Author Affiliations: Center for Practice Management and Outcomes Research, Veterans Affairs Ann Arbor Health Care System, Ann Arbor, Mich; and the Department of Internal Medicine and Michigan Diabetes Research and Training Center, University of Michigan, Ann Arbor.
Prescription drug costs constitute a burden for many chronically ill adults and are strongly related to patients’ likelihood of using less medication than prescribed. We examined the extent to which patients’ trust in their physicians may moderate the impact of economic constraints and other risk factors for cost-related adherence problems.
A total of 912 patients with diabetes recruited from 5 Veterans Affairs health systems completed a detailed cross-sectional survey. Patients reported their level of physician trust using a standard scale and were stratified into high-trust (n = 557) and low-trust (n = 355) groups. We fit multivariate logistic regression models with terms characterizing the interactions between physician trust and (1) patients’ out-of-pocket medication costs and (2) patients’ income, as well as other possible risk factors for cost-related underuse.
Patients with higher out-of-pocket costs were more likely to forgo medications because of cost pressures when physician trust levels were low. Having a low income was only associated with cost-related adherence problems in the context of low physician trust. Patients who reported medication underuse for reasons other than cost were 4 times as likely as other patients to also report cost-related underuse, and those with significant depressive symptoms had more than twice the risk of cost-related underuse compared with those without depression (both P<.05).
These findings suggest that a trusting physician relationship may moderate the impact of cost pressures on patients’ medication adherence. More generally, addressing noncost barriers to adherence may reduce rates of cost-related medication underuse.
Prescription drug costs constitute a burden for many chronically ill adults and are strongly related to patients’ likelihood of using less medication than prescribed.1- 5 Nevertheless, most patients with chronic illnesses report using their medication as prescribed despite cost pressures, even when their incomes are low and they lack prescription drug coverage.1- 3 Studies have identified differences in rates of cost-related adherence problems across racial groups, even when controlling for patients’ ability to pay.1,2 Each of these findings suggests that factors other than cost may either buffer or accentuate the impact of financial pressures on patients’ adherence behavior.
Patients’ decisions to forgo treatment owing to cost are made within the context of other influences on their self-care, such as the value they place on their treatments, their health priorities, and their ability to manage self-management tasks. Patients who are more likely to be nonadherent overall may also be more likely to respond to cost pressures by nonadherence, although this relationship has not been documented. Mental states such as depression can affect patients’ overall regimen adherence6- 8 and may affect how patients respond to medication costs. Understanding patients’ cost-related medication adherence behaviors in the context of other psychosocial correlates could suggest strategies for bolstering adherence even when costs themselves cannot be reduced.
Although clinicians often play a central role in determining patients’ adherence to treatment plans,9,10 physicians’ role in influencing patients’ response to medication costs is not well understood. In addition to the provision of concrete advice for coping with medication costs,11,12 the quality of patient/clinician interactions can have an independent effect on patients’ investment in their self-management goals, as well as their ability to perform specific tasks.13- 15 Patients’ level of trust in their physician may be especially important in shaping their medication adherence decisions in the face of high out-of-pocket medication costs. Physician trust has been found to be even more important than treatment satisfaction in predicting subsequent adherence to recommendations and overall satisfaction with care.16- 18 Physician trust also correlates positively with acceptance of new medications,19 intentions to follow physicians’ advice, perceived effectiveness of care, and improvements in self-reported health status.18,20- 23 Thus, patients who trust their clinicians may place a higher value on their prescription drugs and be more likely to maintain adherence, at least when costs are within a financially feasible range. To our knowledge, no study to date, however, has assessed how patients’ level of physician trust might influence decisions about taking prescription medications in the face of cost pressures.
Accordingly, the purpose of the present study was to determine whether patients who reported high levels of physician trust were less likely to underuse medications because of cost pressures compared with low-trust patients facing similar costs. We also examined variation in cost-related medication underuse across racial groups and the impact of 2 potentially modifiable risk factors: patients’ non–cost-related medication adherence problems and depressive symptoms.
