Provider and System-Level Barriers to Deprescribing: Interconnected Problems and Solutions

Driven by an ambition to provide high-quality care, the U. S. healthcare system is focused on safe and appropriate medication use. With an aging population and a high rate of multimorbidity, there has been a concomitant rise in polypharmacy, often defined as five or more medications (Anderson, 2010). Polypharmacy has steadily and significantly risen over the past decade (Kantor, Rehm, Haas, Chan, & Giovannucci, 2015), and potentially inappropriate medications (PIMs) may be seen in up to 79% of older adults (Hill-Taylor et al., 2013). Categories of PIMs include unwarranted medication-related risks, low likelihood of benefit, or treatment misaligned with patients’ goals of care. Polypharmacy and PIMs are each associated with negative health outcomes, including reduced quality of life and increased risk of falls, hospitalizations, and mortality (Reeve, Thompson, & Farrell, 2017). One mechanism by which exposure to polypharmacy and PIMs can be reduced is deprescribing. Deprescribing has been defined as the “systematic process of identifying and discontinuing drugs when existing or potential harms outweigh existing or potential benefits within the context of an individual patient’s care goals, functional status, life expectancy, values, and preferences” (p. 827) (Scott et al., 2015). Deprescribing has been shown to be generally safe and effective. Additionally, in a consumer-driven (i.e., patient-driven) healthcare system, many patients prefer fewer medicines (Linsky, Simon, & Bokhour, 2015). Defining deprescribing as a process highlights the challenges to proactive, intentional discontinuation. There are patient-, provider-, and system-level barriers to deprescribing, many of which are interconnected. Therefore, innovations and initiatives to address these barriers will frequently require a multi-level approach. While patient-level barriers are important to address—especially given that patient involvement has been identified as a key component of successful deprescribing interventions (Reeve et al., 2017)—they are not the sole barriers to safe medication use. Herein, we explore the barriers and opportunities to improve deprescribing at both the provider and system levels. Provider-level barriers to and enablers of deprescribing have been well characterized (Ailabouni, Nishtala, Mangin, & Tordoff, 2016; Anderson, Stowasser, Freeman, & Scott, 2014). In this paper, we utilize the previously-established domains delineated by Ailabouni and colleagues: (a) knowledge and skill, (b) professional role and identity, (c) motivation and goals, (d) environment, and (e) social factors. As expected, many provider-level barriers to deprescribing and their associated potential solutions are interconnected. The discordance between the provider beliefs supporting medication discontinuation and a failure to implement deprescribing in practice may result from a lack of self-efficacy, confidence, and/or competence to deprescribe (Jubraj et al., 2015). Principles of deprescribing are not typically integrated into curricular standards for health professions. A lack of evidence, guidelines, and training may hinder the development of provider self-efficacy to deprescibe. Similarly, despite evidence demonstrating its benefits, many healthcare providers do not routinely practice a critical component of effective deprescribing: shared decision-making (SDM; Gravel, Légaré, & Graham, 2006). Inadequate professional training in the practice of SDM, low motivation to utilize the technique, and time constraints may contribute to the limited observed use. While medical cultures differ across practice settings, regions, and countries, autonomy and professional jurisdiction to deprescribe are inextricably linked to professional identity. A perceived medical hierarchy, where physicians are considered to have greater authority, may lead other disciplines to defer decision-making (Katon, Von Korff, Lin, & Simon, 2001). Similarly, there may be friction regarding the responsibility to deprescribe between primary care providers and specialists. Regardless of professional title, clinicians may be concerned about misperceptions that they are critiquing a colleague’s medical acumen or decision-making if they provide a conflicting care recommendation. Several provider barriers can be described with theories from behavioral economics, many of which also apply to non-medical situations (Tversky & Kahneman, 1992). Loss aversion describes how individuals perceive a loss in value to outweigh the equivalent gain in value. Clinical inertia is when maintaining the status quo is viewed as the easier or more beneficial option. Taken together, discontinuing a medicine once it has been started may be viewed less favorably than continuing it or not initiating it in the first place. Medical training does not inoculate clinicians from these issues. Properties of healthcare infrastructure can act as barriers to deprescribing by introducing uncertainties regarding a medication’s indication, intended duration, and continued appropriateness. In turn, providers may be hesitant to discontinue medications, particularly when the consequences of discontinuation are also frequently unknown. These uncertainties may result from missing scientific evidence about how to safely and effectively withdraw medications. Uncertainty is further compounded when access to a particular patient’s information is not readily available to make real-time clinical decisions. Providers may spend substantial time gathering patient data instead of reviewing the patient’s medication regimen and partaking in SDM. Further, competing visit and system priorities, such as an acute illness or addressing performance metrics, may detract from a provider’s ability to invest time in assessing medication appropriateness. Clinicians