physicians to focus their efforts on ensuring a positive

social impact. Awareness of this relationship can change patient, physician

and healthcare policy maker decision making and reduce the likelihood of

adverse clinical and financial implications. This could provide a significant

opportunity to help the least healthy of society to greatly improve their

health, quality of life and ability to return to work if they have been hindered

by their poor health status.

Additionally, health care policy has a financial impact and drives

corporate and individual decisions and behavior. This study provides a

clearer understanding of what aspects of DTCA should be carefully

considered in the development of local and national health care policy. This

understanding will reduce the probability of unintended negative

consequences from legislation and policy.

Other social change implications of this study include consumer

protection in a risky market and the prevention of unnecessary and

expensive drug- and health care seeking-behavior. The discovered link

between the advertising component of marketing campaigns and health care-

seeking behavior can potentially drive an overuse of unnecessary

medications and a subsequent unnecessary reduction in health status and

added cost burden for patients. FDA guidance and regulations, federal and

state health care policy, and patient interaction with health care professionals

are all documented factors that may be touched by DTCA. As a result, it is

important to understand the impact of DTCA because it can be used to

improve the health status and economic prosperity of society. This

knowledge can also be used to better manage the potential for abuse, given

the industry financial incentives. Using the identified significant aspects of

DTCA, physicians have a better understanding of the aspects of DTCA that

impact patient decision making. Having this information could help

physicians to develop more effective ways of communicating with patients

and developing more effective care plans.

Recommendations for Further Study

Future researchers can address the limitations of this study by

investigating a larger sample size that is not limited by Internet access,

college affiliation, or age. Although the type of medium used in this study

was not restrictive geographically, cultural and geographic preferences and

differences may still have existed.

This study focused on participants associated with an academic

community. A study that collects information regarding students’ major and

current occupation might also provide additional insights regarding attitudes

and behaviors relative to DTCA. Behaviors and attitudes may vary

depending upon type of occupation, medical experience, and training. As an

example, undergraduate nursing students may provide different answers than

undergraduate business students. The degree to which patients truly

comprehend and understand the information in DTCA could provide an

additional research opportunity. Finally, an assessment of how much

information patients retain from DTCA and whether that retention is

tempered or enhanced by health status or the length of the patient-physician

relationship. Further research may reveal that different degrees of retention

exist for patients with severe conditions such as terminal cancer, and patients

with other conditions such as a minor cold. Additional research may also

reveal that the length of an established physician-patient relationship may

mitigate the impact of

DTCA.

Quantifying the financial impact of DTCA could provide insight into

corporate incentives. Research that better clarifies the incentives and

disincentives for pharmaceutical companies could help to drive safer

corporate decisions and potentially influence health care policy. Research in

this area could help to ensure that corporate incentives are aligned with

desired positive patient outcomes.

The FDA study (HHS, 2004) was one of the initial studies to focus on

the doctorpatient relationship and its broad implications for health care. The

OPDP continues research projects on this topic, with current studies

including an examination of online DTC drug promotion, experimental

study of format variations in the brief summary of DTC print

advertisements, and health care professional survey of prescription drug

promotion. With students as participants, this study provided information

that could enhance the overall patient care experience.

Finally, given the full implementation of the unprecedented employer,

health care industry, and individual changes required under the Affordable

Care Act, future research may produce different results. Signed into law by

President Barack Obama in 2010, the Affordable Care Act seeks to improve

access to the U.S. health care system and increases individual responsibility

as it relates to a persons’ health care. Once fully implemented in the coming

years, employer requirements to provide coverage for older children,

individual mandates to obtain coverage or be subjected to a penalty, and

health care industry performance requirements for payment could likely

produce different research findings and opportunities. Future research could

also help with the assessment of the impact of the Affordable Care Act.

Concluding Remarks

A few of the hypothesized associations were not supported

quantitatively, so this study can contribute to the development of new

knowledge by specifying the aspects of DTCA that are associated with

patient behaviors and perceptions. The findings of this study outline the

specific models that more clearly explain the impact of DTCA. Social

change implications of the study include consumer protection in a risky

market and the prevention of unnecessary, expensive, and potentially

dangerous drug- and health careseeking behaviors.

Direct-to-Consumer Advertising of Drugs and Patients' Health Care Seeking Behaviors

Students also viewed