A qualitative study to understand clinical providers' attitudes and beliefs about polypharmacy and their perceived barriers to and facilitators of appropriate medication discontinuation.
ABSTRACTObjectives: Polypharmacy and adverse drug events lead to considerable healthcare costs and morbidity, yet there is little to guide clinical providers in the area of discontinuing medications that may not be necessary. We sought to understand providers’ beliefs and attitudes about polypharmacy and medication discontinuation.
Study Design: Qualitative study using semi-structured interviews of 20 providers with prescribing privileges at 2 US Veterans Affairs Medical Centers, from April 2012 to October 2012.
Methods: Transcribed interviews were analyzed using grounded thematic analysis, a systematic approach to deriving qualitative themes from textual data.
Results: We identified 10 themes within 4 domains of medication discontinuation. Within the first domain (medication factors), we identified 2 themes: 1) medication characteristics, and 2) uncertainties of why a patient was taking a particular drug. Within the second domain (patient factors), we identified 3 themes: 3) clinical picture of the patient, 4) clinicians’ understanding of the patients’ knowledge and beliefs, and 5) patients’ adherence. Within the third domain (clinical provider factors), we identified 2 themes: 6) professional identity, and 7) providers’ decisions related to their own beliefs about medications. Within the fourth domain (system factors), we identified 3 themes: 8) multiple providers, 9) workload, and 10) external directives and policies such as structural components of a healthcare system.
Conclusions: Provider decisions to discontinue medications are affected by factors at all levels of the clinical encounter. Our findings have implications for development and implementation of interventions to improve appropriate medication discontinuation via enhanced medication reviews, enriched patient-provider communication, and better system-level structures. This, in turn, may reduce the continued prescribing of potentially inappropriate medications that can lead to adverse outcomes or increased healthcare costs.
Am J Manag Care. 2015;21(4):277-283
Clinical providers express concerns about polypharmacy but receive little guidance on how to make decisions about medication discontinuation. Until now, there has been little understanding of providers’ perceptions of factors affecting these decisions.
Polypharmacy, often defined as a patient taking 5 or more medications,1 is common. Roughly 40% of adults 65 years or older experience polypharmacy, with a similar prevalence in active users of Department of Veterans Affairs (VA) facilities.2,3 A greater number of medications is associated with risk of adverse drug events (ADEs), which leads to increased healthcare utilization, costs, and morbidity.4,5 Moreover, physicians may interpret subtle adverse physiologic effects as a distinct problem and add medication to treat them; such a “prescribing cascade” can lead to increased rates of polypharmacy.6
Providers’ perceptions of polypharmacy, as well as barriers to and facilitators of mitigating it through proactive medication discontinuation, are unclear. Clinicians lack evidence of best practices for evaluating patients’ medication regimens to determine which drugs, if any, may safely be discontinued, especially for ambulatory care patients. Several studies have demonstrated the feasibility and safety of withdrawing particular classes of medications, but many focus on inpatients and residents of long-term care facilities.7,8 While medication adherence and medication reconciliation have emerged as health system priorities, little attention has been paid to promoting the discontinuation of medications that may not be necessary or whose benefits no longer outweigh associated risks. Experts have suggested that the “time until benefit” of a medication be compared with life expectancy in elderly populations to aid in decisions about its continued use.9 Whether clinicians consider such risk-benefit balances in the general adult population is unknown.
Appropriate cessation of a medication could lead to a reduction in ADEs and to improved health outcomes, yet little is known about factors that inhibit or facilitate providers’ discontinuation of medication. Therefore, we sought to understand how providers make decisions about medications in a complex population by identifying their beliefs and attitudes about polypharmacy, as well as their perceptions of the factors that influence decisions and actions related to medication discontinuation. Based on prior literature, we hypothesized 4 domains related to discontinuation: medication, patient, provider, and system (); and we sought to identify key themes within each domain to inform future interventions to improve medication discontinuation.
