Perspectives on deprescribing in palliative care

Abstract

Introduction:

Pharmacotherapy plays a critical role in the delivery of high-quality palliative care, but the intersection of palliative care and deprescribing has received little attention.

Areas covered:

We conducted a scoping review of English language articles using PubMed to identify relevant publications between January 1, 2000 to July 31, 2022 using search terms of deprescribing, palliative care, end of life, and hospice. We summarize current definitions and developments in palliative care and deprescribing from both clinical and research perspectives. We highlight key challenges and outline proposed solutions and needed research.

Expert opinion:

The future of deprescribing in palliative care requires the development and adoption of individualized approaches to medication management, including a reconsidered approach to communication about deprescribing. Evidence from high-quality clinical outcomes studies is lacking, and the field needs new approaches to coordination of care delivery. This review article will be of interest to both clinical and research-based pharmacists, physicians, and nurses interested in improving care for patients with serious illness.

1. Introduction

Palliative care improves the lives of patients with serious and life-limiting illness and their families through patient-centered clinical care [1]. A serious illness is a health condition that carries a high risk of mortality and either impairs a person's daily functioning or quality of life or excessively strains his or her caregivers [1]. The most common of these conditions are cancer (poor prognosis, metastatic or hematologic), chronic obstructive pulmonary disease (COPD) and heart failure requiring home oxygen or hospitalization, advanced dementia, and end-stage renal disease [2]. An estimated 12 million adults in the United States and 40 million worldwide are living with serious illness and in need of palliative care [3,4].

A growing and consistent body of evidence demonstrates that palliative care leads to better quality of life and quality of care for this population [5-7]. Pharmacotherapy plays a critical role in the delivery of palliative care, and is essential for pain control, symptom management, and optimizing quality of life [8-10]. The first major papers describing prescribing in palliative care were published in the early 2000’s [9,10], and describe the challenge of managing comorbidities in patients approaching the end of life. More recently, ‘deprescribing’, commonly defined as “a process of medication withdrawal, supervised by a health care professional, with the goal of managing polypharmacy and improving outcomes” [11], has emerged as a clinical strategy to address this challenge. While research at the intersection of deprescribing and palliative care is rapidly growing, most of the recent literature focuses on cancer, dementia, and guidance for how to prescribe for end-of-life symptom management while discontinuing unnecessary non-symptom related medications [12-17].

The goal of this manuscript is to provide an updated perspective on palliative care and prescribing, including deprescribing, in the context of the continuum of care of persons with life-limiting and serious illness. We argue that the future of deprescribing in palliative care is grounded in adopting individualized approaches to prescribing and will demand a reconsidered approach to communication, including positivist terminology that focuses on providing care centered in the patients’ goals, which will ultimately allow the patient and family caregiver to play a key role in determining the appropriate timing of deprescribing as a process to achieve goal concordant medication use in serious illness [17].

2. Materials and methods

Using PubMed, we conducted a scoping review to identify English language articles published between January 1, 2000 and July 31, 2022 using search terms of deprescribing, palliative care, end of life, and hospice. Results were screened by two researchers (MK and JT). Additional studies were identified by searching bibliographies. Based on our collective clinical and research experience, as well as this empirical literature, we highlight how palliative care differs from end-of-life care, and then describe developments in palliative care and ‘deprescribing’ from both clinical and research perspectives. We then highlight key challenges for the field, and outline proposed changes in clinical education, clinical practice, and research to move forward.

3. The setting: palliative care, hospice and end-of-life care in the continuum of life-limiting and serious illness

Pharmacotherapy plays a critical role in the delivery of high-quality care for persons with life-limiting illness at each stage of the continuum from wellness to death. Table 1 summarizes key aspects of care along the continuum of a life-limiting illness, including priorities for care, the types of medications used to achieve care goals, prescribing criteria, the unit of care (i.e. patient or patient and family), and the types of clinicians involved at each stage. This continuum of life-limiting illness spans prevention of disease to end-of-life and hospice care. Because high-quality palliative care and appropriate prescribing begins with a clear understanding of the role of medications used at each stage of care, the role of prescribing criteria in guiding medication choice at each stage, and whether they are able to achieve the patients’ goals for care, we review each of these briefly below.

Table 1.