often practice in settings with a “more is better” mentality. In the United States, increased spending on healthcare is perceived to correlate with better outcomes, regardless of the data showing a more complex relationship (Baicker & Chandra, 2004; Rajkumar & Swaroop, 2008). Social pressure from patients, peers, and society can lead providers to order more diagnostic and therapeutic options. Additionally, reductions in intensity of care may be viewed as rationing, which is a generally unpopular position in healthcare, regardless of whether done explicitly (e.g., certain Medicaid programs) or implicitly (e.g., via cost-sharing). Clinicians have varying beliefs about the role of medications, promotion of complementary and integrative health modalities (e.g., yoga, acupuncture), and ability to provide access to these options for their patients. As such, medications may be one of the few tangible mechanisms for a clinician to demonstrate their concern for a patient and desire to provide a solution to a patient’s problem. Many solutions to provider-level deprescribing barriers address more than one domain. For example, an increase in the cultural acceptability and expectation for clinicians to deprescribe could mitigate several motivational and social barriers, including principles of loss aversion and clinical inertia, a “more is better” mentality, and knowledge of deprescribing and non-pharmaceutical treatments. Such solutions should capitalize on marketing techniques to enhance the awareness and acceptability of deprescribing to both providers and patients. An additional method may include academic detailing, an approach that provides individualized education and anchoring to peer practices to change provider behavior. Educational initiatives can also overcome deficits in knowledge or skill by intervening upon both practicing providers and healthcare trainees. These can take the form of guidelines, algorithms, or interactive tools to help prioritize medication(s) for deprescribing (Reeve et al., 2017). Similarly, efforts to incorporate SDM into routine practice may lead to increased deprescribing rates by addressing patients’ preferences and providing understanding and support. Shared decision-making may help providers allocate resources to meet patient needs, potentially alleviating concerns related to perceived rationing of care. Inherent to the complexity of current healthcare systems is the frequent multidisciplinary aspect of care. Patients may have multiple providers, with variations in professional training, specialty, and practice location. Enhanced training in and utilization of high-quality interprofessional communication can improve patient care through a better understanding of the rationale for medication-related decisions between providers. Moreover, a deep understanding of professional roles may aid in cross-discipline derivation of deprescribing responsibility. System-level barriers to and enablers of deprescribing are rooted in multiple domains; among them are (a) healthcare policy and finance, (b) information technology, and c) research environment and related incentives. Many system-level barriers to deprescribing and their associated potential solutions have substantial overlap. Issues in healthcare policy and finance, such as resource availability, performance metrics, and reimbursement, can preclude appropriate deprescribing (Anderson et al., 2014). Because of its complexity, deprescribing presents an opportunity for interprofessional engagement, with each professional discipline contributing unique value. For example, pharmacists possess medication knowledge to identify PIMs and develop discontinuation plans; nurses can leverage their relationships with patients to support the implementation and monitoring process; and physicians can provide trust and reassurance to patients. Many practices lack access to these diverse interprofessional team members, as well as social workers and care coordinators. Without well-developed teams to distribute the complex steps of deprescribing, all responsibility remains with an often-burdened individual provider. Moreover, even with access to varied professionals, the autonomy and professional jurisdiction of non-physician healthcare providers to engage in and enact medication management decisions varies across states (Herman, 2017). Performance metrics, intended to facilitate the delivery of evidence-based care, often preclude patient-specific goal-setting and decision-making, directly hindering deprescribing opportunities. For example, not all hypertension and diabetes treatment guidelines and metrics incorporate patient prognosis and goals of care, leading to a conflict between provider performance on quality metrics and what may be best for an individual patient. Deprescribing requires adequate time and reimbursement for cognitive services. It takes more time within a clinic visit to reconsider medication appropriateness and to discuss regimen changes with a patient than it does to maintain the status quo and renew medications. Such changes to medications may also require follow-up monitoring, placing additional pressures on busy clinic environments. Providers may forego deprescribing and SDM due to perceptions of low economic value of these time-intensive cognitive clinical services. Health information technologies, integrated into many aspects of healthcare delivery, can function as both a barrier and facilitator to deprescribing. As noted earlier, uncertainty regarding continued medication appropriateness often results from limited access to patient-related data; without complete information, providers may be reluctant to make regimen changes. Wide-scale interoperability of medical and pharmacy records is limited by a multitude of factors, including a free-market economy for health information technologies. Similarly, while electronic prescribing