Data were collected as part of a multisite study examining the quality of diabetes care for patients using the Department of Veterans Affairs (VA) health care system.24 Human studies committees at participating medical centers approved the study. Patients with diabetes and at least 1 VA outpatient visit were identified from diagnostic, laboratory, and pharmacy records in 5 regional VA health care systems. Potential participants were contacted by mail with a follow-up telephone call, and eligible patients were asked to provide consent to a computer-assisted telephone interview. Patients who reported receiving most of their care outside of the VA were ineligible for the study. Using a calculation endorsed by the Council of American Survey Research Organizations,25 the study enrollment rate among potentially eligible patients was 57%. Compared with the population of potentially eligible patients, study participants were somewhat younger (mean age, 65.4 vs 67.8 years), more likely to be white (63% vs 58%), and more likely to have 4 or more primary care visits in the prior year (45% vs 38%). Data for the present study were primarily obtained through a written survey conducted 18 months after the initial telephone interview. Of the 1359 initial study participants, 993 (73%) responded to the follow-up survey. Compared with participants who did not complete the follow-up survey, respondents were older, had better reported health status at baseline, and were more likely to be white and have at least completed high school. We excluded 81 follow-up survey respondents because of missing data on their out-of-pocket medication costs.
In the follow-up survey, patients were asked “Not counting the costs paid by your insurance or the VA, how much do your prescription medications cost you and your family each month? In other words, how much do you typically pay ‘out-of-pocket’ per month for prescription medications?” Responses were grouped into 3 categories of usual monthly medication costs: less than $51, $51 to $100, and more than $100. Patients were then asked “In the past 12 months, have you ever taken less of any medications than prescribed by your doctor because of the cost?” (emphasis in the original survey) and “Other than the cost, have you ever taken less medication than prescribed for any other reason?” We examined cost-related medication underuse (yes vs no) as the primary outcome and non–cost-related underuse (yes vs no) as a potential risk factor. The measure of cost-related underuse we used is similar to items used in other national surveys,3,26 in which patients’ responses were strongly associated with their out-of-pocket medication costs, the availability of prescription drug coverage, income, and subsequent health problems.
Patients’ trust in their physicians was measured using the following 5 validated items from the Primary Care Assessment Survey18: “I can tell my doctor anything”; “I completely trust my doctor’s judgment about my medical care”; “My doctor cares more about holding costs down than doing what is needed for my health”; “My doctor would always tell me the truth about my health, even if there was bad news”; and “If a mistake were made in my treatment, my doctor would try to hide it from me.” Participants responded to each item using a 1 to 5 scale from “strongly agree” to “strongly disagree.” After reversing some item scores, scale scores were computed as the average of individual items (mean ± SD, 1.95 ± 0.66; median [range], 2.0 [1.0-4.4]). We divided patients into high-trust and low-trust subgroups at the mean of the distribution. In ancillary analyses, we examined the sensitivity of our findings to the choice of cutpoint and found no differences in the study findings.
We measured patients’ demographic characteristics (age, educational attainment, race, and income) using standard items. Because the sample was heavily weighted toward low-income patients, annual household income was categorized as less than $10 000, $10 000 to $14 999, $15 000 to $24 999, and $25 000 or more. Depressive symptoms were measured using the 10-item Center for Epidemiological Studies Depression Scale. Using the standard convention, scores of 10 or greater were used to identify patients with probable depression.27,28 Patients’ burden of comorbid disease was measured as the number of chronic conditions identified in their baseline medical record review.
In initial bivariate analyses, we used χ2 tests to determine the relationship between patient characteristics and the likelihood of cost-related medication underuse, within strata defined by physician trust. We then constructed multivariate logistic regression models to evaluate the impact of physician trust, medication cost pressures, and other risk factors on patients’ likelihood of reporting cost-related medication underuse. We focused on 2 hypothesized interactions: (1) physician trust and out-of-pocket medication costs and (2) physician trust and income. Because both interactions included physician trust and were thus explicitly correlated, each was evaluated in a separate logistic model. Other statistical strategies were considered (eg, fitting a single multivariate model that included all possible interactions between trust, medication costs, and income) but ruled out because such higher-order logistic regression models often are difficult to interpret.29 Both models controlled for potential clustering of patients’ responses by VA facility.
