Skip to main content

Factors Affecting the Pen Habits of Physicians: An Exploratory Factor Analytic Study

Case
By: & Published: 2017 | Product: SAGE Research Methods Cases Part 2
+- LessMore information
Search form
No results
Not Found
Download Case PDF

Abstract

This case is based on a research project conducted by us, during summers of 2014. The rising competition in the Indian pharmaceutical industry prompted us to explore the pen habits of Indian doctors in the northern region of India. The case explains the methodological approach used by the two authors to conduct the project and overcome research barriers. The case is specifically intended to explain to students as to what aspects should be taken into consideration while designing a questionnaire and how factor analysis should be conducted. Furthermore, the case explains how the two authors draw the managerial implications with respect to the pen habits of doctors, based on factors derived from the study.

Learning Outcomes

By the end of this case, students should

  • Be able to understand the significance of research design
  • Have a better understanding of in-depth interviews
  • Be able to understand sampling techniques like judgmental sampling
  • Be able to understand the process of questionnaire design
  • Have an understanding of the use of exploratory factor analysis and its usage
Introduction

The Indian pharmaceutical industry is highly competitive with approximately 20,000 companies operating. Although 250 firms control 70% of the market, yet competition among the top players is severe. Some of the leading players are Cipla, Ranbaxy, Dr. Reddy’s Laboratories, Lupin Pharmaceuticals, and Sun Pharmaceuticals. As consumers’ expenditure on healthcare is increasing, the industry is growing at a rapid pace. However, on the other hand, mergers and acquisitions are also increasing in the industry and top players are struggling with their market share.

Marketing Strategy of Pharmaceutical Firms in India

One of the major marketing strategies of pharmaceutical firms consists of promoting their drugs to doctors through medical or pharmaceutical representatives. This could be understood from the fact that medical representative expenditure is somewhere between 15%-20% of revenues. In medical industry, drug choice does not lay in the hand of the end consumer, typically the patient, but doctors who prescribe the drugs to the patients. Hence, most of the marketing efforts are targeted toward doctors. Although in rare cases companies might communicate directly with end consumers, but that is generally to raise awareness of the disease rather than the drug.

Challenges for Indian Pharmaceutical Firms

There is an intense competition among the top few pharmaceutical players in India, who are struggling to sustain their market share and position. To help these firms in devising their marketing strategy through managerial implications of our project, the authors thought that firms need to understand their customers (i.e. doctors) very well. Recent completion of dissertation prompted the authors to use their freshly learned research techniques and methodological approaches to solve this problem. Thus, the broad objective of authors’ project was to explore through survey as to what influences drug prescribing behavior of doctors, which in the pharmaceutical industry is referred to as “pen habits” of doctors. For example, to cure fever, a doctor may prescribe a drug manufactured by Cipla, Ranbaxy, Dr. Reddy’s, or may be some other firm. How do doctors decide on which firm and which brand to choose? This is referred to as pen habit of doctors.

Since this project was our self-initiative, covering pan-India would have been a much expensive proposition. Hence, we decided to focus on major cities of north India only. We believed that conducting this research was need of the hour as many pharmaceutical firms in India were blindly pouring money to promote their drugs without exactly knowing how doctors choose and prescribe drugs. We knew that once the factors affecting doctor’s choice/s are known, firms would be better able to channelize its resources to promote its drugs and hence expand its market share.

The Study

We started the project in March 2014 and completed it by August 2014, which explained in detailed the factors determining the pen habit of doctors’ primarily in the north zone of India. We decided to target our study toward the general physicians only as the core competency of many Indian pharmaceutical firms lies in generic drugs, and patients normally visit a general physician only for such drugs.

Research Design

The first step toward the study was to develop the framework for conducting the project. The framework and research design were rather complex, justifying the 6 months of time which we took to complete this project. Research design should clearly indicate the research question; information required; the exploratory, descriptive, and causal phases of research; the measurement and scaling procedure; development of the questionnaire; specifying the sampling process and sample size; and developing a plan of data analysis. For pharmaceutical firms, the information that was required was factors which affected the pen habits of the doctor. This implied use of exploratory research to identify the variables or items which were relevant for the study.

