Social influences on practice (SIP)
|
Principal Investigator
|
PI Home Department
|
Source of Support
|
Dates
|
Total Cost
|
|
Richard Kravitz, M.D.
|
Internal Medicine
|
National Institute of Health (NIH)
|
9/2002 - 8/2006
|
$2,004,151
|
a) to estimate the effect of direct to consumer (DTC) advertisement driven requests on physicians’ prescribing of antidepressants;
b) to assess whether direct-to-consumer advertising driven requests facilitate or impede the provision of high quality medical care
c) to evaluate the effect of the SP request style on physicians’ communication behaviors.
Data collection and data analyses for the major research question (the impact of DTC advertising on physician behavior) are complete. This study involved trained standardized patients (SP) who completed office visits to enrolled primary care physicians under six different conditions. The six SP presentations varied by condition (depression versus adjustment disorder) and by request style (direct request for a specific antidepressant, request for an antidepressant--brand unspecified, or no medication request at all). Year 3 activities focused on completion of data entry and management, coding variables, data analysis, and preparation of manuscripts. All audible visit recordings (286) have been transcribed. Data collection is complete: 298 standardized patient visits were collected from 152 participating physicians. Physicians were debriefed in two stages. First an initial debriefing regarding the aims of the study; in the second debriefing physicians were actually provided data with which they could compare their sites’ performance with other sites. Physicians were also provided the option of receiving copies of the recorded visits. Coding of data relevant to several key analyses is completed: physician-patient interaction coding, physician participatory style for treatment/ education, and chart reviews.
A summary of initial results follows:
Aim a (effect of requests on prescribing) has been accomplished. Role fidelity in 298 visits was excellent; physicians were “suspicious” that the SP was not a real patient in 13% of visits. Among 149 primary care visits by SPs with major depression, antidepressants were prescribed in 53% when brand-specific requests were made, in 76% when general requests were made, and in 31% when no requests were made (p<.001). Among 149 visits by SPs trained to portray adjustment disorder, antidepressants were prescribed in 55% when brand-specific requests were made, in 39% when general requests were made, and in 10% when no requests were made (p<.001). The results were confirmed in multivariate (mixed effect) models. Minimally acceptable initial care (any combination of an antidepressant, mental health referral, or follow-up within 2 weeks) was offered to 98% of SPs in the major depression role making a general request, 90% of those making a brand-specific request, and 56% of those making no request.
These results were published in the April 27, 2005 issue of JAMA, where they were accompanied by an editorial. The results were also presented during the Opening Plenary Session of the Society of General Internal Medicine (May, 2005) and during the “Best Abstracts” Session of AcademyHealth (June, 2005). Findings from the study were reported by the Washington Post, Los Angeles Times, National Public Radio, and CNN; results were also featured on the NIMH home page in late April and early May, 2005.
Aim b (effect of requests on quality of care) has been substantially accomplished. As noted above, the JAMA paper addressed the issue of clinical quality by measuring the proportion of major depression visits in which physicians provided minimally acceptable initial care (MAIC). Rates of MAIC were much higher among SPs who requested medication compared with those who presented with symptoms only. In a companion manuscript (Kravitz et al., submitted to Annals Intern Med), we examined patient-, physician-, and system-level factors influencing physicians’ recommendations for mental health referral. Among 298 SP visits, 107 (36%) resulted in mental health referrals (most to non-psychiatrists). Physicians or their office staff provided help with securing an appointment in 18% of referrals and recommended a specific person or list in 34%; otherwise patients were told to call their health plan (20%) or were offered no specific assistance (28%). Patient requests for antidepressants increased the likelihood of referral. Physician age, gender, race/ethnicity, and specialty had little relationship with referral, but referrals were less likely among physicians with greater self-confidence in their ability to manage antidepressant therapy (adjusted odds ratio [AOR] 0.39, 95% confidence interval [CI] 0.17 to 0.86), and referrals were more likely if physicians typically spent >=10% of professional time on non-clinical activities (AOR 3.42 , 95% CI 1.45 to 8.07) or had personal life experience with psychotherapy for depression (AOR 2.74, 95% CI 1.15 to 6.52). Perceived availability of mental health services (usually being able to obtain mental health consultation within two weeks) was also associated with the likelihood of referral (AOR 2.94, 95% CI 1.26 to 6.92).