In the first model, we constructed a 6-level categorical variable representing all possible combinations of physician trust (high vs low) and monthly out-of-pocket medication costs (<$51, $51-$100, and >$100). Using the high-trust + low-cost group as the referent, we estimated the increased odds of cost-related underuse among patients with each of the 5 other possible combinations of trust and cost values. In this model, we also examined the main effects of income, non–cost-related adherence problems, depressive symptoms, and patient race (white vs other). Finally, we controlled for potential confounding by factors that were not the focus of the present study including patients’ age, marital status, and health status (insulin use, number of comorbid conditions, and perceived health status). Patients’ age, marital status, perceived health status, and educational attainment were not significantly associated with cost-related underuse in either bivariate or multivariate analyses (all P>.30) and were dropped from the final model.
The second multivariate model tested the interaction between physician trust (high vs low) and the 4-level ordinal measure of patients’ income. We used the high-trust + highest income group as the referent and evaluated the increased risk of cost-related medication underuse within each of the 7 other less favorable combinations of the 2 risk factors. In this model, patients’ usual monthly out-of-pocket costs were included as a covariate along with the other risk factors and potential confounders.
For interaction effects in both models, we computed statistics testing for differences between the referent category and each comparison group; we also statistically tested the equivalence of odds ratios (ORs) comparing patients with similar cost pressures but different levels of physician trust. P<.05 was considered statistically significant.
More than half of all patients (55%) reported monthly out-of-pocket prescription drug costs of less than $51, while 22% reported monthly prescription drug costs of more than $100 (Table 1). Most study participants had annual incomes of less than $25 000 and were white (81%), male (98%), and 65 years or older (53%). Overall, there were few significant differences in the characteristics of high- and low-trust patients (Table 1). However, low-trust patients were more likely to report non–cost-related medication adherence problems and significant depressive symptoms.
Among patients with high levels of physician trust, rates of cost-related underuse increased from 4% among patients with monthly out-of-pocket costs of less than $51 to 11% among patients with monthly costs of more than $100 (P = .01) (Table 2). Rates of underuse increased from 4% to 30% over this same range of costs in the low-trust group (P<.001). Among patients with high levels of physician trust, there were no differences in rates of cost-related underuse across income levels (P = .63). However, the proportion of patients reporting cost-related underuse was significantly greater among lower-income patients in the subgroup reporting low trust (P = .04). In both trust strata, cost-related underuse was significantly more common among patients reporting adherence problems for other reasons, as well as among patients with significant depressive symptoms.
In model 1, testing the interaction between trust and patients’ out-of-pocket medication costs, costs were a stronger determinant of medication underuse when trust levels were low (Table 3). Relative to patients with high levels of physician trust + low monthly medication costs, patients with high trust but moderate costs ($51-$100) were not more likely to report cost-related underuse (adjusted OR, 2.4, 95% confidence interval [CI], 0.8-7.4). In contrast, patients with moderate cost but low physician trust had a significant increase in their risk for underuse (adjusted OR, 3.3; 95% CI, 1.0-11.5). Patients with the highest out-of-pocket costs had significantly increased rates of cost-related underuse regardless of trust levels. However, the magnitude of the increased risk was significantly greater when trust levels were low (adjusted OR, 14.0) than when trust levels were high (adjusted OR, 4.8; test of equivalence, P<.001).
In model 2, testing the interaction between trust and income, we found that lower-income patients were at increased risk for cost-related medication underuse only when trust levels were low (Table 4). Relative to patients with high levels of physician trust and incomes of $25 000 or more, patients with incomes of $10 000 to less than $15 000 only had an increased risk of cost-related underuse when trust levels were low (adjusted OR, 2.6; 95% CI, 1.6-4.3). Patients with high trust and the lowest incomes had no increased risk of underuse relative to the referent group (adjusted OR, 0.6; 95% CI, 0.1-3.7), whereas patients with the lowest incomes had substantially increased risk of cost-related underuse when trust levels were low (adjusted OR, 4.5; 95% CI, 1.6-12.7).