Exploratory Research: Review of Literature

To identify the variables influencing the pen habit of doctors, we first reviewed several research papers to explore how such a study has been conducted in the past and what have been the results. For example, while reviewing the study of Venkatraman and Stremerch (2007), we found that marketing efforts of the pharmaceutical firm and side effects of drug and clinical trials impacted the drug prescribing behavior of doctors. However, this study was based on secondary data collected from the United States. Since such secondary data were not available in India, same was not applicable in the Indian context. However, we got some idea about what factors could influence the drug prescribing behavior of doctors. Through the study of Denig and Haaijer-Ruskamp (1994), we also came to understand that for common diseases, doctors use “thumb of rule practices” and prescribe traditional age old medicines rather than trying any new medicines. Thus, choice behavior was highly contextual. Not only this, the results of a study by Segal and Wang (1999) indicated that several demographic variables like age, experience or recency of graduation were also found to influence doctor’s drug prescribing behavior. Furthermore, in a study by Brett, Burr, and Moloo (2003), some ethical concerns regarding gift schemes were also shown by the doctors. Doctors felt that in terms of promotion, giving free samples of the drug was a more appropriate practice, as it reduced the cost of treatment at least for some patients. Furthermore, Symm, Averitt, Forjuoh, and Preece (2006) also found in a study that it also helped in checking toleration level of the drug, so that dosage can be fixed accordingly.

As we were looking at factors in developed markets, we found that interestingly health insurance policies also impacted drug prescribing behavior of doctors. A study by Gonul, Carter, Petrova, and Srinivasan (2001) indicated that doctors prescribed more expensive medicines to patients when they have private health insurance coverage rather than medicare coverage. In yet other studies, we found significant role played by opinion leaders in drug prescribing behavior. For example, in a study by Nair, Manchanda, and Bhatia (2010), we observed that when government introduced new healthcare policies, doctors relied on the opinion of highly reputed experts in their field to decide on new drugs which suited the new regulations formed by the government. Junior doctors met these experts in conferences or read their interview in trade journals to form their opinion. Similarly, the Manchanda and Chintagunta (2004) study reflected that constant reminders by sales representatives influenced doctors’ drug prescribing behavior. On similar lines, we also identified from Denig’s (1994) study the role of patients’ perspective and clinical trials.

This process took us about 1 month. By the beginning of April 2014, we had some idea of what factors could influence the drug prescribing behavior of doctors. One of our ex-student who was working with a pharmaceutical company was aware of our project. Nevertheless, she got confused when we informally shared with her our findings from the review of literature stage. The ex-student asked “now when you know what factors are important for doctors while prescribing a drug, why to conduct the study again?” We then explained to her that marketing drugs could vary from country to country. Since all these studies were conducted in developed markets and moreover were based on secondary data, direct conclusions cannot be drawn for the Indian market. We further explained to her that it would be ideal to conduct a qualitative research like in-depth interviews and then combine the findings of it with the review of literature to derive the important items for our study.

Exploratory Research: In-Depth Interviews

Continuing further, the next step was to conduct in-depth interviews with a few doctors. Robson and Foster (1989) emphasized on the use of in-depth interviews when the topic is personal and sensitive. In-depth interview is also preferred when depth and comprehensive information is required. The objective of the in-depth interview was to not only generate some additional variables influencing pen habit of doctors but to also validate the variables which have been obtained from the past review of literature. We started with preparation of open-ended questionnaire for in-depth interviews. Some of the questions in our final draft for the in-depth interviews were as follows:

  • When a patient comes to you, (a) you first listen to the patient for symptoms, (b) examine/screen the patient, (c) take history if any, (d) come to a decision on the disease, (e) think for the drugs or the molecule, (f) what next (how do you decide on the brand?)
  • Today, the entire pharmaceutical market is overcrowded. Pharmaceutical companies are promoting the same molecule or product to you under different brand names (me too). How do you select a particular brand or brands for a new product launched?
  • Sir, when you prescribe a B complex C combination, perhaps the name Becosules comes to your mind. Again when you may prescribe an antibiotic, perhaps the name of Cifran comes to your mind. What is the secret of these brands with you?
  • When you prescribe a brand of drug for a disease, what is it that comes to your mind? Is it the pharmacology of the product, clinical trial records of the product, presentation by a pharmaceutical sales representative/officer (PSO), cost, or anything else? Briefly explain.

Interviews with 10 renowned physicians across three hospitals of North India were conducted. These 10 physicians were selected based on a random draw from a list of physicians which was provided to us by a medical insurance company. These physicians were first consulted through emails for prior appointment. Then, we visited these physicians. We asked the above-mentioned questions and noted their responses. For example, one of the doctors from a leading hospital in New Delhi, India, told that generally he does not blindly rely on what PSOs of a pharmaceutical company are telling about the drug. He refers to his colleagues and friends in other hospitals prior to making a brand decision. Furthermore, he also relies on general reputation of the pharmaceutical company. Elaborating his point, the physician further explained that he learned about mal-manufacturing practices adopted by Ranbaxy in one of its plant in Baddi. He has thereafter stopped relying on Ranbaxy drugs.

Another leading physician from Gurgaon, India, mentioned that for him, cost of the drug is very important, as many middle income group patients visit him. Thus, if he has to choose between a very high quality expensive drug and a reasonable quality lower price drug, he would choose the second option. Furthermore, a newly appointed doctor from a reputed hospital in East Delhi mentioned that since she is not much experienced as a practitioner, she relies on the opinion of experts and senior doctors in the field before prescribing any new drug. Furthermore, she also conducts clinic trials before giving medicine to patient. Another experienced doctor said that in his 15 years of experience, he has realized many drugs cure the patient rapidly, which also enhances reputation of the doctor prescribing the drug, but in the long run, few such drugs have severe side effects. He explained one of his experiences. A patient visited him few months back with complaint of severe back pain. He wanted immediate relief. The drug that the doctor prescribed did more harm than good. Although the medicine was quick enough to cure his pain problem, it gave him nausea feeling all the time, which continued for 3 months even when the drug was discontinued. Since then he decided that for him, most important criteria of selecting any drug would be side effect or adverse effect of the drug. One of the doctors we interviewed was a Chief Medical Officer of a renowned hospital in Punjab, India. The doctor said that his decision is highly contextual. For patients whom he knows for several months, and knows are loyal to him, he has different approach compared to new patients. Any new drug, he said, he would test on his already existing patient given that the drug is from a reputed pharmaceutical company and sales representative has emphasized on the drug through several visits. This he does as some response has to be given to pharmaceutical companies as well regarding efficacy (effectiveness of the drug). But for new patients, his first concern is about efficacy of the drug and how rigorously clinical trials have been conducted by the pharmaceutical firm. One of the very young physicians in a South Delhi hospital said, “It depends on the convincing power of the sales representative.” Elaborating he said, if sales representative gives detailed description of the new drug and also provide generous free samples, he is not hesitant in trying new drugs.

Thus, after the in-depth interviews, we were able to identify several variables which influenced the drug prescribing behavior of doctors.

