Objective: To describe the frequency and content of physician discussions about the cost and acquisition of new medications.
Study Design: Qualitative and quantitative study combining patient and physician surveys with transcribed audiotaped office visits from 185 outpatient encounters with 15 family physicians, 18 internists, and 11 cardiologists in 2 Sacramento healthcare systems between January and November 1999, in which 243 new medications were prescribed.
Methods: Transcripts were qualitatively analyzed to describe conversations concerning the cost and acquisition of new medication prescriptions, frequencies of discussions were described, and multivariate logistic regression was used to examine the predictors of cost discussions.
Results: For one third of newly prescribed medications, physicians discussed an aspect of acquisition, including cost and insurance coverage for 12%, logistics of obtaining medications for 18%, and medication supply for 9%. Patients initiated cost or insurance concerns for fewer than 2% of the new medications prescribed. There were 5 discussions about patient out-of-pocket costs and 7 discussions offering tips for cost savings. In multivariate analysis, cost discussions were more likely to occur with patients earning less than $20 000 per year compared with those earning more than $60 000 per year, were less likely to occur with medications prescribed by family physicians or internists compared with cardiologists, and were less likely with medications prescribed to older patients.
Conclusions: Physician-patient discussions about new medication cost and other acquisition issues, especially medication affordability, occur infrequently. Because physicians may not recognize patients' financial impediments to acquiring medications, this issue should be raised when prescribing new medications.
(Am J Manag Care. 2006;12:657-664)
The enormous potential of medicine to cure and treat medical conditions is often unrealized when medications are not obtained or are not taken as directed because of cost or acquisition problems. High medication costs are strongly associated with medication underuse.1-7 One quarter of older patients cite cost as a factor for not taking their medications as prescribed.2 Patients also cite cost as a reason for not filling a new prescription.8 Patients exceeding drug benefit caps often underuse medications for chronic health problems such as hypertension, hyperlipidemia, emphysema, and asthma.6 Inconsistent use of essential medications, often related to increased prescription drug cost-sharing, has been shown to increase drug-related serious adverse events and emergency department visits.9
By physician and patient report, physicians infrequently discuss out-of-pocket costs with patients, despite reported patient desire to converse about the topic.10,11 Clinicians and patients cite discomfort and insufficient time as barriers to discussions.11,12 More discussions about cost are reported by patients who have no insurance,13 who are burdened by out-of-pocket costs,10 or who are seen in a community practice.10 Investigations have examined cost as a barrier to medication use, but only a couple of studies2,8 have examined the actual conversations occurring during physician-patient encounters. Most observational studies14-18 do not comment on discussions of cost or insurance, although some have shown infrequent communication about the cost of new medication prescriptions (in 11% of their sample)19 and antibiotics (in 0.5% of their sample).20
Other organizational factors may affect whether a patient is able to access medications.21,22 For example, patients fail to refill about 33% of antihypertensive medications,23 and patients lacking knowledge about refills or pharmacy contacts were twice as likely to seek refills from emergency departments.24 Although 72% of patients in a subsidized prescription program appropriately obtained medication refills, only 73% of those requesting refills obtained them on time.25 Therefore, conversations about refills may be important when prescribing chronic medications. Physicians in one study15 were observed asking patients about medication supply or quantity for 51% of current, past, or new medications, and physicians in another study17 made a mean of 2.6 statements per visit about supply and uttered a mean of 0.21 statements per encounter about refilling medications. These studies were not specific to new medications, for which patients may need the most instruction.
The actual content and predictors of observed discussions about cost and insurance have not been well explored, and studies of physician-patient encounters have not documented the frequency at which refills or other medication acquisition issues are discussed for new medication prescriptions, to our knowledge. This article examines whether and how physicians conduct discussions about cost when a new medication is prescribed. It also examines the frequency of discussion about medication acquisition. We expected low frequencies of counseling about cost and other acquisition issues but anticipated that discussion contents would cover several cost and insurance issues. We also hypothesized that more cost discussions would occur among patients with lower income levels and with less prescription drug coverage.