Continuum of Care for Life-Limiting Illness

	Preventive Care	Curative Care	Palliative Care	Hospice Care
Priority for Care	Prevention of disease (e.g. hypertension) or complications of disease (e.g. stroke)	Cure of disease (e.g. eliminate cancer) or minimize disease activity (e.g. flares of multiple sclerosis)	Symptom management & Optimize quality of life	End-of-life symptom management
Predominant Medication Categories	Primary prevention & secondary prevention of disease complication medications (e.g. antihypertensive; glucose lowering)	Disease-modifying and curative medications + Preventive & chronic disease management medications (for comorbid conditions)	Symptom Management + Disease-modifying and curative medications + Preventive & chronic disease management medications	Symptom Management
Clinician Type Involved	Primary care provider	Disease-specialist care provider	Palliative Care specialist +/− Disease-specialist care provider + Primary care provider	Hospice/Palliative Medicine Specialist + Primary care provider
Criteria for Prescribing	Clinical Guideline-directed╲Patient and Family Caregiver Goal-directed
Unit of Care & Decision-making	Patient	Patient	Patient – family caregiver dyad	Patient –family caregiver dyad
Opportunities for Communication	Planting seeds about how some drugs that work now, may not be appropriate later		Aligning with goals to optimize symptom management, stopping drugs with side effects may be better than adding a new drug

For most people throughout their lives, prevention of disease and the complications of conditions like hypertension are the focus, and this goal is supported by the use of medications for prevention, including primary prevention (no disease present) and secondary prevention (disease present but no symptoms), as well as chronic disease medications such as antihypertensives and glucose lowering therapies [9]. For conditions like cancer, drug-based prevention with agents such as chemoprevention medications has received some attention, but the focus of cancer pharmacotherapy mainly involves curative and disease-modifying agents, such as chemotherapy and targeted therapies. Innovations in curative targeted therapies have led to increases in lifespan, disease-free life expectancy, and dramatic improvements in remission and cure rates [18]. For other chronic conditions without known cures, such as multiple sclerosis or rheumatoid arthritis, disease-modifying therapies can significantly improve quality of life [19,20]. For both preventive and curative therapy, clinical guidelines from medical societies and other sources typically guide care delivery and prescribing [21].

Unfortunately, not all diseases have a cure or disease-modifying therapies, and many are accompanied by burdensome symptoms. For incurable or untreatable diseases and conditions that affect quality of life, symptom management eventually becomes the priority for care, and this is the primary domain of palliative medicine [9]. Thus, palliative care is not just about cancer, dementia or dying [16]. Palliative care is the interdisciplinary specialty focused on improving quality of life for persons with serious illness and their families [1]. Sometimes it may be possible to continue curative and disease-modifying treatments along with symptom-focused care in an approach called concurrent palliative care [22,23], which is an approach that has been shown to prolong life expectancy for some patients [24].

Palliative care is often confused with hospice care, which focuses on the period at the end of life for persons with less than 6 months to live who are willing to forgo curative treatments and whose goals of care are comfort. Only focusing on end of life misses enormous opportunities to improve care for patients with serious illness who we may not recognize are dying or who may not be dying. In distinction to hospice care, palliative care may be initiated by the patient with serious illness and their family caregiver when they decide it is appropriate and if services are available [15]. An important distinction for medication prescribing in later stages of life-limiting illness, in contrast to earlier stages in which clinical guidelines drive prescribing decisions, is that patient and family caregiver goals are central to tailoring treatment in palliative care.

4. The problem: the challenge of polypharmacy, adverse drug effects, and drug-related harm in palliative care

Pharmacotherapy is central to care and symptom management across the continuum of life-limiting illness. However, it is important to acknowledge the problems of polypharmacy and the accumulation of symptom management drugs in combination with curative, disease modifying, and preventive drugs in late-stage disease. One challenge is that the complexity of drug regimens increases as symptom targeting drugs are added to regimens of curative, disease-modifying, and chronic condition drugs. The classic study by Currow shows that the average drug burden at end of life increases because of the increasing number of symptom management drugs on a baseline regimen of drugs for comorbid conditions [10]. This dynamic and changing burden of medications and associated disease management presents a number of challenges for patients who are the recipients of these medications.

Polypharmacy can cause new symptoms or exacerbate existing symptoms because of known drug side effects, drug-drug interactions, or adverse drug reactions [25]. A key example in palliative care is opiate-induced constipation [12]. Other common examples include the anticholinergic effects of drugs commonly used in serious illness (e.g. oxybutynin, hydroxyzine) which are well known to contribute to delirium, dry mouth, fatigue and constipation [26]. Other drugs of concern and commonly prescribed include fall risk inducing drugs such as antihypertensives and antidepressants [27].

