Journal of General Internal Medicine

, Volume 23, Issue 3, pp 236–241

Patients’ Beliefs and Preferences Regarding Doctors’ Medication Recommendations


    • Meyers Primary Care Institute
    • Baystate Medical Center
  • Kathleen M. Mazor
    • Meyers Primary Care Institute
    • University of Massachusetts Medical School
  • Vanessa Meterko
    • Meyers Primary Care Institute
  • Katherine Dodd
    • Meyers Primary Care Institute
  • James Sabin
    • Harvard Pilgrim Healthcare
Original Article

DOI: 10.1007/s11606-007-0470-3

Cite this article as:
Goff, S.L., Mazor, K.M., Meterko, V. et al. J GEN INTERN MED (2008) 23: 236. doi:10.1007/s11606-007-0470-3



An estimated 20–50% of patients do not take medications as recommended. Accepting a doctor’s recommendation is the first step in medication adherence, yet little is known about patients’ beliefs and preferences about how medications are prescribed.


To explore patients’ beliefs and preferences about medication prescribing to understand factors that might affect medication adherence.


Fifty members from 2 health plans in Massachusetts participated in in-depth telephone interviews. Participants listened to an audio-vignette of a doctor prescribing a medication to a patient and were asked a series of questions related to the vignette. Responses were reviewed in an iterative process to identify themes related to participants’ beliefs and preferences about medication prescribing.


Participants’ beliefs and preferences about medication prescribing encompassed 3 major areas: patient–doctor relationships, outside influences, and professional expertise. Important findings included participants’ concerns about the pharmaceutical industry’s influence on doctors’ prescribing practices and beliefs that there is a clear “best” medication for most health problems.


Patients’ beliefs and preferences about medication prescribing may affect medication adherence. Additional empiric studies that explore whether doctors’ relationships with pharmaceutical representatives impact medication adherence by affecting trust are indicated. In addition, it would be worthwhile to explore whether discussions between patients and doctors regarding equipoise (no clear scientific evidence for 1 treatment choice over another) affect medication adherence.


medication adherencequalitativepatient preferencescommunicationmedication prescribingpharmaceutical detailingshared decision making


Approximately 20–50% of patients do not take medications as prescribed.1 Medication nonadherence can lead to disease progression, adverse drug events, additional prescriptions, unnecessary hospitalizations, and increased health care costs.24 Patients may not take medications out of fear of drug interactions, perceived lack of effectiveness, adverse effects, misunderstanding regarding necessity, or concerns about costs.5 Patients who report better general communication with their doctor, better instructions on how to take a medication, and who receive more medication information are more likely to take medications as prescribed.68 The Agency for Healthcare Research and Quality has recommended counseling techniques to improve medication adherence and reduce medical errors,9 but doctors’ medication-counseling habits have been shown to be suboptimal.10

The first conversation between a doctor and a patient about a new medication is not only an opportunity for the doctor to communicate information about the medication but also a chance to explore patient values that may influence whether a patient is receptive to a medication. A patient’s existing beliefs and preferences about medication prescribing may affect satisfaction with communication about a medication, potentially influencing medication adherence. For example, if a patient believes a prescribing decision is driven by motives other than optimizing his or her individual health, there may be a higher risk for nonadherence. To better understand how patients view the prescribing process, we explored patients’ beliefs and preferences through in-depth interviews.


Study Design and Sample

The study was conducted using in-depth telephone interviews with patients recruited from 2 New England health plans as part of a broader study exploring patient responses to cluster randomized trials. We chose qualitative interviews because of the dearth of existing information on patients’ views and to allow in-depth exploration of beliefs and preferences. The Institutional Review Boards of the participating sites approved the study.


We randomly sampled adult patients at participating health plans and sent 1,000 invitation letters to recruit 50 patients. The letters described the study’s purpose and offered a $20 incentive. Interested patients called and scheduled an interview time; no follow-up contact was made with patients who did not respond. Once an interview was scheduled, we sent a confirmation letter and obtained informed consent.

