- Email: [email protected]
Clinician identification of chronically ill patients who have problems paying for prescription medications
SPECIAL ARTICLE
Clinician Identification of Chronically Ill Patients Who Have Problems Paying for Prescription Medications Michele Heisler, MD, MPA, Todd H. Wagner, PhD, John D. Piette, PhD PURPOSE: Little is known about whether health care providers are effectively identifying patients who have difficulty covering the costs of out-of-pocket prescription medications. We examined whether and how providers are identifying chronically ill adults who have potential problems paying for prescription medications. METHODS: We conducted a cross-sectional survey of a national sample of 4050 adults aged 50 years or older who use prescription medications for at least one of five chronic health conditions. The primary outcome measure was patient report of being asked by a doctor or nurse in the prior 12 months whether the patient could afford the prescribed medication. The measures of prescription cost burden were cost-related underuse of medications, cutting back on other necessities to pay for medications, and worries about medication costs. We adjusted for patient income, education, race/ethnicity, age, sex, health status, number of prescribed medications, pharmacy benefits, frequency of outpatient visits, having a regular health care provider, and sampling weights.
RESULTS: In the weighted analyses, 16% (547/4050) of respondents reported that they had been asked about potential problems paying for a prescribed medication. Only 360 (24%) of the 1499 respondents who reported one or more burdens from out-of-pocket medication costs reported being asked this question. After adjusting for potential confounders, patients who had cut back on medication use or other necessities to cover payments were no more likely than other patients to be asked about the ability to pay for prescription medications. Concerns about medication costs, being a racial/ethnic minority, taking seven or more prescription medications, and having no prescription coverage were independently associated with a greater likelihood of being asked about possible problems with prescription costs. CONCLUSION: Few chronically ill patients who are at risk of or experiencing problems related to prescription medication costs report that their clinicians had asked them about possible medication payment difficulties. Am J Med. 2004;116: 753–758. ©2004 by Excerpta Medica Inc.
M
Despite the scope and severity of the problem, little is known about whether health care providers are effectively identifying patients who experience burdens associated with prescription medication costs. Identifying such patients would enable clinicians to better help these patients make informed decisions about how to manage their health conditions in the context of financial constraints, for example, through assistance programs or, when possible, use of less expensive treatment alternatives. We have previously examined whether patients discussed with their doctors or nurses their underuse of prescription medications during the previous year because of cost, clinician responses when this problem was raised, and which responses patients found useful (14). In that study, approximately two thirds of respondents who had underused medications reported informing their clinicians, and only 28% of these respondents had been asked if they could afford a medication when it was prescribed. In light of the importance of clinician awareness of potential patient financial problems, we examined responses from a national sample of 4050 chronically ill adults who take prescription medications to determine
any adults in the United States who have chronic illnesses face financial burdens from the cost of out-of-pocket prescription medications (1– 4). High out-of-pocket medication costs can lead to patients underusing medications (5–7) and consequently experiencing adverse health outcomes (8 –10), increasing their use of acute health care services (11,12), and forgoing other necessities to cover costs (13). From the Department of Veterans Affairs Center for Practice Management and Outcomes Research (MH, JDP), Ann Arbor, Michigan; Department of Internal Medicine, and Michigan Diabetes Research and Training Center (MH, JDP), University of Michigan, Ann Arbor; Department of Veterans Affairs Health Economics Resource Center (THW), Palo Alto, California; and Department of Health Research and Policy (THW), Stanford University, Palo Alto, California. Dr. Heisler is a recipient of the Veterans’ Affairs Health Services Research and Development Career Development award. Dr. Piette is a VA HSR&D Career Scientist. The views expressed in this article are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs. Requests for reprints should be addressed to Michele Heisler, MD, MPA, Center for Practice Management and Outcomes Research, VA Ann Arbor Health Care System, P.O. Box 130170, 11H, Ann Arbor, Michigan 48113-0170, or [email protected]. Manuscript submitted June 23, 2003, and accepted in revised form January 29, 2004. © 2004 by Excerpta Medica Inc. All rights reserved.
0002-9343/04/$–see front matter 753 doi:10.1016/j.amjmed.2004.01.013
Identifying Patients With Problems Paying for Medications/Heisler et al
whether and how health care providers are identifying patients who may have problems paying for out-ofpocket medication costs.
Outcome Measure Participants were asked: “In the past 12 months, did a doctor or nurse ask you whether you could afford the medication when they gave you a prescription?”
Independent Variables METHODS Study Participants Study participants were drawn from a national sample identified and recruited by Knowledge Networks (Menlo Park, California) using random-digit dialing among the entire U.S. population. Subjects were contacted via mailings and telephone calls, and potential participants were offered WebTV and free monthly Internet access in exchange for completing a short Internet-based survey several times a month. At the time of the current study (November to December 2002), the sample’s acceptance rate was approximately 48%, and the sample included more than 40,000 members. On most sociodemographic characteristics, key health behaviors (e.g., smoking), and the prevalence of chronic illnesses, the sample was consistently found to be within a few percentage points of estimates from the U.S. Census Bureau’s current population survey and the National Health Interview Survey (15,16). The Human Subjects Committees at the authors’ respective institutions approved the study. We identified 5644 subjects aged 50 years or over who reported taking prescription medications for diabetes, depression, cardiac problems, hypertension, or hypercholesterolemia. After three e-mail requests, 4264 subjects completed the online informed consent and questionnaire; 185 participants who reported no longer taking medications for any of the five index conditions were excluded. The Council of American Survey Research Organization (CASRO) survey response rate was 76% (17,18). We excluded 24 additional respondents because of missing data on income, leaving a final sample of 4055 participants. Compared with nonrespondents, respondents were more likely to be white (88% [3752/4264] vs. 81% [1118/ 1380], P ⬍0.001), older (mean age, 65 vs. 63 years, P ⬍0.001), and to have some college education (66% [n ⫽ 2814] vs. 60% [n ⫽ 828], P ⬍0.001). Respondents and nonrespondents were similar in terms of sex (P ⫽ 0.29), home ownership (P ⫽ 0.44), marital status (P ⫽ 0.16), and income (P ⫽ 0.41). In all analyses, poststratification weights were used to adjust the distribution of respondents to match that of the U.S. population in terms of age, sex, race/ethnicity, education, region, and metropolitan residence to account for oversampling and nonresponse. The Current Population Survey for October 2002 provided data on the distribution of the U.S. population (19). All reported results reflect the weighted sample size of 4050.