Further, new symptoms from polypharmacy can contribute to even more polypharmacy through a well-described phenomenon called prescribing cascades [28]. A prescribing cascade occurs when new drugs are prescribed to treat the side effects of other drugs; in such cases, it is preferable to stop the initial offending agent when possible. Leg swelling associated with amlodipine, a dihydropyridine calcium channel blocker, is a common example. While a common response is to start a diuretic to reduce the leg swelling, the more appropriate remedy is to stop the amlodipine. Thus, while the goal in palliative care is to decrease symptom burden, the complicated cocktail of prescribed medications may inadvertently be leading to drug-related harm and an unintentional cycle of increasing polypharmacy. Thus, minimizing harmful medications and prescribing cascades is essential to high-quality palliative care.

5. Current state of the proposed solution: “deprescribing” in palliative care

Reducing unnecessary medication burden is commonly termed “deprescribing”, a term that has no consensus definition but is described as “a process of medication withdrawal, supervised by a health care professional, with the goal of managing polypharmacy and improving outcomes” [11]. Another definition which we prefer, particularly for palliative care, is that it is “the systematic process of identifying and discontinuing drugs in instances in which existing or potential harms outweigh existing or potential benefits within the context of an individual patient’s care goals, current level of functioning, life expectancy, values and preferences” [29]. We note that this also includes dose reduction to a minimally effective dose to minimize potential for side effects. The emphasis in this definition of central importance to palliative care is that the goal is to optimize medications to maximize quality of life and that the key driving factors are patients’ values and preferences in the setting of their remaining life expectancy.

While selective medication withdrawal, dose reduction and discontinuation are critical to high-quality prescribing in palliative care, there are several challenges which need to be understood and addressed to move the field forward. (Table 2) These fall into five categories: communication, drug classification, care coordination; clinician confidence; and clinician time.

Table 2.

Key Challenges for Deprescribing in Palliative Care

	Domain of Challenge	Issues
1.	Communication and messaging	‘Deprescribing’ is a scary term and concept
2.	Classification criteria for identifying medications to withdraw and stop	Explicit criteria (Disease or age-specific) versus Implicit criteria (Patient-centric or goal-oriented) Lack of lack of high-quality clinical outcome studies on deprescribing
3.	Care coordination	No clear model for clinician ownership of medication deprescribing
4.	Clinician confidence	Lack of education for communication approaches Lack of safety data after medication discontinuation Lack of high-quality clinical outcome studies on deprescribing’
5.	Clinician time	Lack of time and reimbursement to review medications, discuss discontinuation, and monitor outcomes

5.1. Challenge #1 – patient-centered communication: “deprescribing” is a frightening term

Communication is central to palliative care. Communication is essential to clearly understanding the goals of care of the patient and their caregiver, to creating a safe space in which to negotiate treatment options, and to providing emotional and spiritual support while presenting treatment plans. This stage of the disease continuum, in which patients and families are adjusting to the possibility of no cure and the growing risk of death, is emotional, stressful, and frightening. The psychological concerns of stopping long term treatments are not well researched, but there is a growing literature which agrees that ‘deprescribing’ reinforces many patient fears such as that clinicians are giving up [9,30].

We can learn from the work that the field of palliative care has done to address their messaging and imaging, and build on those insights [31,32]. For example, a prominent palliative care center in the US offers the following message: “Palliative care is more than end-of-life care; it’s patient and family focused care, delivered with excellence and compassion.” [33] Moreover, it provides “choices in serious illness”. The importance of messaging is underscored by recent work from a stakeholder panel who refined the design of a hospice-oriented deprescribing intervention [34]. One of the stakeholders’ main concerns was the serious negative connotation with the term ‘deprescribing’ and its apparent emphasis on taking therapy away from patients, which exacerbated patient and family fears about death and abandonment. Alternative terminology felt to be more acceptable was to reframe deprescribing as “goal concordant prescribing”, in which prescribing choices are designed to be in alignment with patient goals for care.

5.1.1. Proposed solutions

One strategy would be for the field to draw from communications research, in which studies show that word choice and issue framing matter. For example, in one trial, positively framed messages were associated with greater short-term adherence to a medical intervention than negatively framed messages and usual care [35]. In another experiment in which participants were randomized to one of four different message conditions (i.e. gain, non-loss, non-gain, and loss messages) related to the outcomes of an health behavior, results revealed that gain and non-loss messages were associated with different emotional, cognitive and attitudinal reactions than loss and non-gain messages [36].