In both multivariate models, cost-related underuse was strongly associated with whether patients reported adherence problems for other reasons (adjusted ORs, 4.0 [model 1] and 3.9 [model 2]; both P<.01). Also in both models, patients with significant depressive symptoms had more than twice the odds of cost-related underuse relative to nondepressed patients (both P<.01). The magnitude of the adjusted OR associated with nonwhite race was similar in both models, reaching statistical significance in model 1 (P = .04) but not in model 2 (P = .06).
We found that patients with diabetes who reported low levels of physician trust were at significantly higher risk of underusing medications in response to medication cost pressures than were patients with similar cost pressures but greater trust in their physicians. Low income was a risk factor for cost-related underuse among low-trust patients but not among high-trust patients with similar incomes. These findings suggest that physicians’ efforts to enhance patients’ trust may contribute to decreased rates of cost-related medication underuse even when patients’ prescription drug benefits or medication regimens cannot be changed. It is important to emphasize, however, that out-of-pocket drug costs remained a significant risk factor for nonadherence even when levels of physician trust were high.
The mechanism explaining the possible moderating effect of physician trust on the relationship between cost pressures and adherence problems warrants further study. It is possible that greater physician trust fosters more effective patient-physician communication, which in turn heightens patients’ understanding of their medications’ importance for their health and well-being. It is also possible that greater physician trust is associated with patients’ willingness to raise problems they face from medication costs during clinical encounters.30 Future research, particularly studies that include qualitative data collection, should be conducted to shed light on these important pathways.
In both multivariate models, the risk of cost-related adherence problems was significantly greater among patients who reported forgoing medications for other reasons besides cost. These findings suggest that the impact of medication costs on patients’ self-management should be viewed in the context of their overall adherence behaviors and attitudes, rather than solely as an isolated economic phenomenon. The predictors of overall medication underuse remain poorly understood31,32; however, limiting the complexity of patients’ regimens, providing simple reminders, encouraging patients’ input in treatment decision making, and enhancing communication with clinicians have all been shown to be beneficial.31 These same strategies may be particularly important when patients face the added adherence burden of high treatment costs. Similarly, while effectively addressing other psychosocial correlates of cost-related medication underuse such as depression is not a solution to the problem of cost-related medication underuse, addressing these cofactors may be an important component of an overall plan to assist patients in taking their medication as prescribed.
The strongest predictor of patients’ trust in their physicians is physicians’ communication style,16,18,21,33 including behaviors such as active listening, providing emotional support, providing clear and thorough information, eliciting patients’ input in treatment decision making, and allowing adequate time for patients to ask questions.17,34,35 Health system factors also can promote patients’ trust in their physicians. For example, encouraging patient choice in their selection of physicians, continuity of care, accessibility of physicians, and providing sufficient encounter time all can boost trust levels.33,36,37
Participants in this study were recruited from VA health care systems, and the VA has more generous prescription coverage compared with most other public or private payers in the United States. This coverage explains the large proportion of respondents who reported relatively low monthly medication costs, as well as the low overall proportion of patients with cost-related adherence problems compared with patients who have Medicare1,5 or private prescription coverage.38,39 Although we have no reason to expect that the interplay of trust, medication cost, and adherence would be different among non-VA patients, future studies should replicate these findings in non-VA samples. Within this study’s target population, response rates for the baseline and follow-up surveys were relatively low. We suspect that study participation may be associated with greater overall satisfaction with care, greater physician trust, and better treatment adherence.