The next stage involved developing the questionnaire. For this, the variables obtained from the review of literature and those obtained from the qualitative research, that is, in-depth interviews, were combined. Two MBA students who were not part of the study were told about the objective of the study and were asked to combine the list of variables obtained from the review of literature and in-depth interview. This enhanced the validity of the study. Finally, 13 items which measured the variables obtained from the review of literature and in-depth interviews were developed, such as efficacy of the drug, opinion of the co-worker, and so on (see Table 1 for the list of items). Now the next challenging task was to investigate whether the same factors influenced the pen habits of other doctors at large and, second, whether these numerous factors influencing doctors’ decision could be summarized into a few major factors. For example, one doctor relied on the opinion of senior doctors, while the other relied on opinion of co-workers. There seems to be a common underlying factor of “opinion of others,” which influences drug prescribing behavior of the two doctors. Thus, can the two factors, that is, opinion of co-workers and opinion of experts be put under the same category say “opinion.” To achieve these objectives, we decided to undertake a large-scale questionnaire-based survey followed by an exploratory factor analysis to summarize the data into few categories/factors which influenced pen habit of doctors.

Table 1. Provides the list of basic information questions.

Items/questions

Scale

Proven efficacy of the drug

SA A InD D SD

Proven toleration profile of the drug

SA A InD D SD

Knowledge about the adverse effects of the drug

SA A InD D SD

Cost associated with the drug

SA A InD D SD

Use by opinion lead

SA A InD D SD

Use by co-worker

SA A InD D SD

Detailing by the PSO

SA A InD D SD

Promotion material provided by the pharmaceutical company

SA A InD D SD

Frequency of visit of the PSO

SA A InD D SD

Free samples provided by the company

SA A InD D SD

Proven reputation of the pharmaceutical company

SA A InD D SD

Proven reputation of the drug

SA A InD D SD

Proven clinical trials conducted by the pharmaceutical company

SA A InD D SD

PSO: pharmaceutical sales representative/officer.

Measurement and Scaling and Questionnaire Design

In the next stage, we developed a structured questionnaire. The questionnaire had three parts, namely, the basic information questions, the classification questions, and the identification questions. Questions pertaining to basic information are those relating to the 13 items obtained from review of literature and in-depth interviews; questions relating to the classification questions were those pertaining to demographic questions such as age, academic qualifications, and so on; and questions relating to identification were name, email address, phone number, and so on. The basic information questions were measured using 5-point Likert scales—1 being Strongly Agree (SA), 2 being Agree (A), 3 being Indifferent (InD), 4 being Disagree (D), and 5 being Strongly Disagree (SD). The physicians were asked to rate the basic information questions. Using past literature, the questionnaire was worded as follows:

Dear Doctor, in the section below is a list of few items which may be related to your pen habits. Please be kind enough to tick the appropriate option for each item (where SA is Strongly Agree, A is Agree, InD is Indifferent, D is Disagree and SD is strongly Disagree).

When you prescribe a Drug to a patient certain brands/or may be a single brand comes to the top of your mind immediately because of the …

Sampling Process and Sample Size

Now came the most demanding and tedious task of data collection at large scale. We had to decide on the sampling technique, which could be used to identify the cities, the hospitals, and the respondents. According to Malhotra (2009) and Kalton (1983), if a study is exploratory in nature, the population is largely homogenous, and also both time and cost of data collection are limited, a nonprobability sampling technique like judgmental sampling can be used. Since the present study was exploratory in nature and only general physicians were to be considered and time and cost were also limited, nonprobability sampling was preferred. Kalton (1983) additionally suggested that such a nonprobability sampling technique where experts’ opinions are considered for selecting a representative sample is a judgmental sampling technique. Furthermore, for selecting the cities, hospitals, and more importantly the doctors, we took the help of an expert, Mr Rathin Ahuja, who was a North Zone Regional Manager with a leading pharmaceutical company in India and was associated with pharmaceutical industry for more than 30 years. Based on Mr Rathin’s advice, we started first by exploring major cities in North Zone of India. Using the Census Survey (India) of 2011, we identified all the major cities of Northern India which had a population of at least two million. Three cities, namely, Delhi, Lucknow, and Kanpur, qualified to be in this category. In the next stage, we obtained the list of all the hospitals across these three cities. This list was obtained from a major medical insurance service provider known to Mr Rathin. The list also provided information on the patient carrying capacity of these hospitals. From this list, 40 hospitals were identified which had a bed capacity of 100 or more. In the third stage, we checked the website of all these hospitals to identify the number of Physicians or Internal Medicine doctors. A total of 213 doctors were identified.