Some attention to messaging in a positivist manner was used in the OPTIMZE deprescribing trial for persons with dementia in the outpatient setting [37]. For serious illness, we have proposed a positive messaging approach in our ‘Goal Concordant Prescribing’ program which tries to avoid the word deprescribing [34]. In this program, we train clinicians to consider that goal concordant prescribing, which we define as prescribing that is concordant with the goals of the patient, is centered on optimizing outcomes for the patient, aligning medication changes with patients’ and families’ goals of care, and simplifying medication regimens to achieve the goal of improved quality of life without reducing longevity. This concept builds on foundational issues in palliative care in which the goals of the patient are central to determining therapy; this is particularly critical when a patients’ limited prognosis is understood by the patient and their family [38].

5.1.2. Research needed

There is ample opportunity for this line of research to understand what messaging is most effective and for whom. There are at least two clear opportunities. First, recent work at the intersection of behavioral psychology and mass communications [39], which borrows from the game book of Madison Avenue advertisers, looks at consumer physiologic responses to messaging and the impact of message crafting on tobacco cessation. Such research approaches could be adapted for medication optimization communication. Second, further work needs to empirically test whether a communication reframe to deprescribing can lead to more medication reductions and better outcomes for patients.

5.2. Challenge #2 – drug classification: identifying potentially inappropriate medications is drug focused and not patient-centered

A major challenge is the identification of either unnecessary or potentially inappropriate medications (PIMs) to stop in palliative care. Most work has focused on developing ‘explicit criteria’ for classifying these medications. Explicit criteria lists are lists of drugs or drug classes developed from literature reviews and expert consensus, and are crucial for assessing patients’ medication appropriateness and preventing adverse drug reactions [12,40]. The best known of these is the Beers criteria list which describes drugs harmful to older adults to be avoided, but also include STOPP and START [41,42]. Unfortunately, these do not apply to the end of life and contain many medications appropriate for symptom management in life-limiting illness such as haloperidol and lorazepam.

Two explicit lists are currently in use for palliative care patients: OncPal and STOPPFrail. OncPal targets palliative cancer patients with a life expectancy of six months or less [43]. OncPal was externally validated in 61 patients against expert opinion with high agreement (kappa = 0.83) with an expert panel method and detection of detected one or more PIMs in 70% of 61 cancer patients [43]. A recent retrospective study of 50 patients reported a 75% positive predictive value (PPV) and 91% negative predictive value (NPV) for OncPal compared to expert review gold standard [14]. Limitations of OncPal include lack of guidelines on the prolonged use of thrombosis prophylaxis, lack of guidance on whether specific drugs can be stopped abruptly or need to be tapered, and lack of evidence showing impact on patient-reported outcomes such as quality of life and symptom control [12].

STOPPFrail is an explicit criteria list of 27 medications and is applicable to patients aged 65 or above with limited life expectancy, cognitive impairment, and/or physical disabilities [44]. STOPPFrail was developed using a consensus validation process [44,45]. A recent retrospective review showed a 100% PPV and 90% NPV for STOPPFrail compared to a gold standard of expert review [14]. In a randomized controlled trial in older adults with advanced frailty who were transferring to nursing home care, the use of STOPPFrail-guided deprescribing plans presented to physicians lead to greater reductions in medication use and mean medication costs compared to control, but no change in falls, quality of life, or hospitalizations [46]. A strength of STOPPFrail is that, after appropriate training, it can be applied to complicated medication regimens of 10 or more medications in less than 3 minutes to identify medications that could be stopped [47]. Further, there is published guidance on whether specific drugs can be stopped abruptly or need to be tapered [48]. However, this tool also needs further evalution to demonstrate effect on patient-reported outcomes and use in palliative care.

Alternatively, implicit criteria, which are judgement-based and focused on patient/caregiver preferences, may also be a solution to ensure patients’ medication appropriateness and safety [40, 49, 50]. The Medication Appropriateness Index (MAI) is the globally-recognized implicit drug classification method [51,52]. Using the simplified MAI called the “Unnecessary Drug Use Measure” in a geriatric palliative care unit, an observational study identified medications without an indication, without evidence of effectiveness for the prescribed indication, or that were being are used for extended period beyond what was indicated [53]. The study showed a decrease in the number of unnecessary medications from 1.7 ± 1.5 at baseline to 0.6 ± 0.8 at study end (p=0.003); the baseline number of daily medications was 9.7 ± 4.3. A decrease in all three categories of the Unnecessary Drug Use Measure was measured [53].