We conducted a qualitative study using in-depth interviews of primary care providers and pharmacists at 2 US VA Medical Centers. Qualitative methods are useful for understanding new areas and generating data to inform future research. We explored attitudes toward medication reconciliation, polypharmacy, patient-provider communication, and decision making about discontinuing medication. Institutional review boards at both sites approved the study.
At primary care staff meetings, the lead investigator (AL) introduced the study and distributed recruitment materials describing the project objective of understanding decision making, communication, and factors associated with medication discontinuation. Providers with prescribing privileges—including physicians, nurse practitioners (NPs), and pharmacists—were eligible to participate. Within the VA, clinical pharmacists are authorized to initiate, titrate, and discontinue medications. All eligible providers (n = 63) were subsequently invited to participate, and a convenience sample was scheduled for interviews. All participants provided written informed consent; no stipend was provided.
The lead investigator conducted individual, in-person, in-depth, semi-structured interviews of 30 to 60 minutes’ duration between April 2012 and October 2012. We queried providers’ views and conduct of medication reconciliation, communication with patients, interactions with colleagues, beliefs about medications, polypharmacy, and experience with discontinuing medications. The interviewer used the Interview Guide for Providers (see , available at www.ajmc.com) flexibly and followed up on unanticipated issues raised by participants.
Interviews were audio-recorded, transcribed verbatim, and reviewed for accuracy. Transcripts were analyzed qualitatively using procedures informed by grounded theory methodology,10 a systematic approach to deriving qualitative themes from textual data. This approach begins with open coding, in which investigators identify key concepts emerging from the language used by participants and assign codes (descriptive phrases) to segments of text. We began analyses using sensitizing concepts surrounding polypharmacy and medication discontinuation.11 Using NVivo qualitative analysis software (QSR International, Doncaster, Australia),12 the lead investigator and a research assistant (TBM) coded 4 interviews independently, and through discussion came to consensus on code definitions to develop a coding dictionary. The remaining interviews were coded by the lead investigator, who refined the dictionary as needed. Throughout, detailed coded text segments were reviewed, discussed, and revised if needed by 2 investigators (AL, BB) before the text was condensed into broader themes. Prominent themes and exemplifying quotes were discussed by the research team, and each theme was mapped to one of the 4 domains (medication, provider, patient, and system factors) that comprised our conceptual model. Themes that could have been mapped to multiple domains were assigned the best fit through consensus discussion.
We interviewed 20 providers (8 male, 12 female; 15 site A, 5 site B), with a range of experience in clinical practice (2 to 39 years; median 21) and in the VA (2 to 30 years; median 15). The majority (n = 11) were physicians, with 3 NPs and 6 pharmacists. Twenty interviews were sufficient to achieve thematic saturation, where at a certain point no new concepts appeared. Ten themes emerged, each mapping to 1 of 4 domains influencing the process of proactive medication discontinuation (Figure). Exemplifying quotes that best represent the general sense of each theme can be seen in the .
Medication characteristics. Characteristics of medications, such as the number of daily doses, therapeutic duplication, or whether 1 medication could address 2 problems, influenced clinicians’ decisions when reviewing regimens. Many reported paying particular attention to specific medication classes as potential targets for discontinuation (eg, proton pump inhibitors, statins, opioid pain medications). Reducing or discontinuing pain medications was described as more complicated due to perceived patient reluctance.
Indication uncertainty. Providers spoke at length about the difficulties of decision making with unclear medication indications. Multiple reasons for uncertainty were suggested, such as a patient’s inability to state the reason for his medication, or the presence of an associated indication that might not reflect the true rationale for the medication’s use. For some of their patients, the default for some providers was to maintain the status quo in the absence of a clear reason to stop a medication, despite uncertainty. For other patients, a trial of discontinuation would evaluate its necessity.