Audio-vignette Development and Conduct of Interviews

We designed an audio-vignette to give study participants a common point of reference when answering the interview questions.11,12 The vignette was piloted with a convenience sample of colleagues to be sure the information presented in the vignette was clear. In the 3-minute vignette, actors portrayed a doctor and patient discussing a new prescription for either depression or hypertension (Appendix). Two vignette versions were used so that comments were not restricted to a single condition; version was assigned at random just before the interview.

A trained interviewer (Meterko) conducted interviews via telephone. The interviewer introduced herself, reiterated the purpose of the interview, confirmed consent to participate, and then played the vignette. After the vignette, the interviewer posed open-ended questions using an interview script as a guide (Appendix). The interviewer followed the script verbatim deviating as needed to clarify participant comments. Interviews lasted approximately 30 minutes, and demographic information was collected at the end of interviews. We audio-taped and professionally transcribed all interviews.


The study goals and interview questions provided an initial organizing framework for results. One investigator (Goff) read a set of 5 transcripts, generating a primary list of codes. A second investigator (Mazor) read 3 additional transcripts and recommended modifications to the primary code list. After the primary code list was developed, all transcripts were reviewed iteratively, and the code list was revised as new concepts were discovered. The research team met to review the code list and to adjudicate any variation in transcript coding. This process continued until the 4 members of the research team agreed that the code list captured all major issues raised by participants. The individual codes were then grouped into themes. Microsoft Word and the Statistical Package for Social Sciences were used for coding and organizing results. Theoretical saturation was reached, meaning no new concepts emerged in the last 20 transcripts reviewed.

Two research team members (Meterko and Dodd) coded a subset of transcripts (N = 12) to assess inter-rater reliability. Two-hundred and forty-three out of 300 codes identified in these transcripts matched, yielding 81% agreement.


We included all 50 participants in the analysis. Participant demographic information showed the following: 56% female, 52% aged 50–70, 96% Caucasian, and 68% held at least a college degree. Responses for both beliefs and preferences regarding medication prescribing encompassed 3 major areas: (1) patient–doctor relationships, (2) outside influences, (3) professional expertise. Each major area contained related themes and subthemes (Table 1), which are described in detail. Representative quotes are provided to illustrate key components of the findings.
Table 1

Major areas, Themes, and Subthemes Pertaining to Patients’ Beliefs and Preferences Regarding How Doctors Decide to Recommend a Medication

Major areas



Patient–doctor relationship


Communication Honesty

Familiarity with patient

Shared decision making


Outside influences

Pharmaceutical detailing

Positive view

Negative view


Favorable view

Unfavorable view

Drug equivalency

HMO Regulations



Professional expertise

Medication knowledge


Side effects

Lifelong learning


Medical journals

Patient–Doctor Relationships

Participants discussed patient–doctor relationships in response to both “belief” and “preference” questions. Trust emerged as a key theme related to patient–doctor relationships. Participants felt that when 1 person holds more information (the doctor) than the other (the patient), trust was necessary for accepting a medication recommendation. Some participants felt trust existed because of respect for the doctor’s position, whereas others described trust evolving over time with a particular doctor and relating to that doctor’s honesty.

Familiarity with the patient was discussed independent of trust in response to both prescribing “belief” and “preference” questions. Participants wanted to feel that their doctor was familiar not only with their medical problems but also their health beliefs and home and work circumstances. They hoped that this familiarity would inform the doctor’s recommendation.

Well I had to change doctors a year ago because I changed to Plan X, and I was not happy about that because I was with a doctor that I did trust; I was with him for 25 years. He knew how I felt about going on medications, which is I only do it as a last resort and he knew that I expected knowledgeable answers from him and that if he didn’t have the answers that was okay, but I expected him to say I really and truly don’t know that but I will find out… He was very honest and I really respected that about him. I am with a new doctor who I don’t trust yet and I haven’t decided whether or not I like her…as a person I like her….my other doctor knew my fears [of being on a machine for asthma]… I haven’t been with her long enough and so she doesn’t know what my fears are.

