Physician Perspectives on Deprescribing Cardiovascular Medications for Older Adults

Parag Goyal, MD, MSc; Timothy S. Anderson, MD, MAS, MA; Gwen M. Bernacki, MD, MHSA; Zachary A. Marcum, PharmD, PhD; Ariela R. Orkaby, MD, MPH; Dae Kim, MD, MPH, ScD; Andrew Zullo, PharmD, PhD; Ashok Krishnaswami, MD, MAS; Arlene Weissman, PhD; Michael A. Steinman, MD; Michael W. Rich, MD


J Am Geriatr Soc. 2020;68(1):78-86. 

In This Article


Study Sample

We surveyed a random sample of 750 geriatricians, 750 general internists, and 750 cardiologists from the ACP membership list. ACP is the second-largest medical-specialty organization in the United States, comprising approximately 120 000 physicians who have completed internal medicine residency training; this includes geriatricians, general internists, and cardiologists, among other medical specialties. From July to September 2018, an ACP administrator contacted potential respondents using a standardized email invitation with an embedded link to complete a 24-question web-based survey inquiring about their practice and perspectives relating to deprescribing cardiovascular medications. Participants were offered a $10 honorarium to complete the survey. Between July and September 2018, potential respondents were contacted every 2 to 3 weeks up to eight times until either they completed the survey or the survey period ended. Responses were collected anonymously, and member email addresses were not released to the study team. We excluded respondents who reported that they were not clinically active, did not provide ambulatory care, were trainees, and did not practice geriatrics, general internal medicine, or cardiology. This research study was approved by the Weill Cornell Institutional Review Board.

Survey Design

This survey was designed by investigators who attended a multidisciplinary workshop on pharmacotherapy in older adults with cardiovascular disease, cosponsored by the National Institute on Aging, the American College of Cardiology, and the American Geriatrics Society.[16] Study team members spanning multiple disciplines (geriatrics, general internal medicine, cardiology, and pharmacy) and training levels (fellows-in-training, early-stage investigators, and senior faculty) from across the United States jointly developed a 24-question survey assessing perspectives on deprescribing cardiovascular medications in older adults. The survey was internally tested prior to dissemination.

Respondents were first asked four screening questions to confirm eligibility. Respondents were then asked a series of questions designed to assess deprescribing practices in the prior month, the most common reasons for deprescribing, and the most common barriers to deprescribing. Because patients frequently receive medications from multiple physicians,[17] respondents were asked about whether they considered deprescribing in their practice and whether they considered discussing deprescribing with another physician. Answer choices were based on previously published physician-based facilitators and barriers to deprescribing.[9,10,12,13] The sequence of answer choices was randomly assigned for each respondent to mitigate bias related to the order of answer choices.

Acknowledging that deprescribing practice could result from differences in the patient populations cared for by different specialists (eg, geriatricians often care for patients who are older and/or more frail than do general internists), we sought to characterize differences in deprescribing practices across specialties by providing identical hypothetical patient scenarios to each respondent. Respondents were presented a clinical case of a 79-year-old woman with multiple chronic conditions who took several medications, including four cardiovascular medications. Respondents were asked to identify which (if any) cardiovascular medications they would consider deprescribing for nine clinical scenarios of the same patient with varied concerns, clinical events, and additional medical history. Clinical scenarios incorporated several circumstances where the patient had no concerns, where the patient was symptomatic from a possible adverse drug reaction (lightheadedness with a recent fall, orthostatic symptoms, or hypotension), and where the patient had a limited life expectancy (recent metastatic recurrence of breast cancer, transfer to a skilled nursing facility with a new diagnosis of Alzheimer dementia, increased difficulty in activities of daily living, and age of 90 years).

Last, to determine the generalizability of our findings, all respondents were asked to provide demographic information, years in practice, prior training, and current clinical practice setting. The complete survey is provided in the Supplementary Appendix S1.

Statistical Analysis

We determined descriptive characteristics of respondents by specialty and compared them to the entire survey sample using χ2 for categorical variables. We determined the proportion of each specialty who considered deprescribing, their reported reasons for deprescribing, and their reported barriers to deprescribing. For hypothetical cases, we determined the proportion of respondents in each specialty who would deprescribe any cardiovascular medication and the proportion who would deprescribe any of the following cardiovascular medications: aspirin, atorvastatin, lisinopril, and metoprolol. We used χ2 analysis to test for statistical significance of differences between specialties, using a threshold P < .05. We performed all analyses using Stata, version 14.