Clinical picture. Providers noted that the presence of multiple comorbid conditions—each of which might require 1 or more medications—inhibited discontinuation. Conversely, a geriatric patient’s age sometimes triggered discontinuation because of concerns related to polypharmacy, decline in drug metabolism, or shifts in the risk-benefit ratio due to changes in perceived life expectancy—especially for patients with terminal cancer. Impaired patient cognition also made clinical decision making difficult because of uncertainties about whether the recommended changes would be accurately understood and implemented.
The intended and unintended medication effects were key to medical decision making. Providers felt it was important to elicit whether a medication was providing symptomatic relief, despite noting that paradoxically, such evaluations were infrequent. In a related process, medications were often stopped on a trial basis to see if symptoms returned after discontinuation. Providers thought it important to assess for side effects warranting reactive discontinuation, but patient-reported side effects (eg, dizziness) were felt to be more difficult to ascertain than measurable ones (eg, low blood pressure).
Perception of the patient’s knowledge and beliefs. Another important factor influencing providers’ decisions to discontinue medications was their perception of the patient’s knowledge, beliefs, and preferences about taking medications. Clinicians spoke about how patients’ better comprehension of their conditions and medications often facilitated conversations about stopping a medication, albeit not always resulting in discontinuation. Patients viewed as having limited understanding made it challenging for clinicians to ensure that medications were taken as directed. At the same time, providers recognized that such patients often would “follow doctor’s orders,” thus facilitating discontinuation if called for.
Providers were influenced by patients’ beliefs about a medication's importance; clinicians felt that patients generally prioritize medications that treat symptoms (eg, pain or heartburn). Moreover, some providers endorsed the notion that patients receive “psychological comfort” from continuing medicines they have been taking for a long time.
Medication adherence. A major factor that challenged providers’ ability to make decisions about discontinuation was uncertainty regarding a patient’s adherence to medication. Clinicians often felt that even if they obtained an accurate medication list, questions remained about how consistently adherent patients were to those medicines. Many clinicians spoke about the need to have additional contact with the patient in order to derive an updated medication list—considered essential for prescribing decisions.
Professional identity. Providers discussed how their background and job descriptions created a sense of identity that enabled them to make prescribing decisions. Many spoke about the responsibility they felt to provide appropriate, comprehensive care with ultimate accountability for medication decisions. Previous experience—and success—in discontinuing medications provided confidence for providers to attempt discontinuation in the future. Other providers expressed hesitation to alter regimens for a medication that fell outside their scope of practice, for which they had little experience prescribing, or that another clinician was actively managing.
Provider beliefs about medications. Providers’ perceptions of polypharmacy—its definition, prevalence, and importance—may be critical to decisions about discontinuation. Deciding at what point patients were taking “too many” medications varied, and polypharmacy was considered to be quite prevalent and even the norm. When asked to define polypharmacy, numeric answers ranged from 2 to 12 medications, and many said it depended on the patient.
System FactorsMultiple providers. For patients who obtained medications from providers in other healthcare systems, participants felt limited in their ability to make changes and noted that communication across systems was often difficult. Providers also discussed how the duration and quality of their relationships with patients influenced the patients’ views on whom they trusted most to make decisions.
Workload. Core job duties and perceived workload were frequently cited as influential factors in prescribing decisions. Clinicians viewed discontinuing a medication as a complex process requiring an explicit decision, patient-provider communication, and possibly longitudinal monitoring to assess outcomes. These extra steps were seen as additional work compared with medication renewal.
Most physicians spoke of the potential for nurses and pharmacists to aid in medication reconciliation—an often time-consuming process—to allow time for other clinical issues; however, 1 physician had reservations about tasking another team member with this role. With better schedule capacity to monitor patients closely longitudinally, pharmacists embraced expanded roles that facilitated medication discontinuation.
External directives and policies. Providers felt burdened by the current electronic clinical reminders, several of which are intended for performance measurement. Policies that dictate benchmarking goals were viewed as inhibiting discontinuation because of the need to use medications to achieve targets, regardless of patient age or treatment preferences.