Patients desired shared decision making in the context of medication prescribing. Specifically, participants stated that they wanted to know their medication options and understand the risks and benefits of each option.

Well doctors should consider the patient as a whole of course and they should respect patient’s wishes. If they say he or she doesn’t like to take a medication for some [amount of] time, if she does not want to be on some kind of medication right now, then they should give them some time. Doctors shouldn’t force the medication on the patient. That’s what I feel.

Communication was a key subtheme in discussions related to patient–doctor relationships. Participants felt that good communication contributed to their comfort with a doctor’s medication recommendation, and lack of communication was felt to reduce confidence in medication recommendations.

They [the doctors] count on my opinion. They look at it and say okay is this working or not? How do you feel about it? So that was the dialogue that I have had with my doctors. It makes sense, so that [dialogue] is what I would be looking for.

Outside Influences

Participants believed that outside influences played a role in doctors’ medication recommendations and generally preferred that these outside influences play as small a role as possible. Some participants viewed the effect of pharmaceutical detailing on prescribing in a positive manner, citing free samples and easy access to information for doctors. Others viewed pharmaceutical detailing as having a negative influence on a doctor’s decision making regarding medication prescribing. These participants expressed concerns about possible financial or social gain for the doctors and a potentially biased view point of the pharmaceutical representatives.

[My doctor decides on a medication] by how good looking the drug sales girl is that sells it to him. That’s what I think. Yeah, he’s a single doctor and I get that feeling from him. I don’t think it’s… I hope he doesn’t sacrifice his professionalism for that. But one never knows.

Participants also discussed outside influences related to cost. Participants’ varied on preferences for how medication costs should influence a doctor’s recommendation about a medication. Some expressed a favorable view, feeling they would be grateful for the opportunity to save money, whereas others expressed an unfavorable view, feeling that the best medication for the situation should be offered regardless of cost. Participants generally believed that the doctor had the knowledge to recommend 1 medication that was the “best” for a given situation.

Participants believed Health Maintenance Organization (HMO) regulations and formularies might play a role in how doctor’s decided what medication to recommend. Preferences for this influence varied, with some participants accepting it and others preferring such regulations not influence prescribing decisions.

I think that people wouldn’t like to think that their doctors were basing it [medication recommendation] on what is cheaper but I think with HMOs that tends to happen a lot more often now.

Transparency regarding medication costs was preferred by participants. Participants preferred to know all the medication options as well as the cost of each medication so that they could participate in making informed decisions. Some participants felt their own doctor did share this information, whereas others did not.

I think that he should probably make it known to the person that this is the cost factor between the number of drugs that would work for you and you decide…

Professional Expertise/Experience

Participants believed and preferred that professional expertise and experience contribute to a doctor’s medication recommendation. Knowledge about a specific medication was believed and preferred to be a factor in making a medication recommendation. Participants expected doctors to be aware of medication effectiveness and side effects and to use this knowledge when making medication recommendations. Participants’ belief that doctors know which medication is the best for a given situation arose in discussions pertaining to professional expertise as well as in the discussions of cost.

I think that they hopefully would have enough information to tell them what is the most effective drug. If one is just as effective as another, then I would think they would know the side effects even though something might be therapeutically effective maybe one might have more troublesome or uncomfortable side effects and they would then prescribe the one that is equally effective but with less side effects.

Participants believed that doctors stayed familiar with current medication information through lifelong learning strategies such as journal reading and conferring with colleagues.

Well I assume my doctor’s advice would be based on reading and research that she had done and contact with other doctors in her field about those treatments or about those medications.