5.2.1. Proposed solutions

An alternative to purely explicit or implicit approaches is a hybrid approach melding the explicit criteria of the STOPPFrail with an implicit approach to assesses and incorporate patient goals into decision-making. This strategy is used in the “Goal Concordant Prescribing” approach developed for palliative and hospice care [34]. The melded hybrid approach is suggested by van Merendonk et al. who proposed that the OncPal guideline could be merged with step-by-step procedures that can facilitate deprescribing in a patient-individualized manner [12].

5.2.2. Research needed

Further research is needed to understand feasibility, usability, fidelity, and ultimately impact of applying different PIM classification schema on outcomes of interest to patients receiving palliative care, including quality of life, independence in daily functioning, cognition, patient satisfaction with their care plan. Since palliative care is at the critical juncture of the transition from curative to symptom management care, we argue that a hybrid approach melding explicit lists and patient and family goals for care needs further development and attention.

5.3. Challenge #3 - care coordination: a model of care coordination across the continuum of serious illness is needed

Patients and caregivers can and should receive symptom-oriented palliative care at any point in their trajectory of their life-limiting disease. However, engagement in palliative care and deprescribing typically occurs late in patients’ disease trajectory, either in hospice care or as death approaches. In addition, patients, caregivers, and health care providers have limited knowledge regarding the potential benefits of deprescribing and are often unprepared for these conversations when they ultimately do occur. As result, there may be delayed uptake of deprescribing and limited associated benefits of reducing medication burden.

Multiple healthcare providers are often engaged in the treatments of patients with multimorbidity, and inadequate patient-provider communications can still be found [54]. According to a qualitative study, inconsistent medical advice, repeated lab testing procedures, and polypharmacy were found among patients who visited more than four providers [55]. Additionally, one qualitative study reported that specialists did not communicate with primary care providers for patients’ medication changes, which can cause inappropriate prolongation of treatments [56].

5.3.1. Proposed solutions

Engaging patients and caregivers earlier in palliative care will require additional or reallocation of resources as well as revisions to processes and procedures to ensure optimal coordination of care. As we describe below, additional resources include increasing overall knowledge regarding palliative care among health care providers and the confidence to use that knowledge. There is currently a critical shortage of specialty trained palliative care providers [57]. However, all health care providers should have the capacity to provide primary palliative care which includes basic symptom assessment and management, communication regarding goals of care, and care coordination [58]. Furthermore, because only half (52%) of hospice patients are referred to hospice care from an acute care hospital, health care providers in non-acute settings (e.g. primary care, nursing homes, and assisted living facilities, should also have this capacity [59].

Earlier engagement in palliative care can “plant seeds” for deprescribing conversations that prepare and familiarize patients and caregivers with the potential benefits of reducing medication burden. (Table 1) This could build on the benefits of decreased symptom burden and improved quality of life among patients and caregivers coming from early palliative care interventions that have not traditionally included a focus on deprescribing [24]. In this way, early converastions about the anticipated changes in medications’ benefit-risk profiles that are attributable to expected changes in metabolism and the pathophysiology of death could help prepare patients and families for downstream medication changes and withdrawl of preventative and non-symptom oridented drugs.

Resources needed for care coordination to support palliative care and deprescribing also include: 1) infrastructure to support earlier identification of patients likely to benefit from palliative care and deprescribing in whom to initiate goals of care/serious illness conversations, 2) a framework to facilitate these conversations, and 3) systems to efficiently and accurately communicate patients’ health information, including goals of care, during hand-offs between health care providers and transition between health care settings.

5.3.2. Research needed

Methods for identification of patients likely to benefit from palliative care and deprescribing are needed. These include tools or algorithms that can be imbedded into the electronic health record and that utilize information including patients’ primary and comorbid diagnoses, incidence of PIM prescribing, medication burden, and other patient factors (e.g. difficulty with daily activities) which may indicate that patients’ symptoms are worsening and impacting their quality of life [9, 60, 61]. Administrative data alone may be inadequate to prospectively identify patients with serious illness who can benefit from palliative care; research to improve the sensitivity of existing methods is needed [2].