Providers had mixed views regarding the influence of information technology (IT) on medication discontinuation. For example, by using the electronic health record, providers could ascertain whether the pharmacy had recently dispensed medications, at times leading them to discontinue medications that patients were no longer taking. In contrast, inaccurate problem lists led to uncertainties about the appropriateness of medications, and complicated decision making. Providers felt there was untapped potential for IT to aid in clinical decision making specifically related to discontinuing medications, but were wary of adding more electronic reminders or alerts.
Polypharmacy and adverse drug events lead to significant healthcare costs and morbidity, yet there is little information to guide clinical providers in the process of discontinuing medications in the general adult population. Providers in our study had clear, but varying, views on issues surrounding medication discontinuation, which supported the hypothesized domains in our conceptual model of factors affecting discontinuation decisions. Although themes were determined to primarily relate to 1 domain, overlap of codes from multiple domains was not uncommon. For example, indication uncertainty—a medication-related factor—also related to patient and system factors. This interrelatedness is informative for critical focus of interventions in the future to improve medication reviews, patient-provider communication, and systemlevel structures that will result in appropriate medication discontinuation.
Providers expressed concern about the adverse consequences of using medications inappropriately and supported the notion of reducing unnecessary drugs. However, they faced many difficulties in a typical clinical encounter, such as patient complexity, uncertainty, interactions with other healthcare providers, and constraints imposed by the healthcare system.
Participants indicated they generally attempt to minimize pill quantity, and thereby improve patient adherence,13 although this approach was complicated by uncertainties about the medication’s original indication—particularly when another provider initiated the prescription. Given the frequency with which uncertainty was discussed, a potential IT-based intervention that requires each prescription to include the indication could facilitate future decision making. Some healthcare systems do this currently, but its effect on decision making is unknown.14 Including the indication would enhance communication between healthcare systems, and would be helpful when a provider “inherits” a patient from another clinician.
A key factor impacting discontinuation decisions was the provider’s perception of what the patient preferred. Providers may not accurately assess beliefs, and they may overestimate the degree to which they understand their patients.15 The resulting knowledge gaps could result in continued prescribing of a particular drug, despite a lack of provider belief in its medical necessity.16 When a patient feels strongly about continuing a medication, understanding the underlying reasons through high-quality patient-provider communication may enable the provider to deliver information targeted to those beliefs, or lead to a compromise, such as a lower dose. Conversely, identifying patients who welcome a reduction in pills may facilitate appropriate discontinuation.17 Finally, given that patients are nonadherent to approximately half of their medications,18 discontinuing medicines with lower potential benefit may improve adherence to those remaining, with concomitant better health outcomes.19
While research and performance measures frequently use a cutoff of 5 medications to define polypharmacy, many study participants felt that patients often require at least that many to manage their comorbidities. This observation is consistent with other research suggesting that defining polypharmacy using a specific numeric threshold is not clinically useful.20 Consequently, any intervention to raise awareness of polypharmacy must be account for the possibility that 5 may not be a sufficient number of medications to prompt clinicians to review regimens for potential discontinuation. Further, discontinuation may also be relevant for patients not meeting a prespecified numeric definition of polypharmacy.
Clinicians may feel empowered by a successful prior experience discontinuing a medication, leading them to make similar decisions in the future. While our study was not intended to discern differences between types of providers, we did note that physicians appeared to be more comfortable discontinuing medications than other clinical providers, even if the medications were initially prescribed by subspecialists. As the scope of practice for NPs expands, their perceptions of autonomy may change and influence empowerment to make specific prescribing decisions.21 Whether a provider felt “ownership” of a medication also varied, and the ability to make recommendations could be influenced by whom the patient trusted most to make those decisions. Identification of provider characteristics that are associated with proclivity to discontinue medications could provide focus for tailored interventions aimed at the clinical provider.