754
June 1, 2004
THE AMERICAN JOURNAL OF MEDICINE威
Volume 116
We first examined patient characteristics that we hypothesized providers might use to target queries about potential problems with medication costs. These were ethnicity (white, African American, Hispanic, or other [which included Native Americans and Asian/Pacific Islanders]), education level (high school degree or less vs. some college or more), annual household income (⬍$20,000; $20,000 to $39,999; $40,000 to $59,999; or ⱖ$60,000), number of current prescription medications (one to two, three to six, or seven or more), and whether respondents reported having any pharmacy benefits (“yes” to the question, “Do you have any health insurance that helps pay for prescription medications?”; or “no” to the question, “Just to be clear, when you get prescription medication, do you have to pay the full amount yourself?”). Additional covariates included patient characteristics that providers might be less likely to use. These were sex, age, and self-reported health status (excellent or very good, good, or fair or poor). We also adjusted for patients’ reported frequency of outpatient visits during the past 12 months (none to one, two to five, six or more) and whether patients had a regular health care provider, since more visits and having a regular provider might provide more opportunities to discuss problems of paying for prescription medications. To identify how successful providers were at identifying patients experiencing medication cost problems, we examined three measures of patient burden: cost-related underuse of medication in the previous 12 months, cutting back on other necessities (e.g., food, heat, or other basic needs) to pay for prescription medication in the last 12 months, and worry about medication costs (“at least once every 2 to 3 months” [“at least once a week”, “at least once a month”, or “more than every 2 to 3 months”] vs. “less than every 2 or 3 months” or “never”).
Statistical Analysis Bivariate tabulations and multivariate logistic regression models were used to identify the prevalence and sociodemographic and clinical correlates of reporting that a clinician had asked about potential problems with medication costs. A multivariate logistic regression model was used to assess the independent association between each burden type and being asked about the ability to pay for prescription medications, adjusting for other patient characteristics. Regression diagnostic procedures yielded no evidence of substantive multicollinearity, heteroskedasticity, or influential outliers in any of the models. We
Identifying Patients With Problems Paying for Medications/Heisler et al
performed all analyses using STATA 7 (Stata Corp., College Station, Texas).
RESULTS The sample of 4050 survey respondents was sociodemographically diverse (Table 1). Most subjects reported multiple chronic conditions, including hypertension (73% [n ⫽ 2957]); diabetes (27% [n ⫽ 1094]); a history of myocardial infarction, angina, or heart failure (33% [n ⫽ 1337]); and depression (28% [n ⫽ 1134]). Overall, in the weighted analyses, 37% of respondents (n ⫽ 1499) reported one or more of the three types of burden associated with their out-of-pocket medication costs; 18% (n ⫽ 672) reported taking less medication than prescribed because of the cost and 22% (n ⫽ 724) reported spending less on food, heat, or other basic needs to pay for their medicines. Eleven percent (n ⫽ 446) reported cutting back on both medication use and other necessities. Thirty-four percent (n ⫽ 1171) reported worrying more than every 2 to 3 months about being able to pay for their prescription medications. Among these 1171 subjects, 75% (n ⫽ 878) reported worrying at least once a month and 20% (n ⫽ 234) reported worrying at least once a week.
Likelihood of Clinicians Asking Patients About Potential Problems Paying for Their Medications Sixteen percent of respondents (547/4050) reported that a doctor or nurse had ever asked them in the preceding 12 months about their ability to afford a prescription. In multivariate logistic regression analyses, taking three or more medications, having an annual income of less than $20,000, reporting no pharmacy benefits, and being African American or of other minority race/ethnicity were independently associated with having been asked about prescription cost problems (Table 2). Respondents who reported one or more medication cost–related burdens were more likely than those who reported no burdens to be asked about potential problems with medication costs (24% vs. 15%, P ⬍0.001). Nevertheless, only 24% (n ⫽ 360) of the 1499 respondents who reported one or more burdens from out-ofpocket medication costs reported being asked about medication cost problems in the prior year. Only 23% (155/672) of those who reported underusing medications were ever asked about their ability to afford prescriptions, 26% (188/724) of those who had cut other necessities had been asked, and 24% (281/1171) of those who worried about medication cost problems more than every 2 or 3 months had been asked. When we added the three burden variables to the
Table 1. Characteristics of the 4050 Respondents* Characteristic
Number (%)
Female sex Age (years) 50–54 55–64 ⱖ65 Race White African American Hispanic Other Education Some college or more High school or less Annual household income ⬍$20,000 $20,000–$39,999 $40,000–$59,999 ⱖ$60,000 Self-reported health status Very good or excellent Good Fair or poor Number of current medications 1–2 3–6 ⱖ7 Have prescription benefits Have a regular doctor Number of outpatient visits in past 12 months 0–1 2–5 ⱖ6
2221 (55) 599 (15) 1371 (34) 2080 (51) 3252 (80) 442 (11) 117 (3) 239 (6) 1811 (45) 2239 (55) 922 (23) 1313 (32) 895 (22) 920 (23) 1081 (31) 1597 (41) 1372 (28) 1002 (25) 2074 (51) 974 (24) 3500 (84) 3929 (96) 306 (7) 2130 (51) 1614 (42)
* All results have been adjusted for the survey design and analytic weights. Because of rounding, percentages may not equal 100. A total of 4055 respondents had completed the survey.