While there are several models to facilitate goals of care conversations, there is little available to guide goals of care with respect to deprescribing. An adaptation of the Serious Illness Conversation Guide, a tool that focuses on patient centeredness, could help align medication decisions with patients’ goals of care [62,63]. Such tools have been developed by Weir at al [64] to guide patient conversations with pharmacists and Tjia et al for the palliative care setting [34]. Current recommendations suggest that goals of care conversations should occur early and often in the disease trajectory. Additional research is needed on how best to implement these strategies, promote uptake, and achieve continuity of care [65]. Lastly, there is limited existing research, models of care, and associated infrastructure for coordination of care among patients with serious illness. Technological advances to include this information in electronic health records and registries and make it available to health care providers within and across health care settings may help with this. Further research can also help better define roles for who should manage and ensure communication of this information.

5.4. Challenge #4 - clinician confidence: the need for more education and evidence

For goal concordant prescribing and deprescribing to advance in clinical settings, there is a need to improve clinician confidence in deprescribing as a major area requiring greater attention [15]. Specifically, to achieve the goal of a patient-centered medication regimen, it is important to advance provider knowledge of and confidence in deprescribing for palliative care patients. Key challenge areas include: lack of high quality evidence for benefit and harms of deprescribing interventions in limited life expectancy; inadequate clinician awareness of available resources and knowledge to engage in deprescribing effectively; insufficient palliative care education for all clinicians; a fragmented healthcare system in need of greater care team integration and interdisciplinary care models including those which involve pharmacists with expertise in medication management among patients with serious illness [66].

5.4.1. Proposed solutions

First, there is a need for high quality randomized controlled trials and evidence-based guidelines to identify and safely stop or taper medications in palliative care, and approach deprescribing conversations with patients and their caregivers in safe and effective way [67-68]. To date, there is a single, high-quality, randomized controlled trial in hospice-eligible patients of statin discontinuation; this trial demonstrated better quality of life and no increased risk of harm among the group discontinuing statins [69]. Second, there is the need for educational interventions for all clinicians involved in prescribing. For example, a cross-sectional study to assess deprescribing knowledge and attitudes of 131 pharmacy students in the US found that fewer than 50% of pharmacy students considered that their didactic training on deprescribing in the clinical setting was sufficient [70]. Third, several authors noted that clinician confidence can be addressed by leveraging the strength of pharmacist expertise either by holding multidisciplinary team meetings or using pharmacist-conducted medication reviews [71,72]. Increasing the availability of geriatric or palliative care pharmacists including by increasing the number of pharmacy residency training opportunities may be needed to meet the demand for team-based care for older adults with serious illnesses. While the recommendation in UptoDate suggests to “[u]se the expertise and services of pharmacists, nurses, and other health care professionals, either to lead the process, to complete time intensive tasks, or as part of a multidisciplinary team.” it is estimated that only 8% of interdisciplinary palliative care teams have a pharmacist [73-79]. Finally, we need a model of care for coordinated communication across the continuum of care and provider specialties to facilitate deprescribing by different members of the prescribing team, including primary care providers, disease-specialists, and palliative care clinicians. Physicians tend to be concerned if they should deprescribe medications given by other clinicians and/or if deprescribing would change their patient-doctor relationships [80].

5.4.2. Research needed

There is ample opportunity for research to increase clinician confidence, starting with the need for more evidence of outcomes following medication discontinuation and withdrawal. While there are several systematic reviews summarizing the benefits of deprescribing, few have focused on prescribing in palliative care or the end of life and there is consensus that more work needs to be done with adequately powered observational and experimental studies [8, 66, 81-87]. Further, the outcome measures for such trials need to be expanded to include not only the number of medications being prescribed and deprescribed, but also adverse drug reactions (ADR) and adverse drug withdrawal events (ADWE), potential adverse emotional or psychological effects for family caregivers, and ultimately whether a deprescribing intervention achieved goal-concordant prescribing.

In addition to clinical trials, there is an opportunity for leveraging existing real-world data to monitor prescribing patterns, identify opportunities for interventions, motivate changes in clinical behavior, and generate evidence to support clinical decisions including through the development of consensus criteria and recommendations [88]. The quality and comprehensiveness of real-world data (i.e., data collected as part of routine care rather than for explicit research purposes) available to researchers is expanding rapidly. Technological advances have increased access to and linkages between various data sources with utility for the study of deprescribing in the palliative care setting such as self-reported data from patients and caregivers, functional status and physiologic data streams from wearable devices, clinical data from electronic health records, and healthcare utilization information from health care payers. These latter traditional administrative datasets including enrollment, claims, and encounter data maintained by payers are a rich and still widely used resource for pharmacoepidemiologic research.