We analyzed data from the Physician Patient Communication Project, a study of outpatient visits to family physicians, internists, and cardiologists in 2 healthcare systems in Sacramento. The study was originally designed to investigate the relationship between request fulfillment and patient outcomes and was not specific to medications. In this study, 44 physicians were recruited, 21 from Kaiser Permanente (a group-model health maintenance organization with salaried physicians) and 23 from the University of California, Davis Medical Group (a multispecialty group practice with discounted fee-for-service compensation for physicians and salaried specialists). Study participants consisted of 15 family physicians, 18 internists, and 11 cardiologists. One family physician who prescribed no new medications was excluded. The study design was previously described.26 The University of California, Davis Institutional Review Board approved the overall study (protocol 992212), and the University of California, Los Angeles, Institutional Review Board approved this analysis (exemption 04-193).
Patients were recruited between January and November 1999. A research assistant reviewed appointment books 1 to 2 days before patient appointments and randomly sampled and telephoned patients to screen for study eligibility. Patients had to be 18 years or older, speak English, and have a new or worsening problem or be somewhat concerned about their health or about having a potentially serious undiagnosed condition. Of 4560 patients selected for telephone contact, the response rate was 32%; 68% (909 of 1332 eligible patients) of eligible patients were enrolled. Study participants provided informed consent and were compensated $10.
Among 860 of 909 patients enrolled in the study who had their visits successfully audiotaped and transcribed, physicians identified 270 patients as receiving new prescriptions via a postvisit survey. A review of 90 randomly selected transcripts indicated that physicians correctly identified 24 of 25 encounters in which new medications were prescribed. A new medication was defined as one that a patient had never before taken or one that was given for an acute condition or symptom (such as an antibiotic or analgesic). Therefore, medications taken in the past and new prescriptions belonging to the same class as a previously used medication were not included. We identified 185 encounters in which 243 new medications were prescribed.
Patient and Physician Characteristics
Patients provided information about their demographic characteristics, whether they had a prior visit with the physician, and about their out-of-pocket prescription drug costs. Physicians provided demographic information and characteristics of their practice and reported on the number of "old drugs" continued.
We used 185 transcripts in which new medications were prescribed to develop a coding framework to describe the content of conversations about new medications. Transcripts were reviewed to identify every piece of conversation relating to any newly prescribed medication. Each identified conversational element was assigned a code. The initial coding categories were based on existing literature16,17,20 and on clinical experience, and we used analytic induction to expand the categories in an iterative process until no further codes could be developed. We split, merged, and adjusted the coding categories27,28 until they were mutually exclusive and represented all communication about new prescriptions. Conversation corresponding to themes concerning medication cost and insurance, whether a medication was a generic or brand name drug, logistics of obtaining medication, medication supply (quantity of medication to be dispensed), and medication refills were assigned to relevant codes. Two coders worked together to develop the codes and to ensure reliable implementation. A third coder independently coded 29 (16%) of the transcripts. The overall mean κ was 0.90 (range, 0.79-0.98). A detailed description of the coding framework and its development is given elsewhere.29 ATLAS/ti 4.2 (Thomas Muhr Scientific Software Development, Berlin, Germany) was used to code the transcripts.
Each code developed inductively through qualitative analysis was assigned a variable corresponding to the conversational content. Using new medication prescriptions as the unit of analysis, we constructed a data set of communication behaviors about each new prescription, which allowed us to determine the occurrence of conversations about different themes. STATA 8.0 (StataCorp LP, College Station, Tex) was used to calculate descriptive frequencies of the communication topics by site, physician specialty, general medication class, over-the-counter (OTC) medications, and "as needed" (pro re nata [PRN]) medication status. A summary of any of the 5 acquisition topics was captured as "any acquisition communication." Whether the patient or physician initiated the conversation was also noted.
Medications were classified into general medication classes according to their purpose. Over-the-counter medications are those that a patient can buy without a prescription. However, OTC medications recommended at prescription strength were classified as non-OTC medications because they required more extensive counseling. A PRN medication was one that a physician recommended be taken on an as-needed basis.30 If no verbal recommendation was made during the office visit, a medication was classified as PRN if it was not prescribed for a chronic condition and if it was presumed acceptable for the patient to stop taking it on his or her own without further physician instructions based on the content of the conversation.