above multivariate regression model, respondents reporting cost-related underuse of medication and those who reported cutting other necessities to pay for medications were not more likely than other respondents to be asked about their ability to pay for prescription medications. Only reported worry about medication costs was associated significantly with being asked about medication affordability (odds ratio [OR] ⫽1.7; 95% confidence interval [CI]: 1.2 to 2.5). Independent of whether they reported experiencing any of the three burdens related to medication costs, patients who were African American (OR ⫽ 1.8; 95% CI: 1.1 to 2.7) or of other minority race/ ethnicity (OR ⫽ 2.4; 95% CI: 1.3 to 4.5) continued to have higher odds of being asked about medication cost problems than white patients. Taking seven or more prescription medications (OR ⫽ 1.7; 95% CI: 1.1 to 2.8) and not having pharmacy benefits (OR ⫽ 1.5; 95% CI: 1.1 to
June 1, 2004
THE AMERICAN JOURNAL OF MEDICINE威
Volume 116 755
Identifying Patients With Problems Paying for Medications/Heisler et al
Table 2. Association between Patient Characteristics and Reporting that a Clinician Had Asked About Problems Paying for a New Prescription Medication in the Past 12 Months (n ⫽ 4050)*
Variable Sex Male Female Age (years) 50–54 55–64 ⱖ65 Race White Black Hispanic Other Education Some college or more High school or less Annual income ⱖ$60,000 $40,000–$59,999 $20,000–$39,999 ⬍$20,000 Self-reported health status Very good or excellent Good Fair or poor Number of current medications 1–2 3–6 ⱖ7 Prescription benefits Yes No Have a regular doctor No Yes Number of outpatient visits in past 12 months ⱖ6 2–5 0–1
Unadjusted Frequency of Being Asked n/N (%)
Odds Ratio† (95% Confidence Interval)
256/1829 (14) 378/2221 (17)
Reference 1.2 (0.9–1.6)
114/599 (19) 247/1371 (18) 270/2080 (13)
Reference 1.1 (0.8–1.6) 0.7 (0.5–1.1)
455/3252 (14) 102/442 (23) 23/117 (20) 65/239 (27)
Reference 1.7 (1.1–2.7) 1.5 (0.8–3.1) 2.6 (1.4–4.8)
235/1811 (13) 381/2239 (17)
Reference 1.2 (0.9–1.6)
110/920 (12) 125/895 (14) 197/1313 (15) 194/922 (21)
Reference 1.2 (0.8–1.7) 1.3 (0.9–1.8) 1.6 (1.1–2.3)
119/1081 (11) 256/1597 (16) 274/1372 (20)
Reference 1.2 (0.9–1.7) 1.3 (0.9–1.9)
100/1002 (10) 332/2074 (16) 205/974 (21)
Reference 1.6 (1.0–2.4) 1.9 (1.2–3.2)
525/3500 (15) 116/550 (21)
Reference 1.8 (1.2–2.6)
13/121 (11) 629/3929 (16)
Reference 1.7 (0.8–3.4)
307/1614 (19) 298/2130 (14) 24/306 (8)
Reference 0.8 (0.6–1.1) 0.5 (0.3–1.0)
* All results have been adjusted for the survey’s design and analytic weights. Because of rounding, percentages may not equal 100. A total of 4055 respondents completed the survey. † Adjusted for other patient characteristics. Multivariate logistic models included all the variables listed in the table and sampling weights.
2.2) also remained independently associated with being asked about possible problems paying for prescription medications.
DISCUSSION We found that 37% of chronically ill respondents reported experiencing burdens associated with covering out-of-pocket costs for their prescription medication. 756
June 1, 2004
THE AMERICAN JOURNAL OF MEDICINE威
Volume 116
However, only a minority of these subjects reported that their health care providers had ever asked them during the prior year about possible problems paying for their medications when prescribing. If screening for medication cost problems were nearly universal, we would not expect strong associations between patient risk factors and whether their clinicians asked them about potential financial problems. Yet, given the overall low rates of such queries, one would hope that
Identifying Patients With Problems Paying for Medications/Heisler et al
clinicians were accurately targeting patients who were most likely to experience problems. Patients with very low annual incomes, no pharmacy benefits, and multiple prescriptions indeed were more likely to be asked than those with higher incomes, pharmacy benefits, and fewer medications. Still, only about 1 in 5 respondents in each of these risk groups reported being asked. Of the three medication cost-related burdens, only patient-reported worry was independently associated with being asked about medication payment difficulties in the multivariate analyses. While it is possible that patients who report experiencing frequent worry about their medication costs are communicating their concerns to their providers more actively than are those who are underusing medications or cutting other necessities, our data do not permit us to draw conclusions about the reasons for these differences. African American and other race/ethnic minority patients were significantly more likely than white patients to be asked, even after adjusting for risk factors such as annual income, pharmacy benefits, number of medications, and whether they actually reported medication cost–related burdens. In light of research documenting the ways in which stereotyping and other biases may influence physicians’ treatment decisions for African Americans and other minorities (20,21), our findings suggest that providers may be making racially based assumptions about patients’ ability to pay, as well as their possible medication adherence. Whether this is indeed the case, and whether such assumptions contribute to ethnic disparities in prescribing effective but expensive treatments, needs to be explored in future studies. Our findings do not allow us to draw conclusions about why providers are not asking most patients who have risk factors, experienced burdens, or both, about problems affording their prescription medications. It may be that, besides facing multiple competing demands during office visits (22), providers often do not ask about issues they feel poorly trained to discuss or for which they do not believe they have ready solutions (23,24). One recent study of 484 patient-provider pairs found that although 79% of physicians believed that patients in general wanted to discuss out-of-pocket health care costs, only 35% of these physicians reported ever discussing this matter with their patients (25). In light of the adverse health outcomes associated with cost-related adherence problems, and evidence that providers can increase patients’ understanding of medications and adherence substantially (26 –29), failure to assess possible burdens and underuse stemming from medication costs represents a serious missed opportunity. Our findings have important implications for both individual clinicians and health care systems. At a minimum, clinicians need to be aware of differences in medication costs (30), discuss with patients possible problems
paying for medications as part of broader assessments of medication adherence, and know about available financial assistance programs and lower-cost vendors to reduce patients’ out-of-pocket expenses. The relatively high prevalence of burden from medication costs documented in this and other studies (4,5) further suggests that medical practices and health care systems should institute formal procedures to identify patients facing medication cost–related burdens. For example, questions about problems paying for prescribed medications could be included in previsit questionnaires or intake interviews, a practice that has been shown to successfully increase the use of preventive services and the identification of other health problems (31–33). Our study has several limitations. First, it was based exclusively on patient survey data and thus may be subject to recall and other forms of self-report bias. Second, we only asked respondents about three specific forms of burden related to prescription medication costs. We did not collect data on other strategies that respondents who are concerned about medication costs might have adopted, such as refusing a new prescription (34). Thus, our study does not reflect the full range of burdens that patients may face from out-of-pocket medication costs. Third, although we had previously asked respondents who had cut back on medication use because of cost whether they had told their health care providers about their underuse (14), we did not ask respondents who had experienced worry or cut back on necessities if they had ever taken the initiative to discuss these problems with their providers. In conclusion, only about one fourth of the respondents who experienced burdens related to prescription medication costs reported that they had been asked in the prior year about possible problems paying for their prescription medications. Clinicians and health care systems need to develop more effective strategies to proactively identify and assist patients facing problems associated with out-of-pocket medication costs.