In this study, we used in-depth patient interviews to explore patients’ beliefs and preferences regarding how doctors prescribe a new medication. Our results suggest there may be modifiable entities influencing medication adherence that have not been previous described in detail. These entities include participant’s concerns about doctors’ relationships with pharmaceutical representatives and beliefs about “best” medications, which are discussed in detail.

Some participants expressed concerns that medication recommendations are influenced by doctors’ relationships with the pharmaceutical industry, specifically by the presence of pharmaceutical representatives in the doctor’s office. Prior studies suggest patients may not be fully aware that pharmaceutical companies give gifts to doctors.13,14 When patients were given examples of different types of gifts, patients felt small gifts or gifts that benefit patients were generally acceptable but gifts directly benefiting doctors were less acceptable.15 Doctors’ relationships with pharmaceutical representatives appear to influence prescribing practices16,17, and concerns exist related to gift-giving ethics.18,19 Recent lay publications have also raised ethical questions about some doctors’ relationships with the pharmaceutical industry.2022 Participants in our study described trust as important in accepting a doctor’s recommendation for a medication. Participants’ concerns that doctors’ medication recommendations are being influenced by pharmaceutical representatives rather than being based solely on patient needs may indicate a threat to patients’ trust in doctors. This finding implies a need to empirically evaluate the impact doctors’ relationships with pharmaceutical companies have on trust and medication adherence. To maintain patient trust, it may be advisable for doctors to disclose their practice’s policies regarding interactions with pharmaceutical representatives. This information could be provided in written details about the practice, e.g., “The doctors in this practice do not accept gifts from pharmaceutical representatives.” Providing this disclosure would be consistent with disclosures of financial relationships between doctors and pharmaceutical companies required in peer-reviewed publications and for many speaking engagements.

Participants believed that doctors have the information needed to determine which medication is safest and most effective for a patient in a specific clinical situation, revealing that participants were not aware of the lack of medication equivalency studies. Addressing equipoise (no clear scientific evidence for 1 treatment choice over another) is an important aspect of shared decision making.23,24 A familiar example of such a discussion might be about generic medication use. Patients may be more likely to accept a new medication recommendation if they actively participate in the decision process. However, conversations about differences between similar medications, such as “out of pocket” costs, rarely happen, although they are desired by patients.25 Further evaluation of how and when such discussions take place in practice, the adequacy of medical school curricula to develop the skills for such discussions, and the actual effect such practices have on adherence are necessary.

This study was conducted in the context of 2 managed care health plans, and thus findings may not be generalizable to patients who belong to other types of plans or who are uninsured. Further, most participants were Caucasian, so we do not know whether members of racial or ethnic minorities would have responded similarly. Prior work in medication adherence shows that adherence rates may be lower for certain ethnic minorities.2629 Differences in health beliefs and use of alternative therapies may explain some differences found in adherence rates for special populations.3032 Replicating this study in special populations could yield novel and clinically useful results. The vignette also may have primed participants to respond to questions differently than they would have without the vignette. Although unlikely to invalidate results, it may be useful for future studies to explore responses with and without vignettes.

Accepting a doctor’s recommendation is a necessary first step in the complex process of taking a medication as prescribed. This study’s findings suggest a need to rigorously evaluate how doctors’ relationships with pharmaceutical representatives impact patient trust and medication adherence in a diverse population. Studies evaluating correlations between a patient’s beliefs and attitudes about pharmaceutical representatives and the patient’s actual medication adherence could provide additional data in this area. In addition, it is important to evaluate how patients’ understanding of equipoise impacts medication adherence. Assessing whether medication equipoise discussions impact adherence and, if so, whether medical school curricula prepare students to discuss medication equipoise with patients would inform future recommendations for medical education curricula and for practitioners.


Funding was provided by AHRQ through a Supplement to the HMO Research Network Center for Education and Research on Therapeutics (HMORN CERT) U 318 HS010391-06S1.

Conflict of Interest

None disclosed.

Copyright information

© Society of General Internal Medicine 2007