Multivariate logistic regression was performed using discussions of cost as the outcome variable. Continuous predictor variables included physician report of the number of continued medications, total length of the visit, and patient and physician age. Categorical variables included the following: patient and physician sex, OTC and PRN medication status, whether the patient had a previous visit to the physician, patient race/ethnicity (white or nonwhite), physician specialty (family medicine, internal medicine, or cardiology), practice site (Kaiser Permanente or University of California, Davis Medical Group), household income (<$20 000, $20 000 to $40 000; >$40 000 to $60 000, or >$60 000 per year), amount of prescription drug costs covered by the patient's health plan (less than half, half or more, or unknown), and medication class (antibiotics, analgesics, cardiovascular medications, genitourinary drugs, psychiatric medications, and ear, nose, and throat preparations, or other). Standard errors were adjusted for clustering of medications within patients, and model goodness of fit was evaluated using pseudo 2.
The 185 patients receiving new prescriptions had a mean age of 55 years, 50% were male, and 83% were white. Almost all had health insurance, and more than 75% paid less than half of their prescription drug costs. Thirty-one percent of the patients were seen by family physicians, 47% by internists, and 23% by cardiologists (Table 1).
Content of Physician Communication About Cost
Physicians discussed cost and insurance issues in 28 encounters involving newly prescribed medications (Table 2), and patients initiated discussions in 4 encounters. In 7 encounters, physicians asked patients what insurance plan they belonged to, and in 5 encounters, they asked about patient out-of-pocket costs, such as "Is your insurance pretty good about covering your pills, or do you buy them all out-of-pocket?" or "How much do you spend on your medications?" Two physicians queried what medications were covered by the patient's insurance plan, and one of them voiced frustration over not knowing which lipid-lowering medication was on the formulary:
Yeah, the insurances do whatever they do. I don't know. I gave up trying to know ahead of time. But if you go to the pharmacy and they say, oh, they don't like to cover this, then ask, 'What do they like to cover that is in this category?' and then let me know.
And I'll change it. But I think [your insurance plan] covers this. I think it does.
Ten discussions centered on insurance formulary restrictions ("[insurance plan] won't pay for it"). In 4 of these cases, physicians suggested ways to get around the restrictions. Physicians prescribing a medication for smoking cessation declared: "Usually I write it for the Wellbutrin, not the Zyban, just because of the insurance," and "When you go down there, you have to tell them it's for depression; otherwise, they don't want to give it to you." Pill splitting was suggested so that patients could obtain a dosage not covered by an insurance plan: "They are 100-mg pills, probably should split them in half. The 100 is the only one that's covered, but you don't need 100, you just need 50, which is half."
Medication price was mentioned in 12 visits. Half of the visits referred to actual dollar amounts, while the other half had qualitative descriptions of the medication being expensive. Seven physicians offered tips for cost savings ("You know, you probably could find it at [pharmacy name] or somewhere cheaper. Shop around."), with 3 physicians mentioning substituting a similar drug in the same category ("If they have any objections to it or don't cover it, there are a number of drugs in that category") and 3 physicians suggesting generic substitution as a means to reducing costs ("The Norvasc you're taking now, that you have to pay $25 for, what I would like to do is to change that to one that is generic").
Physicians made remarks about generic medications for 2% of the new drugs prescribed, but no conversation contained information about the significance of generic vs brand name medications. Physicians sometimes assumed that patients would want only medications that were covered under their insurance plan. For example, a physician seeking to prescribe an antiviral medication discovered that the patient's insurance did not cover the less frequently dosed medications:
Yeah, honestly, unfortunately the famciclovir and the valacyclovir, which are the 2 newer ones, 3 times a day, or twice a day actually. Those work really well, but it is not covered by [insurance plan], and that's what you have, correct?
You have [insurance plan]? So unfortunately you get stuck with the acyclovir, which is 5 times a day.
The medications are equally effective, but the covered medication has to be taken more frequently than its more expensive alternatives. This physician did not ask whether the patient would be willing to pay more for a twice-a-day medication, and the patient did not verbalize a preference.