REFERENCES 1. Steinwachs DM. Pharmacy benefit plans and prescription drug spending. JAMA. 2002;288:1773–1774. 2. Adams AS, Soumerai SB, Ross-Degnan D. The case for a Medicare drug coverage benefit: a critical review of the empirical evidence. Annu Rev Public Health. 2001;22:49 –61. 3. Hwang W, Weller W, Ireys H, Anderson G. Out-of-pocket medical spending for care of chronic conditions. Health Aff (Millwood). 2001;20:267–278. 4. Kitchman M, Neuman T, Sandman E, et al. Seniors and Prescription Drugs: Findings from a 2001 Survey of Seniors in Eight States. Menlo Park, California: The Henry J. Kaiser Family Foundation; 2002. 5. Steinman MA, Sands LP, Covinsky KE. Self-restriction of medications due to cost in seniors without prescription coverage. J Gen Intern Med. 2001;16:793–799. June 1, 2004
THE AMERICAN JOURNAL OF MEDICINE威
Volume 116 757
Identifying Patients With Problems Paying for Medications/Heisler et al 6. Kennedy J, Erb C. Prescription noncompliance due to cost among adults with disabilities in the United States. Am J Public Health. 2002;92:1120 –1124. 7. Strickland WJ, Hanson CM. Coping with the cost of prescription drugs. J Health Care Poor Underserved. 1996;7:50 –62. 8. Piette JD, Wagner TH, Potter MB, Schillinger D. Health insurance status, medication self-restriction due to cost, and outcomes among diabetes patients in three systems of care. Medical Care. 2004;42:102–109. 9. Soumerai SB, Avorn J, Ross-Degnan D, Gortmaker S. Payment restrictions for prescription drugs under Medicaid. Effects on therapy, cost, and equity. N Engl J Med. 1987;317:550 –556. 10. Tamblyn R, Laprise R, Hanley JA, et al. Adverse events associated with prescription drug cost-sharing among poor and elderly persons. JAMA. 2001;285:421–429. 11. Soumerai SB, Ross-Degnan D, Avorn J, McLaughlin T, Choodnovskiy I. Effects of Medicaid drug-payment limits on admission to hospitals and nursing homes. N Engl J Med. 1991;325:1072–1077. 12. Soumerai SB, McLaughlin TJ, Ross-Degnan D, Casteris CS, Bollini P. Effects of a limit on Medicaid drug-reimbursement benefits on the use of psychotropic agents and acute mental health services by patients with schizophrenia. N Engl J Med. 1994;331:650 –655. 13. Safran DG, Neuman P, Schoen C, et al. Prescription drug coverage and seniors: how well are states closing the gap? Health Aff (Millwood). 2002;(suppl web exclusives):W253-W268. 14. Piette J, Heisler M, Wagner T. Cost-related medication under-use: do patients with chronic illnesses tell their doctors? Arch Intern Med. 2004. In press. 15. Krosnick JA, Chang L. A Comparison of Random Digit Dialing Telephone Survey Methodology with Internet Survey Methodology as Implemented by Knowledge Networks and Harris Interactive. Columbus, Ohio: Ohio State University; 2001. 16. Wagner TH. Using the internet to conduct surveys with veterans: Knowledge Network’s sample and sampling validity. Available at: http://www.herc.research.med.va.gov/wagner_CHI.htm. Accessed August 30, 2003. 17. Council of American Survey Research Organizations. Guidelines for survey research quality 1998. Available at: http://www.casro. org/guidelines.cfm. Accessed August 30, 2003. 18. Council of American Survey Research Organizations. CASRO Task Force on Completion Rates. New York, New York: Council of American Survey Research Organizations; 1982. 19. Bureau of Labor Statistics and the Bureau of the Census. CPS (Current Population Survey) basic monthly survey. Available at: http:// www.bls.census.gov/cps/bovrvw1.htm. Accessed June 3, 2003. 20. van Ryn M, Burke J. The effect of patient race and socio-economic status on physicians’ perceptions of patients. Soc Sci Med. 2000;50: 813–828.
758
June 1, 2004
THE AMERICAN JOURNAL OF MEDICINE威
Volume 116
21. van Ryn M. Research on the provider contribution to race/ethnicity disparities in medical care. Med Care. 2002;40:I140 –I151. 22. Yarnall KS, Pollak KI, Ostbye T, Krause KM, Michener JL. Primary care: is there enough time for prevention? Am J Public Health. 2003; 93:635–641. 23. Cabana MD, Rand CS, Powe NR, et al. Why don’t physicians follow clinical practice guidelines? A framework for improvement. JAMA. 1999;282:1458 –1465. 24. Levinson W, Gorawara-Bhat R, Lamb J. A study of patient clues and physician responses in primary care and surgical settings. JAMA. 2000;284:1021–1027. 25. Alexander GC, Casalino LP, Meltzer DO. Patient-physician communication about out-of-pocket costs. JAMA. 2003;290:953–958. 26. Bultman DC, Svarstad BL. Effects of physician communication style on client medication beliefs and adherence with antidepressant treatment. Patient Educ Couns. 2000;40:173–185. 27. DiMatteo MR. Patient adherence to pharmacotherapy: the importance of effective communication. Formulary. 1995;30:596 –598 601-602,605. 28. Rost K, Carter W, Inui T. Introduction of information during the initial medical visit: consequences for patient follow-through with physician recommendations for medication. Soc Sci Med. 1989;28: 315–321. 29. Sherbourne CD, Hays RD, Ordway L, DiMatteo MR, Kravitz RL. Antecedents of adherence to medical recommendations: results from the Medical Outcomes Study. J Behav Med. 1992;15:447–468. 30. Korn LM, Reichert S, Simon T, Halm EA. Improving physicians’ knowledge of the costs of common medications and willingness to consider costs when prescribing. J Gen Intern Med. 2003;18:31– 37. 31. Eakin EG, Glasgow RE, Riley KM. Review of primary care-based physical activity intervention studies: effectiveness and implications for practice and future research. J Fam Pract. 2000;49:158 – 168. 32. Stevens VJ, Glasgow RE, Toobert DJ, Karanja N, Smith KS. Oneyear results from a brief, computer-assisted intervention to decrease consumption of fat and increase consumption of fruits and vegetables. Prev Med. 2003;36:594 –600. 33. Glasgow RE, La Chance PA, Toobert DJ, Brown J, Hampson SE, Riddle MC. Long-term effects and costs of brief behavioural dietary intervention for patients with diabetes delivered from the medical office. Patient Educ Couns. 1997;32:175–184. 34. Federman AD, Adams AS, Ross-Degnan D, Soumerai SB, Ayanian JZ. Supplemental insurance and use of effective cardiovascular drugs among elderly Medicare beneficiaries with coronary heart disease. JAMA. 2001;286:1732–1739.