A third manuscript (Feldman et al., submitted to Annals Intern Med) addresses the charge that DTC-advertising adversely affects quality by derailing the clinical examination. Specifically, we asked whether patients’ antidepressant requests affect: a) history-taking by primary care physicians (PCPs) for patients with depressive symptoms and a coexisting musculo-skeletal disorder, b) the diagnosis of depression, c) the provision of minimally acceptable initial care, and d) visit duration? General antidepressant requests were associated with more depression history-taking (Adjusted Parameter Estimate [APE] = 0.80 more questions out of 10 (95% Confidence Interval [CI] = 0.31, 1.29, p = 0.00); brand-specific requests were marginally associated with depression history-taking (APE = 0.45, 95% CI = -0.04, 0.93, p= 0.07). Requests were not related to musculoskeletal question asking (p > 0.3) or visit length (p > .8). Depression history taking was directly associated with the likelihood of a chart diagnosis of depression and the provision of minimally acceptable initial depression care.
Aim c (effect of requests on physician-patient communication) is underway. This analytic aim is proceeding along two tracks. The Patient Centered Care Team, led by Dr. Epstein, is examining the relationship of patient assertiveness (as represented experimentally by requests) with physicians’ patient-centered communication behaviors. A review article on measurement of patient-centered care has been published (Epstein et al., Soc Sci Med 2005). An additional manuscript (Sheilds et al., in preparation) tests the hypothesis that a patient-centered communication style (asking more questions, probing patients’ concerns) reduces physicians’ reactivity to patients’ requests for marginally indicated clinical services. The Qualitative Team, led by Dr. Paterniti, is using grounded theory to evaluate SPs triple role as actors, “patients,” and expert witnesses of the structure and process of medical care. Dr. Paterniti et al. are also evaluating the nature of physicians’ refusals to comply with patients’ requests.
Several additional manuscripts are submitted or in preparation. Franz et al. (submitted to HSR: Health Services Research) examines the prevalence, predictors, and consequences of SP detection (that is, when the physician becomes suspicious that the patient in the room is not a “real” patient). We found that the prevalence of detection depends on how detection is defined, and that meaningful detection is uncommon but may have some impact on physicians’ propensity to recommend mental health referrals for SPs with depressive symptoms. A second methods paper (Carney et al., in preparation), develops a conceptual framework for evaluating SP role fidelity and reliability. Two manuscripts by Young et al., prepared under a Minority Supplement, further inform our understanding of the physician-patient interaction in the treatment of depression. The first (Young et al., submitted) examines the extent to which primary care physicians educate their patients about antidepressant therapy. The second (Young et al., in preparation) measures the degree to which physicians engage patients in shared decision making as they evaluate treatment options for depressive symptoms.
(1) Patients’ requests for antidepressant medication are a powerful influence on physician prescribing.
(2) While patients’ requests for antidepressants may increase prescribing in clinically marginal situations (e.g., adjustment disorder),
they also reduce under-treatment of major depressive disorder.
(3) Standardized Patients are a powerful tool for examining physician behavior in an experimental context.
SIP Manuscripts Submitted, In Press, or Published
Influence of patients’ requests for direct-to-consumer advertised antidepressants: a randomized controlled trial
Kravitz RL, Epstein RM, Feldman M.D., Franz CE, Azari R, Wilkes MS, Hinton L, Franks P.
JAMA 2005; 293:1995-2002
Measuring patient-centered communication in Patient-Physician consultations: Theoretical and practical issues
Epstein RM, Franks P, Fiscella K, Shields CG, Meldrum SC, Kravitz RL,Duberstein PR.
Soc Sci Med. 2005;1516-1528
Caught in the act? Prevalence, predictors, and consequences of physician detection of unannounced standardized patients
Franz CE, Epstein R, Miller K,Brown A, Song J, Feldman M.D., Franks P, Kelly-Reif S, Kravitz RL
Health Services Research (submitted)
Mental health referrals for depressive symptoms in primary care: patient, physician and system effects
Kravitz RL, Franks P, Feldman M, Meredith LS, Hinton L, Franz C, Duberstein P, Callahan E, Epstein RM.
Annals Intern Medicine (submitted)
Do Patient Requests for Antidepressants Enhance or Hinder Physicians’ Evaluation of Depression: A Randomized Controlled Trial?
Feldman M, Franks P, Epstein RM, Franz CE, Kravitz RL.
Annals Intern Medicine (submitted)
An examination of the quantity and content of physician communication about prescribed antidepressants
Young HN, Bell RA, and Kravitz RL.
JGIM (submitted)