facilitates medication continuation, the ability to notify pharmacies of medication order discontinuation is not commonplace; as a result, patients may inadvertently continue filling a prescription despite contradicting provider intent (Allen & Sequist, 2012). Research-derived clinical practice recommendations that target a single-organ system can steer clinicians toward the aforementioned one-size-fits-all approach to care delivery. There is also a paucity of patient-specific, individualized prescribing guidance. Both can inhibit effective deprescribing. Many research funding agencies are focused on a specific organ or disease rather than more holistic approaches to human-subjects research, which can influence research priority areas. Translational research, a relatively linear concept, refers to bringing knowledge of disease mechanisms elucidated in laboratories into clinical diagnosis and treatment and findings from clinical studies into usual practice (T2; Woolf, 2008). The existence of information does not automatically result in its uptake into everyday care delivery, with known lags in T2 translation (Westfall, Mold, & Fagnan, 2007). Additionally, proximal outcomes resulting from medication prescribing are relatively easier to measure (e.g., lipids with statin use), whereas avoidance of adverse outcomes that result from deprescribing can be more difficult to identify (e.g., falls avoided due to stopping benzodiazepines). Limited ability to reliably measure this type of outcome can then impede research to identify the most effective deprescribing interventions. Analogous to provider-level solutions, many system-level solutions address multiple barriers. Workforce development opportunities, such as the Health Resources and Services Administration’s (an agency of the U.S. Department of Health and Human Services) Geriatric Workforce Enhancement Program, aim to transform clinical training environments and increase the availability of health professionals through education (Busby-Whitehead, Flaherty, & Potter, 2016). These types of initiatives can augment the autonomy and jurisdiction of non-physician prescribers, enhance access to and collaboration among interdisciplinary team members, and support effective deprescribing. Increasing reimbursements and reducing unnecessary paperwork may also mitigate provider concerns about workload sustainability. New tools that tailor performance metrics and treatment goals to individual patients, as opposed to using standard threshold values, can assuage clinicians’ concerns of penalties for pursuing less aggressive therapeutic goals. This individualized approach has gained some traction for decisions related to screening (e.g., for prostate cancer), but can easily be applied to treatment targets (e.g., lipid levels; Glasgow, Peeples, & Skovlund, 2008). Fostering these types of metrics may de-emphasize one-size-fits-all prescribing algorithms. Disincentives to patients and providers to use PIMs can promote safer medication use. This can take the form of quantity or refill limits, requiring review of continued indication for use, or non-medical consequences (e.g., driving restrictions with particular medications; Canadian Deprescribing Network, 2018). Such measures should be coupled with enhanced availability, access, and awareness of safer alternative pharmacologic and non-pharmacologic measures. However, potential unintended consequences of this approach include restricted patient access to necessary care or enhanced utilization of equally poor alternatives (e.g., switching from benzodiazepines to zolpidem for sleep), and thus should be implemented with caution. Linking medications with indications through interoperable records would provide clinicians with the information necessary for decision-making, reduce efforts spent locating information, and facilitate deprescribing (Schiff, Seoane-Vazquez, & Wright, 2016). Barring direct interoperability, other initiatives to enhance medication record transparency include requiring the submission of all medication data—not just controlled substances—to a state’s prescription monitoring program (Monegain, 2018). Clinicians would then have an adequate and complete medication record upon which to base deprescribing decisions. Surescripts CancelRx feature allows providers to discontinue prescriptions at a patient’s pharmacy, preventing the unintended refill of these medications (Uhrig, 2018). Funding is needed to support research along several steps in the translational pathway. There is little incentive for pharmaceutical companies to study the effects of medication withdrawal; other mechanisms to support these types of investigations are required. Research that evaluates best practices to foster deprescribing is similarly needed. Importantly, efforts are needed to overcome the barriers in translating findings from bench to bedside to clinic. Implementation science research aims to identify the mechanisms that expedite the uptake of evidence-based practices into clinical care. Recent establishment of new funding mechanisms and the development of deprescribing research networks (e.g., the Canadian Deprescribing Network) are intended to support scientific exploration and build momentum for explication of evidence-based deprescribing guidance. Deprescribing is central to safe and appropriate medication use. As delineated and described, provider- and system-level barriers to deprescribing—and their potential solutions—are highly intertwined. Organizations and initiatives that acknowledge the connectedness of such barriers, and thus employ a multi-faceted approach to reducing such barriers, may facilitate effective deprescribing in clinical practice. Dr. Linsky is supported by the Veterans Affairs Health Services Research & Development (grant number CDA 12–166), which had no role in the preparation, review, or approval of the manuscript.