Clinician Identification of Chronically Ill Patients Who Have Problems Paying for Prescription Medications Michele Heisler, MD, MPA, Todd H. Wagner, PhD, John D. Piette, PhD PURPOSE: Little is known about whether health care providers are effectively identifying patients who have difficulty covering the costs of out-of-pocket prescription medications. We examined whether and how providers are identifying chronically ill adults who have potential problems paying for prescription medications. METHODS: We conducted a cross-sectional survey of a national sample of 4050 adults aged 50 years or older who use prescription medications for at least one of five chronic health conditions. The primary outcome measure was patient report of being asked by a doctor or nurse in the prior 12 months whether the patient could afford the prescribed medication. The measures of prescription cost burden were cost-related underuse of medications, cutting back on other necessities to pay for medications, and worries about medication costs. We adjusted for patient income, education, race/ethnicity, age, sex, health status, number of prescribed medications, pharmacy benefits, frequency of outpatient visits, having a regular health care provider, and sampling weights.
RESULTS: In the weighted analyses, 16% (547/4050) of respondents reported that they had been asked about potential problems paying for a prescribed medication. Only 360 (24%) of the 1499 respondents who reported one or more burdens from out-of-pocket medication costs reported being asked this question. After adjusting for potential confounders, patients who had cut back on medication use or other necessities to cover payments were no more likely than other patients to be asked about the ability to pay for prescription medications. Concerns about medication costs, being a racial/ethnic minority, taking seven or more prescription medications, and having no prescription coverage were independently associated with a greater likelihood of being asked about possible problems with prescription costs. CONCLUSION: Few chronically ill patients who are at risk of or experiencing problems related to prescription medication costs report that their clinicians had asked them about possible medication payment difficulties. Am J Med. 2004;116: 753–758. ©2004 by Excerpta Medica Inc.
M
Despite the scope and severity of the problem, little is known about whether health care providers are effectively identifying patients who experience burdens associated with prescription medication costs. Identifying such patients would enable clinicians to better help these patients make informed decisions about how to manage their health conditions in the context of financial constraints, for example, through assistance programs or, when possible, use of less expensive treatment alternatives. We have previously examined whether patients discussed with their doctors or nurses their underuse of prescription medications during the previous year because of cost, clinician responses when this problem was raised, and which responses patients found useful (14). In that study, approximately two thirds of respondents who had underused medications reported informing their clinicians, and only 28% of these respondents had been asked if they could afford a medication when it was prescribed. In light of the importance of clinician awareness of potential patient financial problems, we examined responses from a national sample of 4050 chronically ill adults who take prescription medications to determine
any adults in the United States who have chronic illnesses face financial burdens from the cost of out-of-pocket prescription medications (1– 4). High out-of-pocket medication costs can lead to patients underusing medications (5–7) and consequently experiencing adverse health outcomes (8 –10), increasing their use of acute health care services (11,12), and forgoing other necessities to cover costs (13). From the Department of Veterans Affairs Center for Practice Management and Outcomes Research (MH, JDP), Ann Arbor, Michigan; Department of Internal Medicine, and Michigan Diabetes Research and Training Center (MH, JDP), University of Michigan, Ann Arbor; Department of Veterans Affairs Health Economics Resource Center (THW), Palo Alto, California; and Department of Health Research and Policy (THW), Stanford University, Palo Alto, California. Dr. Heisler is a recipient of the Veterans’ Affairs Health Services Research and Development Career Development award. Dr. Piette is a VA HSR&D Career Scientist. The views expressed in this article are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs. Requests for reprints should be addressed to Michele Heisler, MD, MPA, Center for Practice Management and Outcomes Research, VA Ann Arbor Health Care System, P.O. Box 130170, 11H, Ann Arbor, Michigan 48113-0170, or [email protected]. Manuscript submitted June 23, 2003, and accepted in revised form January 29, 2004. © 2004 by Excerpta Medica Inc. All rights reserved.
0002-9343/04/$–see front matter 753 doi:10.1016/j.amjmed.2004.01.013
Identifying Patients With Problems Paying for Medications/Heisler et al
whether and how health care providers are identifying patients who may have problems paying for out-ofpocket medication costs.
Outcome Measure Participants were asked: “In the past 12 months, did a doctor or nurse ask you whether you could afford the medication when they gave you a prescription?”
Independent Variables METHODS Study Participants Study participants were drawn from a national sample identified and recruited by Knowledge Networks (Menlo Park, California) using random-digit dialing among the entire U.S. population. Subjects were contacted via mailings and telephone calls, and potential participants were offered WebTV and free monthly Internet access in exchange for completing a short Internet-based survey several times a month. At the time of the current study (November to December 2002), the sample’s acceptance rate was approximately 48%, and the sample included more than 40,000 members. On most sociodemographic characteristics, key health behaviors (e.g., smoking), and the prevalence of chronic illnesses, the sample was consistently found to be within a few percentage points of estimates from the U.S. Census Bureau’s current population survey and the National Health Interview Survey (15,16). The Human Subjects Committees at the authors’ respective institutions approved the study. We identified 5644 subjects aged 50 years or over who reported taking prescription medications for diabetes, depression, cardiac problems, hypertension, or hypercholesterolemia. After three e-mail requests, 4264 subjects completed the online informed consent and questionnaire; 185 participants who reported no longer taking medications for any of the five index conditions were excluded. The Council of American Survey Research Organization (CASRO) survey response rate was 76% (17,18). We excluded 24 additional respondents because of missing data on income, leaving a final sample of 4055 participants. Compared with nonrespondents, respondents were more likely to be white (88% [3752/4264] vs. 81% [1118/ 1380], P ⬍0.001), older (mean age, 65 vs. 63 years, P ⬍0.001), and to have some college education (66% [n ⫽ 2814] vs. 60% [n ⫽ 828], P ⬍0.001). Respondents and nonrespondents were similar in terms of sex (P ⫽ 0.29), home ownership (P ⫽ 0.44), marital status (P ⫽ 0.16), and income (P ⫽ 0.41). In all analyses, poststratification weights were used to adjust the distribution of respondents to match that of the U.S. population in terms of age, sex, race/ethnicity, education, region, and metropolitan residence to account for oversampling and nonresponse. The Current Population Survey for October 2002 provided data on the distribution of the U.S. population (19). All reported results reflect the weighted sample size of 4050.