With essentially complete capture of medications that are covered by a health plan, administrative data are well-suited for research that addresses questions regarding medication utilization patterns. Ongoing research to monitor prescribing patterns is needed among those with serious illnesses to identify potentially hazardous drug use with adequate prevalence to warrant a response from guideline promoting professional societies. Further, as health systems implement increasingly sophisticated forms of clinical decision support and data-driven quality improvement initiatives, it may be desirable to analyze administrative data sources and provide timely feedback to clinicians regarding the extent to which their prescribing patterns for patients with serious illnesses deviate from their peers and from best practice recommendations in a manner consistent with elements of clinical performance feedback intervention theory [89]. This would require payer targeted education regarding palliative prescribing, to avoid inadvertently sending messages to clinicians about “inappropriate” medications which are actually effectively palliating symptoms. Beyond tracking prescribing patterns, with continued advances in statistical and pharmacoepidemiologic methods real-world data are often suitable to answer questions of safety and effectiveness [90]. By leveraging real-world data sources, preliminary studies can be performed relatively efficiently which may justify the need for an expensive clinical trial, and in other circumstances rigorous pharmacoepidemiologic research may be the only feasible or ethical approach to generating novel information to guide care for complex patients such as candidates for palliative care and deprescribing. However, challenges remain with identifying appropriate comparison groups for observational studies of deprescribing since such research is particularly susceptible to confounding by indication by unmeasured sources that are difficult to address analytically. Recommendations for conducting observational deprescribing research and overcoming such challenges are presently absent, but are under development within the geriatric pharmacoepidemiologic research community.

In addition, we need research about the optimal educational strategies to enhance confidence that reaches a wide array of clinicians, particularly those who do not specialize in palliative care. For example, one prospective pre-post study used small-group osteoarthritis management workshops and a ‘decision tree’ tool for 249 family doctors that increased provider confidence [91]. Comparable workshops to increase clinician confidence in palliative and end of life care exist but are not widely available [34].

Finally, we need implementation science models for how to best utilize the limited and valuable resource of palliative care-trained pharmacists. For example, a randomized controlled intervention reported that interdisciplinary meetings with various prescribers, clinical pharmacologists and/or trained pharmacists who supported physicians’ deprescribing practice were successful education intervention to reduce inappropriate prescribing [77]. While various models of pharmacist involvement have shown promising results, such as multidisciplinary team meetings, and pharmacist-conducted medication reviews, the limited availability of such resources necessitates and increase in the supply of such professionals and demands we understand which models are most effectiveness, feasible, and sustainable [87]. We know of few studies of direct physician-driven deprescribing in palliative care [92].

This highlights the importance of systematic and ongoing training of pharmacists in palliative care. There are increasing, but still insufficient, opportunities to provide specialty training for pharmacists in palliative care and pain management. There are currently only 25 accredited post-graduate year 2 pharmacy residency programs in palliative care and pain management in the US with 6 additional programs seeking accreditation [93]. The American Society for Health-System Pharmacists provides educational goals and objectives for these programs and there is also a professional society, the Society of Pain & Palliative Care Pharmacists which helps support continuing education for pharmacists in the workforce [94].

5.5. Challenge #5 – clinician time: reimbursement and support for clinician time are needed

Inadequate practitioner time and a mismatch between value-enhancing care activities and what is reimbursable under traditional fee-for-service payment models is a pervasive and insidious problem in healthcare. In a fragmented healthcare system, taken together with challenges to coordination of care described above, time constraints can lead to an unintentional practice of expecting the other clinician to do the time-intensive work of ascertaining patient goals, reviewing medications, and then aligning the medications with the patients’ goals of care. Several approaches can be of help.

5.5.1. Proposed solutions

Recommendations for the practice of intentional, patient-centered, medication management include, “Setting aside separate visit(s) to focus on medication review and deprescribing and using reimbursement/billing codes can include the diagnosis that the medication is intended to treat or general wellness/annual health visits. And starting the process; deprescribing doesn’t need to be conducted all in one visit” [73]. Further, to reimburse for the time required for medication review, programs such as Medicare Part D Medication Therapy Management programs in the US and Home Medicines Reviews in Australia may be options for reimbursed reviews [95,96]. A transitions of care medication review supported by a transition of care billing code or the advanced care plan billing code in the US may be an additional avenue for paid time for review [97-100]. This type of review could also be built into the scope of work of palliative care clinicians who can bill for visit time or expertise for their role helping the transitions of care of patients with serious illness. Finally, the shift towards alternative payment models such as accountable care organizations represents an opportunity for delivery systems to realign their delivery systems to augment high-value activities like palliative care that benefit patients and the system, rather than continuing with the relentless pursuit of increasing the volume of billable services as an exercise in revenue maximization.