Next, we sent emails and postal mails to all these 213 doctors inviting them to participate in the study. The invitation clearly explained the objective of the study. Out of these 213 doctors, 151 accepted the invitation which they confirmed by either email or postal mail. These 151 doctors were spread across 27 hospitals across the three cities, namely, Delhi, Lucknow, and Kanpur.

Finally, we started the data collection process at large scale, which went on for one and a half months, after which we returned with 151 filled-in questionnaires.

Plan of Data Analysis: Quantitative

We decided that for the data on the classification questions, descriptive statistics will be used. Also, the data on the 13-item instrument, which we have developed by combining the review of literature and in-depth interview results, will be submitted for exploratory factor analysis. Factor analysis is a data reduction and a data summarization correlation based technique. Groups of items which are similar load on one factor and are different from groups of items which have loaded in another factor. The primary objective of factor analysis in the present study was to obtain the underlying dimensions or factors. We also decided to perform reliability test on the final instrument and the factors obtained. To conduct descriptive statistics, exploratory factor analysis, and reliability analysis, we used IBM-SPSS Ver. 21.0. In SPSS, for conducting exploratory factor analysis, we used the option of principal component analysis with varimax rotation.

Results and Analysis
Descriptive Statistics: Classification Information

We next coded the data. The classification questions revealed that out of 151 physicians, 58 were women and the remaining 93 were male. Also, the average physician age was 39 years and the average years in practice was 14 years (excluding internships done by the physicians at different time periods). Also, 94 physicians were at least MBBS and MD, while the remaining 57 were MBBS. We next conducted exploratory factor analysis on the 13 basic information questions (or items).

Exploratory Factor Analysis

Next factor analysis with varimax rotation was conducted on the sample dataset. One of the primary requirements of conducting factor analysis is to maintain the sample size criteria. As factor analysis was conducted on the 13 items, the required sample size was in excess of 130. Hair, Anderson, Tatham, and Black (1998) proposed that for exploratory factor analysis, an item to respondent ratio of 1:10 was a requirement. We were satisfied in this regard as the sample size of our study was 151. In the next step, exploratory factor analysis was conducted using varimax rotation. Varimax is a rotation technique which assumes that the factors are not correlated. There are other rotation techniques also, such as oblique rotation which assumes that the factors are un-correlated. Rotation in factor analysis is also used because the rotated factor matrix is easier to interpret. Once factor analysis is conducted as a first step, the Kaiser–Meyer–Olkin (KMO) test of the factor analytic model and the Bartletts’ test of sphericity should be checked. From Table 2, it is evident that the KMO is 0.640 and the chi-square in the Bartlett’s Test of sphericity is significant. The significance of the chi-square in Bartlett’s test indicates that the correlation matrix is not an identity and factor analysis can be conducted. However, the KMO statistic which indicates the measure of sampling adequacy should be in excess of 0.70, which was not the case in our study. We needed to explore the reason behind this.

Table 2. KMO and Batlett’s Test.

Kaiser–Meyer–Olkin (KMO) measure of sampling adequacy

0.640

Bartlett’s test of sphericity

Approx. chi-square

504.664

df

78

Sig.

0.000

When a factor analytic model is conducted, one of the outcomes is the Anti-Image Correlation matrix. In the anti-image correlation matrix, the diagonal consists of individual-item KMOs, and the off-diagonals are negative partial correlations. In this matrix, a researcher should try to identify any variable/variables in the diagonal which has/have a KMO of less than 0.50. It was observed that one item, which was labeled as “cost” had a KMO of less than 0.50 (see Table 3). As the item “cost” had a KMO of less than 0.50, this item was removed from subsequent analysis. Thus, 12 items were left. Once the item “cost” was removed, factor analysis was again conducted for the second time (we call this the second run and the previous one as first run), and it was observed that the KMO of the model was in excess of 0.70 and the chi-square of the Bartlett’s Test of Sphericity was also significant (see Table 4). Also, the anti-image correlation matrix revealed that all the 12 items had KMO in excess of 0.60.