754
June 1, 2004
THE AMERICAN JOURNAL OF MEDICINE威
Volume 116
We first examined patient characteristics that we hypothesized providers might use to target queries about potential problems with medication costs. These were ethnicity (white, African American, Hispanic, or other [which included Native Americans and Asian/Pacific Islanders]), education level (high school degree or less vs. some college or more), annual household income (⬍$20,000; $20,000 to $39,999; $40,000 to $59,999; or ⱖ$60,000), number of current prescription medications (one to two, three to six, or seven or more), and whether respondents reported having any pharmacy benefits (“yes” to the question, “Do you have any health insurance that helps pay for prescription medications?”; or “no” to the question, “Just to be clear, when you get prescription medication, do you have to pay the full amount yourself?”). Additional covariates included patient characteristics that providers might be less likely to use. These were sex, age, and self-reported health status (excellent or very good, good, or fair or poor). We also adjusted for patients’ reported frequency of outpatient visits during the past 12 months (none to one, two to five, six or more) and whether patients had a regular health care provider, since more visits and having a regular provider might provide more opportunities to discuss problems of paying for prescription medications. To identify how successful providers were at identifying patients experiencing medication cost problems, we examined three measures of patient burden: cost-related underuse of medication in the previous 12 months, cutting back on other necessities (e.g., food, heat, or other basic needs) to pay for prescription medication in the last 12 months, and worry about medication costs (“at least once every 2 to 3 months” [“at least once a week”, “at least once a month”, or “more than every 2 to 3 months”] vs. “less than every 2 or 3 months” or “never”).
Statistical Analysis Bivariate tabulations and multivariate logistic regression models were used to identify the prevalence and sociodemographic and clinical correlates of reporting that a clinician had asked about potential problems with medication costs. A multivariate logistic regression model was used to assess the independent association between each burden type and being asked about the ability to pay for prescription medications, adjusting for other patient characteristics. Regression diagnostic procedures yielded no evidence of substantive multicollinearity, heteroskedasticity, or influential outliers in any of the models. We
Identifying Patients With Problems Paying for Medications/Heisler et al
performed all analyses using STATA 7 (Stata Corp., College Station, Texas).
RESULTS The sample of 4050 survey respondents was sociodemographically diverse (Table 1). Most subjects reported multiple chronic conditions, including hypertension (73% [n ⫽ 2957]); diabetes (27% [n ⫽ 1094]); a history of myocardial infarction, angina, or heart failure (33% [n ⫽ 1337]); and depression (28% [n ⫽ 1134]). Overall, in the weighted analyses, 37% of respondents (n ⫽ 1499) reported one or more of the three types of burden associated with their out-of-pocket medication costs; 18% (n ⫽ 672) reported taking less medication than prescribed because of the cost and 22% (n ⫽ 724) reported spending less on food, heat, or other basic needs to pay for their medicines. Eleven percent (n ⫽ 446) reported cutting back on both medication use and other necessities. Thirty-four percent (n ⫽ 1171) reported worrying more than every 2 to 3 months about being able to pay for their prescription medications. Among these 1171 subjects, 75% (n ⫽ 878) reported worrying at least once a month and 20% (n ⫽ 234) reported worrying at least once a week.
Likelihood of Clinicians Asking Patients About Potential Problems Paying for Their Medications Sixteen percent of respondents (547/4050) reported that a doctor or nurse had ever asked them in the preceding 12 months about their ability to afford a prescription. In multivariate logistic regression analyses, taking three or more medications, having an annual income of less than $20,000, reporting no pharmacy benefits, and being African American or of other minority race/ethnicity were independently associated with having been asked about prescription cost problems (Table 2). Respondents who reported one or more medication cost–related burdens were more likely than those who reported no burdens to be asked about potential problems with medication costs (24% vs. 15%, P ⬍0.001). Nevertheless, only 24% (n ⫽ 360) of the 1499 respondents who reported one or more burdens from out-ofpocket medication costs reported being asked about medication cost problems in the prior year. Only 23% (155/672) of those who reported underusing medications were ever asked about their ability to afford prescriptions, 26% (188/724) of those who had cut other necessities had been asked, and 24% (281/1171) of those who worried about medication cost problems more than every 2 or 3 months had been asked. When we added the three burden variables to the
Table 1. Characteristics of the 4050 Respondents* Characteristic
Number (%)
Female sex Age (years) 50–54 55–64 ⱖ65 Race White African American Hispanic Other Education Some college or more High school or less Annual household income ⬍$20,000 $20,000–$39,999 $40,000–$59,999 ⱖ$60,000 Self-reported health status Very good or excellent Good Fair or poor Number of current medications 1–2 3–6 ⱖ7 Have prescription benefits Have a regular doctor Number of outpatient visits in past 12 months 0–1 2–5 ⱖ6
2221 (55) 599 (15) 1371 (34) 2080 (51) 3252 (80) 442 (11) 117 (3) 239 (6) 1811 (45) 2239 (55) 922 (23) 1313 (32) 895 (22) 920 (23) 1081 (31) 1597 (41) 1372 (28) 1002 (25) 2074 (51) 974 (24) 3500 (84) 3929 (96) 306 (7) 2130 (51) 1614 (42)
* All results have been adjusted for the survey design and analytic weights. Because of rounding, percentages may not equal 100. A total of 4055 respondents had completed the survey.
