Clinician and Family Caregiver Perspectives on Deprescribing Chronic Disease Medications in Older Nursing Home Residents Near the End-of-Life

Abstract

Introduction:

Nursing home (NH) residents with limited life expectancy (LLE) who are intensely treated for hyperlipidemia, hypertension, or diabetes may benefit from deprescribing.

Objective:

This study sought to describe NH clinician and family caregiver perspectives on key influences on deprescribing decisions for chronic disease medications in NH residents near the end-of-life (EOL).

Methods:

We recruited family caregivers of Veterans who recently died in a VA NH, known as Community Living Centers (CLCs), and CLC healthcare clinicians (physicians, nurse practitioners, physician assistants, pharmacists, registered nurses). Respondents completed semi-structured interviews about their experiences with deprescribing statin, antihypertensive, and antidiabetic medications for residents near EOL. We conducted thematic analysis of interview transcripts to identify key themes regarding influences on deprescribing decisions.

Results:

Thirteen family caregivers and 13 clinicians completed interviews. Key themes included: 1) Clinicians and caregivers both prefer to minimize drug burden; 2) Clinical factors strongly influence deprescribing of chronic disease medications, with differences in how clinicians and caregivers weigh specific factors; 3) Caregivers trust and rely on clinicians to make deprescribing decisions; 4) Clinicians perceive caregiver involvement and buy-in as essential to deprescribing decisions, which requires time and effort to obtain; 5) Clinicians perceive conflicting care from other clinicians as a barrier to deprescribing.

Conclusions:

Findings suggest a need for efforts to encourage communication with and education for family caregivers of residents with LLE about deprescribing, and to foster better collaboration among clinicians in CLC and non-CLC settings.

1. Introduction

Older adults residing in nursing home (NH) settings experience a high medication burden that often persists in the context of advanced dementia (AD) and/or other life-limiting conditions (i.e., progressive conditions with an expected survival of months or less) [1]. Previous studies of NH residents with limited life expectancy and/or advanced dementia (LLE/AD) have found polypharmacy to be common, with 20–80% taking 10 or more medications [2–5]. The use of potentially inappropriate medications (e.g., drugs and/or dosages of medications to avoid due to age, comorbidity, or risk of drug-drug interactions) is also common in NH residents [6–8], with an estimated prevalence of 43% according to a systematic review [8] of studies using explicit criteria (e.g., AGS Beers Criteria^® or STOPP/START [9,10]). Continued prescribing of chronic disease medications with reduced potential for benefit in context of LLE/AD contributes to unnecessary polypharmacy in NH residents [11]. Attention to deprescribing, defined as discontinuing or reducing the dose of medications when risks outweigh expected benefits in context of the patient’s goals of care, life expectancy, values, and preferences, is therefore necessary for NH residents with LLE/AD [12].

Recent data suggests that efforts are specifically needed to reduce overtreatment for diabetes, hypertension, and lipid control through deprescribing in NH residents with LLE/AD. Guidelines recommend against overtreatment of diabetes in this population, defined as receiving antidiabetes medications to maintain glycated hemoglobin (HbA1c) of <7.5%, because of concerns about associated hypoglycemia and increased cardiovascular risk [13–16]. However, a recent Veterans Affairs (VA) study identified 44% of NH residents with diabetes and LLE/AD as potentially overtreated for diabetes [17]. These findings concur with previous studies in both VA and non-VA settings [18–20]. Hypertension guidelines are less clear about optimal blood pressure (BP) targets in older adults with LLE/AD, but acknowledge potentially increased risks (e.g., postural hypotension, falls, fractures, mortality) and reduced benefits of intensive treatment [21]. Yet, a recent VA study found that 34% of NH residents with LLE/AD were being treated to a systolic BP of <120 mmHg [22], a widely used standard for potential overtreatment in this population [23–25]. Similarly, statins are prescribed to more than one-third of NH residents with LLE/AD [3,4], despite guidance from geriatric and palliative care groups recommending against using statins in older adults with less than two years life expectancy or AD due to lack of clear benefit [26–29].

Recent studies evaluating health outcomes associated with deprescribing antidiabetes agents, antihypertensives, and statins provide some evidence that these medications can be safely deprescribed in potentially overtreated patients with LLE/AD [25,30–32], although evidence is limited and, in the case of statins, inconsistent [33]. In practice, only a minority have these medications de-intensified or discontinued after NH admission [4,17,22,34]. In prior quantitative studies, we identified many clinical, family caregiver, clinician, organizational, and policy factors that were associated with deprescribing of chronic disease medications in potentially overtreated VA NH residents with LLE/AD [4,17,22,35]. However, how clinicians and family caregivers communicate and weigh different factors when making decisions to deprescribe or continue these medications is less understood.

To address this gap and inform the development of interventions to support appropriate deprescribing of chronic disease medications for NH residents with LLE/AD, we conducted semi-structured interviews with family caregivers of recently deceased NH residents, as well as NH clinicians. Our specific objective was to identify key themes regarding influences on deprescribing decisions for chronic disease medications in NH residents near the end-of-life.