Examination of a physician-patient interaction, in which the physician prescribed an antihypertensive medication, provides an illustration of how a physician can create a medication regimen taking patient cost concerns into consideration (Table 3). By introducing the topic of potential cost savings with a generic medication, this physician allows the patient to express previous difficulties with medication cost that led to medication nonadherence. The physician also ensures the acceptability and practicality of the medication regimen by querying the patient about his ability to take a twice-a-day medication. Comments about the equivalent efficacy of the generic medication inform the patient about the medication.
Other Aspects of Medication Acquisition
Overall, for 33% of new medications, physicians communicated about any aspect of medication acquisition, including cost and insurance, generic and brand name, logistics, supply, and refills (Table 2). Medication supply, or the quantity of medication to be dispensed to the patient, was mentioned for 9% of the new prescriptions and statements about refills for 5%. Conversations about the logistics of obtaining medications occurred for 19% of the medications prescribed. Typical logistics discussions centered on pharmacy choice and on the need for the patient to go to the pharmacy for a prescription vs buying something OTC. Physicians sometimes gave patients tips about efficiently obtaining medication.
Relationship of Cost Discussions to Medication, Patient, and Physician Characteristics
In multivariate analyses, discussions about medication cost and insurance differed significantly by medication class and site of care. Cost was less likely to be discussed if the medication was prescribed by a family physician (odds ratio [OR], 0.003; 95% confidence interval [CI], 0.000-0.150) or by an internist (OR, 0.02; 95% CI, 0.00-0.49) compared with a cardiologist, as well as by a Kaiser Permanente physician (OR, 0.20; 95% CI, 0.04-0.97) compared with a University of California, Davis, Medical Group physician (Table 4). In addition, cost discussions were less likely to arise with older patients (OR, 0.57; 95% CI, 0.35-0.93) and were more likely with patients earning less than $20 000 per year compared with those earning more than $60 000 per year (OR, 8.27; 95% CI, 1.29-52.80). The relationship of these variables with cost discussion did not change substantially with the addition of visit length or physician race/ethnicity to the model (data not shown).
Physicians infrequently discuss medication cost and acquisition issues when prescribing new medications, and patients rarely initiate these conversations. The content of these discussions consists of physician inquiries about patient insurance plans and out-of-pocket costs, formulary restriction conversations, and strategies for affording medications and overcoming barriers to medication access. Cost and insurance coverage are particularly important points of discussion because high medication costs and copayments are associated with greater patient nonadherence.2,7,8,31-33 Inadequate communication about cost as a barrier to patient medication adherence is problematic because clinicians frequently fail to identify patients with difficulties purchasing prescription medications34 and have poor knowledge regarding whether prescribed medications are covered by patient insurance formularies.35
Insurance formularies delineate the medications to which patients have low cost access. Tiered incentive-based pharmacy benefit plans attempt to steer patients toward generic or preferred medications, for which patients have smaller or no copayments.36 Patients in 3-tier pharmacy plans prescribed generic or preferred pharmaceuticals were more adherent to their medications than those given nonpreferred brand name medications.37 Unfortunately, physicians often do not guide patients toward appropriate medication substitution38 and rarely educate patients about what to do if their medications are not covered by their insurance. As a result, pharmacists telephone physician offices for clarification about more than 40% of prescriptions.39 Of these, up to 26% are for formulary issues, and 37% are for prior authorization issues.40 Conversely, absent discussions, physicians may be unaware of patient preferences to pay an additional cost for medications with more convenient dosing schedules or fewer adverse effects.
This study shows rates of cost discussions about new medications that are similar to those found in a Canadian study19 of physician-patient encounters, suggesting that these issues are not limited to the United States. Medication supply and refills were mentioned for fewer than 10% of the new drugs in this study. However, physicians may address these issues in more detail during subsequent visits, as suggested by a prior observational study15 of new and continued medications, in which more than 50% of patients were asked about medication supply. When chronic medications are initiated, supply and refill issues are important to address. Patients receiving only a 30-day supply of a chronic medication may not realize the need to continue the medication when it runs out and may not seek refills. Other patients not understanding the process of obtaining refills may seek care in the emergency department when their prescriptions run out.24 Future research should query whether overall efficiency of care improves with better discussion about medication cost and acquisition at the time a new medication is prescribed.