above multivariate regression model, respondents reporting cost-related underuse of medication and those who reported cutting other necessities to pay for medications were not more likely than other respondents to be asked about their ability to pay for prescription medications. Only reported worry about medication costs was associated significantly with being asked about medication affordability (odds ratio [OR] ⫽1.7; 95% confidence interval [CI]: 1.2 to 2.5). Independent of whether they reported experiencing any of the three burdens related to medication costs, patients who were African American (OR ⫽ 1.8; 95% CI: 1.1 to 2.7) or of other minority race/ ethnicity (OR ⫽ 2.4; 95% CI: 1.3 to 4.5) continued to have higher odds of being asked about medication cost problems than white patients. Taking seven or more prescription medications (OR ⫽ 1.7; 95% CI: 1.1 to 2.8) and not having pharmacy benefits (OR ⫽ 1.5; 95% CI: 1.1 to
June 1, 2004
THE AMERICAN JOURNAL OF MEDICINE威
Volume 116 755
Identifying Patients With Problems Paying for Medications/Heisler et al
Table 2. Association between Patient Characteristics and Reporting that a Clinician Had Asked About Problems Paying for a New Prescription Medication in the Past 12 Months (n ⫽ 4050)*
Variable Sex Male Female Age (years) 50–54 55–64 ⱖ65 Race White Black Hispanic Other Education Some college or more High school or less Annual income ⱖ$60,000 $40,000–$59,999 $20,000–$39,999 ⬍$20,000 Self-reported health status Very good or excellent Good Fair or poor Number of current medications 1–2 3–6 ⱖ7 Prescription benefits Yes No Have a regular doctor No Yes Number of outpatient visits in past 12 months ⱖ6 2–5 0–1
Unadjusted Frequency of Being Asked n/N (%)
Odds Ratio† (95% Confidence Interval)
256/1829 (14) 378/2221 (17)
Reference 1.2 (0.9–1.6)
114/599 (19) 247/1371 (18) 270/2080 (13)
Reference 1.1 (0.8–1.6) 0.7 (0.5–1.1)
455/3252 (14) 102/442 (23) 23/117 (20) 65/239 (27)
Reference 1.7 (1.1–2.7) 1.5 (0.8–3.1) 2.6 (1.4–4.8)
235/1811 (13) 381/2239 (17)
Reference 1.2 (0.9–1.6)
110/920 (12) 125/895 (14) 197/1313 (15) 194/922 (21)
Reference 1.2 (0.8–1.7) 1.3 (0.9–1.8) 1.6 (1.1–2.3)
119/1081 (11) 256/1597 (16) 274/1372 (20)
Reference 1.2 (0.9–1.7) 1.3 (0.9–1.9)
100/1002 (10) 332/2074 (16) 205/974 (21)
Reference 1.6 (1.0–2.4) 1.9 (1.2–3.2)
525/3500 (15) 116/550 (21)
Reference 1.8 (1.2–2.6)
13/121 (11) 629/3929 (16)
Reference 1.7 (0.8–3.4)
307/1614 (19) 298/2130 (14) 24/306 (8)
Reference 0.8 (0.6–1.1) 0.5 (0.3–1.0)
* All results have been adjusted for the survey’s design and analytic weights. Because of rounding, percentages may not equal 100. A total of 4055 respondents completed the survey. † Adjusted for other patient characteristics. Multivariate logistic models included all the variables listed in the table and sampling weights.
2.2) also remained independently associated with being asked about possible problems paying for prescription medications.
DISCUSSION We found that 37% of chronically ill respondents reported experiencing burdens associated with covering out-of-pocket costs for their prescription medication. 756
June 1, 2004
THE AMERICAN JOURNAL OF MEDICINE威
Volume 116
However, only a minority of these subjects reported that their health care providers had ever asked them during the prior year about possible problems paying for their medications when prescribing. If screening for medication cost problems were nearly universal, we would not expect strong associations between patient risk factors and whether their clinicians asked them about potential financial problems. Yet, given the overall low rates of such queries, one would hope that
Identifying Patients With Problems Paying for Medications/Heisler et al
clinicians were accurately targeting patients who were most likely to experience problems. Patients with very low annual incomes, no pharmacy benefits, and multiple prescriptions indeed were more likely to be asked than those with higher incomes, pharmacy benefits, and fewer medications. Still, only about 1 in 5 respondents in each of these risk groups reported being asked. Of the three medication cost-related burdens, only patient-reported worry was independently associated with being asked about medication payment difficulties in the multivariate analyses. While it is possible that patients who report experiencing frequent worry about their medication costs are communicating their concerns to their providers more actively than are those who are underusing medications or cutting other necessities, our data do not permit us to draw conclusions about the reasons for these differences. African American and other race/ethnic minority patients were significantly more likely than white patients to be asked, even after adjusting for risk factors such as annual income, pharmacy benefits, number of medications, and whether they actually reported medication cost–related burdens. In light of research documenting the ways in which stereotyping and other biases may influence physicians’ treatment decisions for African Americans and other minorities (20,21), our findings suggest that providers may be making racially based assumptions about patients’ ability to pay, as well as their possible medication adherence. Whether this is indeed the case, and whether such assumptions contribute to ethnic disparities in prescribing effective but expensive treatments, needs to be explored in future studies. Our findings do not allow us to draw conclusions about why providers are not asking most patients who have risk factors, experienced burdens, or both, about problems affording their prescription medications. It may be that, besides facing multiple competing demands during office visits (22), providers often do not ask about issues they feel poorly trained to discuss or for which they do not believe they have ready solutions (23,24). One recent study of 484 patient-provider pairs found that although 79% of physicians believed that patients in general wanted to discuss out-of-pocket health care costs, only 35% of these physicians reported ever discussing this matter with their patients (25). In light of the adverse health outcomes associated with cost-related adherence problems, and evidence that providers can increase patients’ understanding of medications and adherence substantially (26 –29), failure to assess possible burdens and underuse stemming from medication costs represents a serious missed opportunity. Our findings have important implications for both individual clinicians and health care systems. At a minimum, clinicians need to be aware of differences in medication costs (30), discuss with patients possible problems
paying for medications as part of broader assessments of medication adherence, and know about available financial assistance programs and lower-cost vendors to reduce patients’ out-of-pocket expenses. The relatively high prevalence of burden from medication costs documented in this and other studies (4,5) further suggests that medical practices and health care systems should institute formal procedures to identify patients facing medication cost–related burdens. For example, questions about problems paying for prescribed medications could be included in previsit questionnaires or intake interviews, a practice that has been shown to successfully increase the use of preventive services and the identification of other health problems (31–33). Our study has several limitations. First, it was based exclusively on patient survey data and thus may be subject to recall and other forms of self-report bias. Second, we only asked respondents about three specific forms of burden related to prescription medication costs. We did not collect data on other strategies that respondents who are concerned about medication costs might have adopted, such as refusing a new prescription (34). Thus, our study does not reflect the full range of burdens that patients may face from out-of-pocket medication costs. Third, although we had previously asked respondents who had cut back on medication use because of cost whether they had told their health care providers about their underuse (14), we did not ask respondents who had experienced worry or cut back on necessities if they had ever taken the initiative to discuss these problems with their providers. In conclusion, only about one fourth of the respondents who experienced burdens related to prescription medication costs reported that they had been asked in the prior year about possible problems paying for their prescription medications. Clinicians and health care systems need to develop more effective strategies to proactively identify and assist patients facing problems associated with out-of-pocket medication costs.