Several study caveats result from our inability to fully capture the complexity of patients’ experience with medications and other clinical services in this survey. Most older chronically ill patients use multiple medications, and rates of cost-related underuse may not be consistent for all drugs in a patient’s regimen.40- 42 In the present study, we could not determine whether patients who reported low levels of physician trust were more likely to underuse “essential” medications, less essential medications, or both medication types.43 Similar to most patients with diabetes, those in the present study saw a variety of physicians to manage their glycemic control, complications, and comorbid chronic diseases. However, the trust measure we used included the phrase “your doctor” (ie, singular), and the implications of this discrepancy are unclear. Patients may have responded reliably about trust in their primary care provider, and primary care provider trust may in fact be the operant determinant of patients’ adherence in the context of cost. Alternatively, patients’ trust in their overall health care team may determine their response to medication costs, and it is possible that study participants’ trust reports accurately reflect those global impressions. Regardless of which type of physician trust actually drives patients’ response to cost pressures, the discrepancy between patients’ health care experience and the survey items may mean that the role of physician trust in cost-related adherence problems has been underestimated.
Finally, this survey was cross-sectional; we cannot definitively determine whether low trust increased the risk of cost-related underuse or whether physicians of poorly adherent patients behaved in a manner that decreased these patients’ trust reports. We also were not able to control for clustering of patients’ cost-related underuse by physician. Although we controlled for clustering by VA facility, our lack of information on physician characteristics and inability to cluster by physician is a limitation of these analyses.
With these caveats, we conclude that patients’ trust in their physicians may influence their likelihood of responding to medication cost pressures by underusing medications. More generally, the study suggests that cost-related medication underuse not only is an economic phenomenon but also is influenced by other determinants of patients’ adherence behavior. Along with prescribing cost-effective medication regimens, physicians should support patients’ adherence by forging more trusting relationships and addressing noncost adherence barriers.
Correspondence: John D. Piette, PhD, Center for Practice Management and Outcomes Research, VA Ann Arbor Health Care System, PO Box 130170, Ann Arbor, MI 48113-0170 (email@example.com).
Accepted for Publication: April 4, 2005.
Financial Disclosure: None.
Funding/Support: This study was supported by grant SDR 01-019 from the Department of Veterans Affairs, Health Services Research and Development Service (VA HSR&D), Washington, DC. Drs Heisler and Kerr are supported by the VA HSR&D Career Development Program. Dr Piette is a VA HSR&D Career Scientist.
Thank you for submitting a comment on this article. It will be reviewed by JAMA Internal Medicine editors. You will be notified when your comment has been published. Comments should not exceed 500 words of text and 10 references.
Do not submit personal medical questions or information that could identify a specific patient, questions about a particular case, or general inquiries to an author. Only content that has not been published, posted, or submitted elsewhere should be submitted. By submitting this Comment, you and any coauthors transfer copyright to the journal if your Comment is posted.
* = Required Field
Disclosure of Any Conflicts of Interest*
Indicate all relevant conflicts of interest of each author below, including all relevant financial interests, activities, and relationships within the past 3 years including, but not limited to, employment, affiliation, grants or funding, consultancies, honoraria or payment, speakers’ bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued. If all authors have none, check "No potential conflicts or relevant financial interests" in the box below. Please also indicate any funding received in support of this work. The information will be posted with your response.
Register and get free email Table of Contents alerts, saved searches, PowerPoint downloads, CME quizzes, and more
Subscribe for full-text access to content from 1998 forward and a host of useful features
Activate your current subscription (AMA members and current subscribers)
Purchase Online Access to this article for 24 hours
Some tools below are only available to our subscribers or users with an online account.
Download citation file:
Web of Science® Times Cited: 112
Customize your page view by dragging & repositioning the boxes below.
and access these and other features:
Enter your username and email address. We'll send you a link to reset your password.
Enter your username and email address. We'll send instructions on how to reset your password to the email address we have on record.
Athens and Shibboleth are access management services that provide single sign-on to protected resources. They replace the multiple user names and passwords necessary to access subscription-based content with a single user name and password that can be entered once per session. It operates independently of a user's location or IP address. If your institution uses Athens or Shibboleth authentication, please contact your site administrator to receive your user name and password.