Table 3. Anti-image correlation matrix.

Efficacy (1)

Toleration (2)

Knowledge about adverse (3)

Cost (4)

Reputation of company (5)

Reputation of drug (6)

Clinical trials company (7)

Detailing (8)

Free samples (9)

Use by co-worker (10)

Use by opinion lead (11)

Promotion material (12)

Frequency of visit by PSO (13)

1

0.710

-0.299

-0.374

0.124

0.037

-0.039

-0.240

-0.029

0.003

0.089

-0.086

0.113

-0.035

2

 

0.690

-0.149

-0.324

0.091

-0.215

-0.119

0.097

0.150

0.125

-0.152

0.096

-0.009

3

 

 

0.657

-0.023

0.057

-0.090

0.073

-0.176

-0.038

-0.067

0.171

-0.049

0.142

4

 

 

 

0.420

-0.055

0.063

0.024

-0.138

-0.119

0.036

-0.058

-0.021

-0.029

5

 

 

 

 

0.630

-0.579

-0.307

-0.181

-0.025

0.132

-0.048

-0.139

-0.009

6

 

 

 

 

 

0.642

-0.043

0.138

-0.062

-0.070

0.119

-0.060

-0.099

7

 

 

 

 

 

 

0.709

-0.103

0.123

-0.153

0.096

0.053

0.067

8

 

 

 

 

 

 

 

0.732

-0.218

-0.079

0.137

-0.151

-0.077

9

 

 

 

 

 

 

 

 

0.814

-0.232

-0.053

-0.152

-0.234

10

 

 

 

 

 

 

 

 

 

0.626

-0.593

-0.044

-0.075

11

 

 

 

 

 

 

 

 

 

 

0.623

-0.096

0.127

12

 

 

 

 

 

 

 

 

 

 

 

0.833

-0.296

13

 

 

 

 

 

 

 

 

 

 

 

 

0.787

Table 4. KMO and Batlett’s Test.

Kaiser–Meyer–Olkin measure of sampling adequacy

0.703

Bartlett’s test of sphericity

Approx. chi-square

477.610

df

66

Sig.

0.000

Next, we looked into the communalities(see Table 5) which indicates the amount of variance in each variable which is indicated by extracted factors. Researchers observe the communality values to identify item/s which has values much less compared to the average. Such items should be removed from further analysis. However, in this study, we encountered no such values. One of the most important outputs of factor analysis is the total variance explained (see Table 6). From this table the extracted number of factors is determined by the eigenvalue criterion, that is, we consider as many number of factors which have eigenvalues in excess of 1. It was observed that the extracted number of factors is four as the first four factors had eigenvalues in excess of 1.00. Also, the four factors account for 67.179% of variance. In factor analysis, the first factor always accounts for most of the variance followed by the second, third, fourth, and so on. Next, we observed the rotated component matrix. It was observed (see Table 7) that four items have loaded in Factor 1, while three items each have loaded in the second factor and third factor. Also, two items have loaded in fourth factor. An item loading is the correlation between the item and the factor. An item will load in multiple factors but will be considered only belonging to that factor where its loading is above 0.40. Also, the item should not have such high loadings with any other factor. Otherwise, this will lead to a problem of high cross-loading of an item across multiple factors.

Table 5. Communalities.

Extraction

Efficacy

0.691

Toleration

0.583

Knowledge about adverse

0.745

Reputation of company

0.789

Reputation of drug

0.717

Clinical trials company

0.547

Detailing

0.597

Free samples

0.655

Use by co-worker

0.786

Use by opinion lead

0.828

Promotion material

0.605

Frequency of visit by PSO

0.518

PSO: pharmaceutical sales representative/officer.