Although cardiologists and primary care physicians counsel similarly about essential elements of prescribing,41 cardiologists in this study counseled more about cost and insurance than primary care physicians. This discrepancy could be due to cardiologists' prescribing more nongeneric medications, which cannot be detected in this study. Kaiser Permanente physicians, who need to learn only 1 formulary, may discuss medication cost and insurance less if they knowingly prescribe covered medications. It is concerning that older patients received less counseling about cost, because these patients are most likely to be taking multiple medications and to have the most difficulty with financial strain. It is not surprising that cost discussions occurred more often with patients earning less than $20 000 per year, because these patients have the least disposable income and may initiate more cost conversations with their physicians. Because physicians perceive that members of low socioeconomic groups are less adherent with medical advice,42 they may make a special effort to counsel patients who they think fall into this group. Although this study showed no differences in communication about cost with patients of different racial/ethnic backgrounds, the number of nonwhite patients may have been too small for differences to be detected.
This study has several limitations. Cost and acquisition issues are not universally applicable. For example, physicians with prior knowledge of a patient's financial situation or knowledge of the patient's drug formulary may not need to discuss medication cost or insurance when prescribing a new medication. In time-compressed office visits, conversations about the logistics of where to obtain medications may be better left to office staff. Although we were unable to assess whether these discussions occurred during future office visits, it can be argued that addressing these issues at the time of medication initiation might prevent potential medication nonadherence.
The study included only encounters in which a new medication was actually prescribed and may have missed instances in which a new medication was recommended but not prescribed because of cost or other factors. In addition, the Hawthorne effect43 of having a tape recorder on the table may have enhanced physician counseling about medications. Because the study population was mostly insured, white, and had prescription drug coverage, findings about cost and insurance discussions may not be generalizable to populations with a higher burden of uninsured patients, those without prescription drug coverage, or those with more racial/ethnic diversity. Future studies need to study communication about cost with these populations in more depth.
Discussions about cost and medication acquisition are particularly important because they form the context in which patients obtain their medications. Data for this study were collected before the advent of 3-tier formulary systems and the Medicare Part D insurance plan, both of which have increased the complexity of medication decision making related to coverage and cost. Physician and patient discomfort, along with physician lack of knowledge about how to help patients with medication unaffordability, may be barriers to cost discussions.11,12 Interventions should target improving physician knowledge about helping patients achieve affordable medication regimens, as well as stressing the importance of recognizing and addressing cost and acquisition issues with patients.
We appreciate the collaboration of Honghu Liu, PhD, and the technical assistance of Victor Gonzalez.
From the Department of Family Medicine (DMT) and Division of General Internal Medicine and Health Services Research, Department of Medicine (RDH, NSW), The David Geffen School of Medicine at University of California, Los Angeles, and Department of Sociology (JH), UCLA, and University of California, Davis Center for Health Services Research in Primary Care (DAP, RLK) and Department of Sociology (DAP), University of California, Davis, Sacramento.
This study was supported by grant 034384 from the Robert Wood Johnson Foundation and by the UCLA Specialty Training and Advanced Research (STAR) Program. Dr Tarn was supported by training grant PE19001-09 from the Health Resources Services Administration. Dr Hays was supported by grant P20-MD00148-01 from the UCLA/DREW Project EXPORT, National Institutes of Health, and National Center on Minority Health & Health Disparities and by grant AG-02-004 from the UCLA Center for Health Improvement in Minority Elders/Resource Centers for Minority Aging Research, National Institutes of Health, and National Institute of Aging. Dr Kravitz was supported by mid-career research and mentoring award 1K24MH072756-01 from the National Institute of Mental Health.
Address correspondence to: Derjung M. Tarn, MD, PhD, Department of Family Medicine, The David Geffen School of Medicine at UCLA, 10880 Wilshire Blvd, Ste 1800, Los Angeles, CA 90024. E-mail: firstname.lastname@example.org.