Interaction between providers was greatly affected by system-level factors. It was felt that better information transmission occurred between providers within the same healthcare system; however, while meaningful use of electronic health IT aims to break down these barriers, data transmission between sites is still limited.22 Having a complete, accurate record of all prescription and nonprescription drugs for a patient is prerequisite to making informed clinical decisions and capitalizing on electronic decision support systems.23
There is potential to tap into the power of electronic systems to support clinicians in determining whether a medication may be appropriate for discontinuation trial, such as when a relevant indication in the problem list is absent, or when the prescription has surpassed a specified duration of time. This consideration must be balanced with “alert fatigue,” so that the alerts are acted upon rather than ignored.24 Participants were particularly frustrated with the use of alerts and clinical reminders that urged increasing healthcare use in order to achieve treatment targets.
We conducted our study at 2 VA medical centers; providers from other geographic regions or non-VA care settings may have different views. We had a limited number of participants, and thus may not be able to discern differences in beliefs among physicians, nurse practitioners, and pharmacists. Providers who volunteered to participate may differ from those who did not respond, and many participants had long tenure in their current position. Future research to survey a broader sample of providers will enable better understanding of the range of opinions and associations with specific provider characteristics. Finally, this study was limited to the perspectives of primary care providers, and patients’ perceptions of medication discontinuation should be examined.
While discontinuing medication can ostensibly be considered “doing less,” doing so in a clinically responsible manner in fact requires a greater investment of time and energy by the provider. Our findings establish a foundation upon which further research can examine how various providers view medication discontinuation, as well as what they identify as barriers and facilitators. This, in turn, will enable development of interventions to reduce the continued prescribing of potentially inappropriate medications.Author Affiliations: Section of General Internal Medicine (AL, SRS, TBM), VA Boston Healthcare System, Boston, MA; Center for Healthcare Organization and Implementation Research, VA Boston Healthcare System (AL, SRS), Boston, MA, and ENRM VA Medical Center (BB), Bedford, MA; Section of General Internal Medicine, Boston Medical Center (AL, SRS), Boston, MA; Division of General Internal Medicine and Primary Care, Brigham and Women’s Hospital (AL, SRS, TBM), Boston, MA; Department of Health Policy and Management, Boston University School of Public Health (BB), Boston, MA.
Source of Funding: The principal investigator was supported by a Department of Veterans Affairs (VA), Veterans Health Administration, Veterans Integrated Service Network (VISN1) Career Development Award, and the study was conducted using resources of the VA Boston Healthcare System. The views expressed in this article are those of the authors and do not necessarily represent the views of the Department of Veterans Affairs. The funding organization had no role in the design and conduct of the study; the collection, management, analysis, and interpretation of the data; the preparation, review, or approval of the manuscript; or the decision to submit the manuscript for publication.
Author Disclosures: The authors report no relationship or financial interest with any entity that would pose a conflict of interest with the subject matter of this article.
Authorship Information: Concept and design (AL, SRS, TBM, BB); acquisition of data (AL, TBM); analysis and interpretation of data (AL, SRS, TBM, BB); drafting of the manuscript (AL, BB); critical revision of the manuscript for important intellectual content (AL, SRS, TBM, BB); obtaining funding (AL); and administrative, technical, or logistic support (AL, SRS, TBM).
Address correspondence to: Amy Linsky, MD, MSc, VA Boston Healthcare System, 150 S Huntington Ave, Bldg 9, Rm 425 (152G), Boston, MA 02130. E-mail: firstname.lastname@example.org.
1. Bushardt RL, Massey EB, Simpson TW, Ariail JC, Simpson KN. Polypharmacy: misleading, but manageable. Clin Interv Aging. 2008;3(2): 383-389.
2. Gurwitz JH. Polypharmacy: a new paradigm for quality drug therapy in the elderly? Arch Intern Med. 2004;164(18):1957-1959.
3. Preskorn SH, Silkey B, Shah R, et al. Complexity of medication use in the Veterans Affairs healthcare system: part I: outpatient use in relation to age and number of prescribers. J Psychiatr Pract. 2005;11(1):5-15.