Table 6. Total variance explainedx

Factors

Initial eigenvalues

Extraction sums of squared loadings

Rotation sums of squared loadings

Total

% of variance

Cumulative %

Total

% of variance

Cumulative %

Total

% of variance

Cumulative %

1

2.995

24.956

24.956

2.995

24.956

24.956

2.469

20.571

20.571

2

2.534

21.114

46.070

2.534

21.114

46.070

2.022

16.850

37.421

3

1.402

11.686

57.756

1.402

11.686

57.756

1.889

15.741

53.162

4

1.131

9.424

67.179

1.131

9.424

67.179

1.682

14.017

67.179

5

0.795

6.625

73.804

 

 

 

 

 

 

6

0.644

5.364

79.168

 

 

 

 

 

 

7

0.552

4.596

83.764

 

 

 

 

 

 

8

0.506

4.215

87.979

 

 

 

 

 

 

9

0.429

3.575

91.554

 

 

 

 

 

 

10

0.413

3.443

94.997

 

 

 

 

 

 

11

0.340

2.832

97.829

 

 

 

 

 

 

12

0.261

2.171

100.000

 

 

 

 

 

 

Table 7. Component matrix.

Factor

1

2

3

4

Detailing

0.737

Free samples

0.737

Promotion material

0.719

Frequency of visit by PSO

0.677

Reputation of company

0.834

Reputation of drug

0.818

Clinical trials company

0.669

Knowledge about adverse

0.831

Efficacy

0.776

Toleration

0.602

Use by opinion lead

0.905

Use by co-worker

0.827

PSO: pharmaceutical sales representative/officer.

Reliability Analysis

After the factor analysis was performed, a reliability analysis was conducted to test the internal consistency of the instrument. For the 12-item instrument of the study, evidence of strong reliability was suggested as Cronbach’s alpha was in excess of 0.70. Next, reliability analysis was performed for each of the factors separately. Nevertheless Cronbach’s alpha for the fourth factor could not be determined as only two items loaded in this factor. Cronbach’s alpha can only be determined for a factor having three or more items. For the fourth factor, reliability statistics were obtained using correlation between the two items representing it. The reliability statistics for the four factors and the 12 items considered together were all in excess of 0.70 (see Table 8). Once the factor structure is determined, the next important task is to name the factor based on the items which have loaded in each of the factors.

Table 8. KMO and Batlett’s Test.

Reliability statistics

Items/questions

Details

0.745

12

Twelve items (Cronbach’s alpha)

0.733

4

Factor 1 (Cronbach’s alpha)

0.743

3

Factor 2 (Cronbach’s alpha)

0.725

3

Factor 3 (Cronbach’s alpha)

0.783

2

Factor 4 (Correlation between the two items of Factor 4)

Naming the Factors

The first factor consisted of four items, namely, detailing, free samples, promotion material, and frequency of visit by PSO. This factor was named as promotion. This factor explained the maximum variance, that is, contributed most to the pen habit of doctors in the northern region of India. We thus concluded that pharmaceutical firms need to train and invest a lot in sales representative factor by promoting free samples and increasing the frequency of visits by PSOs. The second most important factor was termed as reputation and consisted of factors like reputation of drug, reputation of company, and clinical trials conducted by pharmaceutical firms. Thus, in a pharmaceutical firm, the R&D team needs to be communicated about the significance of trials that doctors perceive and that team should be very careful and conduct such trials in the best possible manner. He further decided to discuss with public relations and marketing department as to what measures should be taken to enhance overall reputation of the firm. He also decided to invest in feedback surveys about individual drugs so that he could know whether individually some drugs are not accepted well in the market and plan remedy measures. The third factor was termed as drug traits and consisted of items like knowledge about adverse effects, tolerance, and efficacy (effectiveness) of the drug. The fourth factor was named as opinion.