REFERENCES 1. Steinwachs DM. Pharmacy benefit plans and prescription drug spending. JAMA. 2002;288:1773–1774. 2. Adams AS, Soumerai SB, Ross-Degnan D. The case for a Medicare drug coverage benefit: a critical review of the empirical evidence. Annu Rev Public Health. 2001;22:49 –61. 3. Hwang W, Weller W, Ireys H, Anderson G. Out-of-pocket medical spending for care of chronic conditions. Health Aff (Millwood). 2001;20:267–278. 4. Kitchman M, Neuman T, Sandman E, et al. Seniors and Prescription Drugs: Findings from a 2001 Survey of Seniors in Eight States. Menlo Park, California: The Henry J. Kaiser Family Foundation; 2002. 5. Steinman MA, Sands LP, Covinsky KE. Self-restriction of medications due to cost in seniors without prescription coverage. J Gen Intern Med. 2001;16:793–799. June 1, 2004
THE AMERICAN JOURNAL OF MEDICINE威
Volume 116 757
Identifying Patients With Problems Paying for Medications/Heisler et al 6. Kennedy J, Erb C. Prescription noncompliance due to cost among adults with disabilities in the United States. Am J Public Health. 2002;92:1120 –1124. 7. Strickland WJ, Hanson CM. Coping with the cost of prescription drugs. J Health Care Poor Underserved. 1996;7:50 –62. 8. Piette JD, Wagner TH, Potter MB, Schillinger D. Health insurance status, medication self-restriction due to cost, and outcomes among diabetes patients in three systems of care. Medical Care. 2004;42:102–109. 9. Soumerai SB, Avorn J, Ross-Degnan D, Gortmaker S. Payment restrictions for prescription drugs under Medicaid. Effects on therapy, cost, and equity. N Engl J Med. 1987;317:550 –556. 10. Tamblyn R, Laprise R, Hanley JA, et al. Adverse events associated with prescription drug cost-sharing among poor and elderly persons. JAMA. 2001;285:421–429. 11. Soumerai SB, Ross-Degnan D, Avorn J, McLaughlin T, Choodnovskiy I. Effects of Medicaid drug-payment limits on admission to hospitals and nursing homes. N Engl J Med. 1991;325:1072–1077. 12. Soumerai SB, McLaughlin TJ, Ross-Degnan D, Casteris CS, Bollini P. Effects of a limit on Medicaid drug-reimbursement benefits on the use of psychotropic agents and acute mental health services by patients with schizophrenia. N Engl J Med. 1994;331:650 –655. 13. Safran DG, Neuman P, Schoen C, et al. Prescription drug coverage and seniors: how well are states closing the gap? Health Aff (Millwood). 2002;(suppl web exclusives):W253-W268. 14. Piette J, Heisler M, Wagner T. Cost-related medication under-use: do patients with chronic illnesses tell their doctors? Arch Intern Med. 2004. In press. 15. Krosnick JA, Chang L. A Comparison of Random Digit Dialing Telephone Survey Methodology with Internet Survey Methodology as Implemented by Knowledge Networks and Harris Interactive. Columbus, Ohio: Ohio State University; 2001. 16. Wagner TH. Using the internet to conduct surveys with veterans: Knowledge Network’s sample and sampling validity. Available at: http://www.herc.research.med.va.gov/wagner_CHI.htm. Accessed August 30, 2003. 17. Council of American Survey Research Organizations. Guidelines for survey research quality 1998. Available at: http://www.casro. org/guidelines.cfm. Accessed August 30, 2003. 18. Council of American Survey Research Organizations. CASRO Task Force on Completion Rates. New York, New York: Council of American Survey Research Organizations; 1982. 19. Bureau of Labor Statistics and the Bureau of the Census. CPS (Current Population Survey) basic monthly survey. Available at: http:// www.bls.census.gov/cps/bovrvw1.htm. Accessed June 3, 2003. 20. van Ryn M, Burke J. The effect of patient race and socio-economic status on physicians’ perceptions of patients. Soc Sci Med. 2000;50: 813–828.
758
June 1, 2004
THE AMERICAN JOURNAL OF MEDICINE威
Volume 116
21. van Ryn M. Research on the provider contribution to race/ethnicity disparities in medical care. Med Care. 2002;40:I140 –I151. 22. Yarnall KS, Pollak KI, Ostbye T, Krause KM, Michener JL. Primary care: is there enough time for prevention? Am J Public Health. 2003; 93:635–641. 23. Cabana MD, Rand CS, Powe NR, et al. Why don’t physicians follow clinical practice guidelines? A framework for improvement. JAMA. 1999;282:1458 –1465. 24. Levinson W, Gorawara-Bhat R, Lamb J. A study of patient clues and physician responses in primary care and surgical settings. JAMA. 2000;284:1021–1027. 25. Alexander GC, Casalino LP, Meltzer DO. Patient-physician communication about out-of-pocket costs. JAMA. 2003;290:953–958. 26. Bultman DC, Svarstad BL. Effects of physician communication style on client medication beliefs and adherence with antidepressant treatment. Patient Educ Couns. 2000;40:173–185. 27. DiMatteo MR. Patient adherence to pharmacotherapy: the importance of effective communication. Formulary. 1995;30:596 –598 601-602,605. 28. Rost K, Carter W, Inui T. Introduction of information during the initial medical visit: consequences for patient follow-through with physician recommendations for medication. Soc Sci Med. 1989;28: 315–321. 29. Sherbourne CD, Hays RD, Ordway L, DiMatteo MR, Kravitz RL. Antecedents of adherence to medical recommendations: results from the Medical Outcomes Study. J Behav Med. 1992;15:447–468. 30. Korn LM, Reichert S, Simon T, Halm EA. Improving physicians’ knowledge of the costs of common medications and willingness to consider costs when prescribing. J Gen Intern Med. 2003;18:31– 37. 31. Eakin EG, Glasgow RE, Riley KM. Review of primary care-based physical activity intervention studies: effectiveness and implications for practice and future research. J Fam Pract. 2000;49:158 – 168. 32. Stevens VJ, Glasgow RE, Toobert DJ, Karanja N, Smith KS. Oneyear results from a brief, computer-assisted intervention to decrease consumption of fat and increase consumption of fruits and vegetables. Prev Med. 2003;36:594 –600. 33. Glasgow RE, La Chance PA, Toobert DJ, Brown J, Hampson SE, Riddle MC. Long-term effects and costs of brief behavioural dietary intervention for patients with diabetes delivered from the medical office. Patient Educ Couns. 1997;32:175–184. 34. Federman AD, Adams AS, Ross-Degnan D, Soumerai SB, Ayanian JZ. Supplemental insurance and use of effective cardiovascular drugs among elderly Medicare beneficiaries with coronary heart disease. JAMA. 2001;286:1732–1739.