5.5.2. Research needed

There is a conspicuous paucity of research about care coordination models, policies, and reimbursement needed to address the challenge of deprescribing. Lessons learned from the US Medicare transition of care billing codes and ACP billing codes, both of which could be used in this setting, are sorely needed [97-101].

6. Conclusion

Palliative care is essential to optimizing clinical care delivery that enhances the quality of life of persons with serious and life limiting illness and their families. Pharmacotherapy is an essential component of high-quality palliative care. Making decisions about which medications to continue or stop as life-limiting illness proceeds toward death is critically important and clinically challenging. Additional training and research is needed to enhance the capacity, communication skills, evidence, and care coordination models available to the multidisciplinary teams providing care for patients with serious illness and their families. Investments in these areas promise to bring the benefits of deprescribing and goal concordant prescribing into greater realization in clinical practice. Our hope is for people with serious illness to live their fullest lives for longer with the support of medications and with greater control over and informed participation in prescribing choices.

7. Expert opinion

This perspective on deprescribing in palliative care highlights the progress made in the area of managing prescription medication use in patients with life-limiting a serious illness in the last 20 years. Our review highlights overall progress that has been made in research and clinical practice. We note that the greatest progress has been achieved in the identification of potentially inappropriate medications (PIM) in life limiting illness, as well as in the development of deprescribing interventions. Explicit criteria to identify PIMs has been developed applicable to patients with cancer (OncPal) and patients aged 65 years and older (STOPPFrail). Implicit criteria approaches based on PIM checklists have been adapted from the Medication Appropriateness Index. Future clinical efforts and research needs to bring both of these approaches together in a melded approach that combines explicit lists with implicit criteria that ascertain and tailor prescribing to patients’ goals of care as it progresses over time. In contrast, less progress has been made in the areas of patient-centered communication, care delivery and care coordination across the continuum of life-limiting illness, and clinician confidence and health system reimbursement. Key opportunities to advance communication in deprescribing in palliative care lie in reframing medication discontinuation in a positive light and in the possibility of using marketing research methods to refine messages for all stakeholders, including patients and family caregivers. For care delivery and care coordination, there are opportunities to upstream deprescribing conversations by ‘planting seeds’ to prepare patients and family caregivers for downstream deprescribing conversations that are aimed at refining medication regimens when symptom management becomes the main goal of care. To improve clinician confidence, there is a continued need for outcomes research based on clinical trial and observational study designs that inform the effectiveness, benefits, and safety of approaches to improve palliative care prescribing and deprescribing. And to improve real world implementation of deprescribing in clinical practice, significant efforts are required to develop feasible and sustainable practice models that allow for periodic, meaningful conversations between patients and clinicians that will ultimately ensure that deprescribing is concordant with patient wishes. This is an exciting time for deprescribing research given the recent establishment of international networks of deprescribing researchers across the globe. Taken together with the growing interest in deprescribing, advances in methodology, and greater appreciation of the importance of addressing the need to optimize prescribing in palliative care, we can expect that there will be some progress for many of these issues in the next 5 years.

Article highlights.

• Pharmacotherapy plays a critical role in the delivery of palliative care, and is essential for pain control, symptom management, and optimizing quality of life.

• Deprescribing has no consensus definition and in palliative care is best conceptualized as “the systematic process of identifying and discontinuing drugs in instances in which existing or potential harms outweigh existing or potential benefits within the context of an individual patient’s care goals, current level of functioning, life expectancy, values and preferences.”

• The future of deprescribing in palliative care is grounded in adopting individualized approaches to prescribing and will demand:

  • a reconsidered approach to patient-centered communication, including messaging and positivist terminology that focuses on providing care centered in the patients’ goals;
  • a melded approach to identifying potentially inappropriate medications in palliative care that is based on both explicit and implicit criteria approaches;
  • clinical care coordination across the continuum of life-limiting illness that ‘plants seeds’ that prepare patients and family caregivers for deprescribing in palliative care;
  • an evidence base that bolsters clinician confidence in the effectiveness and safety of deprescribing; and
  • models of care that address constraints of clinician time and reimbursement.