4. Nebeker JR, Barach P, Samore MH. Clarifying adverse drug events: a clinician’s guide to terminology, documentation, and reporting. Ann Intern Med. 2004;140(10):795-801.
5. Gandhi TK, Weingart SN, Borus J, et al. Adverse drug events in ambulatory care. N Engl J Med. 2003;348(16):1556-1564.
6. Rochon PA, Gurwitz JH. Optimising drug treatment for elderly people: the prescribing cascade. BMJ. 1997;315(7115):1096-1099.
7. Iyer S, Naganathan V, McLachlan AJ, Le Couteur DG. Medication withdrawal trials in people aged 65 years and older: a systematic review. Drugs Aging. 2008;25(12):1021-1031.
8. Garfinkel D, Mangin D. Feasibility study of a systematic approach for discontinuation of multiple medications in older adults: addressing polypharmacy. Arch Internal Med. 2010;170(18):1648-1654.
9. Lee SP, Bain KT, Maio V. Appropriate discontinuation of medications at the end of life: a need to establish consensus criteria. Am J Med Qual. 2007;22(6):393-394.
10. Charmaz K. Premises, principles, and practices in qualitative research: revisiting the foundations. Qual Health Res. 2004;14(7):976-993.
11. Bowen G. Grounded theory and sensitizing concepts. International J Qual Meth. 2006;5(3):12-23.
12. NVivo [qualitative data analysis software; computer program]. Version 9. Victoria, Australia: QSR International Pty Ltd; 2010.
13. Ingersoll KS, Cohen J. The impact of medication regimen factors on adherence to chronic treatment: a review of literature. J Behav Med. 2008;31(3):213-224.
14. Bartlett G, Tamblyn R, Huang A, Kawasumi Y, Petrella L, Dufour E. Evaluation of standardized tasks for primary care physicians using the MOXXI electronic prescribing and integrated drug management system. AMIA Annual Symposium Proceedings. 2003;2003:785.
15. Mulley AG, Trimble C, Elwyn G. Stop the silent misdiagnosis: patients’ preferences matter. BMJ. 2012;345:e6572.
16. Britten N, Stevenson FA, Barry CA, Barber N, Bradley CP. Misunderstandings in prescribing decisions in general practice: qualitative study. BMJ. 2000;320(7233):484-488.
17. Hagendorff A, Freytag S, Müller A, Klebs S. Pill burden in hypertensive patients treated with single-pill combination therapy—an observational study. Adv Ther. 2013;30(4):406-419.
18. Osterberg L, Blaschke T. Adherence to medication. N Engl J Med. 2005;353(5):487-497.
19. Shi L, Hodges M, Yurgin N, Boye KS. Impact of dose frequency on compliance and health outcomes: a literature review (1966—2006). Expert Rev Pharmacoecon Outcomes Res. 2007;7(2):187-202.
20. Viktil KK, Blix HS, Moger TA, Reikvam A. Polypharmacy as commonly defined is an indicator of limited value in the assessment of drug-related problems. Br J Clin Pharmacol. 2007;63(2):187-195.
21. Maylone MM, Ranieri L, Quinn Griffin MT, McNulty R, Fitzpatrick JJ. Collaboration and autonomy: perceptions among nurse practitioners. J Am Acad Nurse Pract. 2011;23(1):51-57.
22. DesRoches CM, Campbell EG, Vogeli C, et al. Electronic health records’ limited successes suggest more targeted uses. Health Aff (Millwood). 2010;29(4):639-646.
23. Romano MJ, Stafford RS. Electronic health records and clinical decision support systems: impact on national ambulatory care quality. Arch Intern Med. 2011;171(10):897-903.
24. van der Sijs H, Aarts J, Vulto A, Berg M. Overriding of drug safety alerts in computerized physician order entry. J Am Med Inform Assoc. 2006;13(2):138-147.