Conclusion

Through this project, we were able to get some clarity regarding what doctors exactly look into while prescribing drugs. However, we were unable to understand as to why sales promotion emerged as the most critical factor and drug trait was not the most important but third most important factor affecting pen habit of doctors.

After further exploring the industry, we realized that this could happen since sales representatives are the primary source of product information in a market which is cluttered with different branded and unbranded drugs. Doctors generally operate under time constraints. They themselves cannot evaluate each and every drug available in a market. Furthermore, sales representatives not only provide them free samples to be tested but also keep updating them about latest happenings in the industry. Although trade journals are better source for such information, sales representatives can give the insider view. In other words, sales representatives are the most critical connection between pharmaceutical companies and doctors. Furthermore, these agents, if they have good influencing power, could convince doctors about price, quality, tolerance, efficacy, and other such traits associated with drug or the company.

Overall, we were convinced with the study. We next looked for a prestigious conference where this article could be presented so that senior managers from large pharmaceutical firms could benefit from the study.

Exercises and Discussion Questions
  • What is in-depth interview? Was it necessary for this study?
  • What is judgmental sampling? Explain in the context of this case study?
  • What are the elements of research design?
  • What is exploratory factor analysis? Why is factor analysis conducted in the present study?
  • Explain how in this project a four-factor solution was achieved. It could have been a two- or five-factor solution also.
  • Explain the significance of naming of factors. How the factors in this case have been named?
Further Reading
Malhotra, N. (2009). Marketing research: An applied orientation. Upper Saddle River, NJ: Pearson Education.
Field, A. (2013). Discovering statistics using IBM SPSS statistics. London, England: SAGE.
References
Brett, A. S., Burr, W., & Moloo, J. (2003). Are gifts from pharmaceutical companies ethically problematic? Archives of Internal Medicine, 163, 22132218.
Denig, P. (1994). Drug choice in medical practice: Rationales, routines, and remedies (Doctoral dissertation, University of Groningen). Retrieved from https://www.rug.nl/research/portal/publications/drug-choice-in-medical-practice(8ddacf01-2f31-4ac4-9399-ad5fb90a091c).html
Denig, P., & Haaijer-Ruskamp, F. M. (1994). “Thinking aloud” as a method analyzing the treatment decisions of physicians. European Journal of Public Health, 4, 5569.
Gonul, F. F., Carter, F., Petrova, E., & Srinivasan, K. (2001). Promotion of prescription drugs and its impact on physicians’ choice behavior. Journal of Marketing, 65, 7990.
Hair, J. F., Anderson, R. E., Tatham, R. L., & Black, W. C. (1998). Multivariate data analysis. Upper Saddle River, NJ: Prentice Hall.
Kalton, G. (1983). Introduction to survey sampling. London, England: SAGE.
Malhotra, N. (2009). Marketing research: An applied orientation. Upper Saddle River, NJ: Pearson Education.
Manchanda, P., & Chintagunta, P. K. (2004). Responsiveness of physician prescription behavior to sales force effort: An individual level analysis. Marketing Letters, 15, 129145.
Nair, H. S., Manchanda, P., & Bhatia, T. (2010). Asymmetric social interactions in physician prescription behavior: The role of opinion leaders. Journal of Marketing Research, 47, 883895.
Robson, S., & Foster, A. (1989). Qualitative research in action. London, England: Hodder & Stoughton.
Segal, R., & Wang, F. (1999). Influencing physician prescribing. Pharmacy Practice Management Quarterly, 19, 3050.
Symm, B., Averitt, M., Forjuoh, S. N., & Preece C. (2006). Effects of using free sample medications on the prescribing practices of family physicians. Journal of the American Board of Family Medicine, 19, 443449.
Venkatraman, S., & Stremerch, S. (2007).The debate on influencing doctors’ decisions: Are drug characteristics the missing link? Management Science, 53, 16881701.

Copy and paste the following HTML into your website