2. Methods

2.1. Study Design

This was a qualitative study using semi-structured telephone interviews with two groups of respondents: 1) family caregivers of Veterans who had recently died while residing in a VA NH, known as Community Living Centers (CLCs); and 2) clinicians practicing in VA CLCs. Enrollment continued until thematic saturation was reached within each group of family caregivers and clinicians. To maximize capture of diverse perspectives on deprescribing, including organizational factors, we sought to enroll roughly equal numbers of respondents (within caregiver and clinician groups) from CLCs with relatively lower vs. higher rates of deprescribing antidiabetic agents, antihypertensives, and statins among residents with LLE/AD observed in prior quantitative studies [4,17,22]. Specifically, recruitment efforts targeted roughly equal inclusion of both clinicians and caregivers across two groups of CLCs: 1) CLCs with rates of deprescribing for all three medication classes that fell into the lowest four deciles across all 134 CLCs; and 2) CLCs with rates of deprescribing in the highest four deciles.

2.2.1. Family Caregiver Recruitment

We reached out to potentially eligible family caregivers of Veterans aged ≥65 years who had recently died in selected CLCs through routine contacts made by the Veteran Experience Center (VEC), a national VA quality improvement program that administers the VA Bereaved Family Survey (BFS). The VEC contacts the next-of-kin listed in the VA electronic health record (EHR) for all Veterans who die in a VA facility, including CLCs. Next-of-kin are mailed the BFS survey approximately 4–6 weeks after the Veterans’ death and are sent a reminder letter about the survey 4–8 weeks later. We included recruitment announcements for the current study in these reminder mailings. Only next-of-kin for Veterans aged ≥65 years, who died at one of the 15 selected CLCs within the past 8–12 weeks, and who met claims-based diagnosis code criteria [36] for coronary artery disease, stroke/transient ischemic attack, diabetes, or hypertension were sent announcements.

Interested next-of-kin participants provided verbal informed consent and were screened to confirm that they met study inclusion criteria: 1) Veteran diagnosis with at least one qualifying chronic condition listed above; 2) they felt able to evaluate the care received by the Veteran in the last month of life; and 3) they were able to complete the interview in English. Eligible participants then scheduled a telephone interview and were paid $50 after completing it. Thirteen family caregivers were enrolled and completed interviews.

2.2.2. CLC Clinician Recruitment

We used the VA Corporate Data Warehouse to identify potentially eligible VA clinicians practicing at one of the selected CLCs. Eligible clinicians included physicians, nurse practitioners (NP), physician assistants (PA), pharmacists, and registered nurses (RN). We sought representation from each type of clinician. An initial recruitment email was sent to potentially eligible clinicians, with follow-up emails sent two and four weeks later. The email instructed clinicians to contact the research coordinator via email or telephone. After providing verbal informed consent, clinicians were screened to confirm that they met inclusion criteria: 1) qualifying clinician type of physician, NP, PA, pharmacist, or RN; 2) currently practice at a VA CLC; and 3) able to complete the interview in English. Interested clinicians then scheduled and completed a telephone interview. We also included a prompt in the interview guide to allow for snowball recruitment of additional clinicians. No monetary participation incentive was provided, in accordance with VA policies regarding clinician research participation. Fourteen clinicians were enrolled, with 13 completing the interview.

2.2.3. Data Collection

Interviews were audio-recorded and conducted by two experienced interviewers (GK and LS), using semi-structured interview guides drafted by the study team and pilot-tested with colleagues outside of the team. Interview length was 30–45 minutes.

The caregiver interview guide began with questions about the caregiver’s relationship to the Veteran, involvement with the Veteran’s care, including end-of-life care at the CLC, and involvement in and satisfaction with medication decision-making. The interviewer then asked more focused questions about the caregivers’ discussions with clinicians about the Veteran starting, continuing, decreasing, and stopping medications in the CLC. Questions encompassed exploration of perspectives about deprescribing in general as well as specific probes regarding medications for lipid, diabetes, and hypertension control. Caregiver demographics, including age, gender, race, Hispanic ethnicity, and education level, were collected at the end of the interview.

Separate clinician interview guides were used for prescribers in the VA (i.e., physicians, NPs, PAs, pharmacists) versus those without prescribing privileges (i.e., RNs). The prescriber guide began by asking about factors they consider when starting/increasing and decreasing/discontinuing statin medications for lipid control in their patients with LLE/AD, including circumstances for making changes, who is involved in this decision, information sources used, and situations in which a statin would be restarted after stopping or reducing. Similar questions were asked about medications to treat hypertension and diabetes and included probes about treatment targets for these conditions, respectively. Prescribers were then asked about experiences discussing deprescribing decisions with patients and family caregivers, including discussion content, barriers and facilitators to having these discussions, barriers and facilitators to actually implementing deprescribing decisions based on these discussions, and VA organizational efforts they have experienced to promote or support deprescribing. The interview guide for RNs was very similar but was phrased to elicit their impressions of the clinician care teams’ experiences with medication decision-making and deprescribing in the CLC rather than personal experiences as a prescriber. The interview concluded with demographics questions (gender, race, Hispanic ethnicity, years in clinical practice, years practicing in VA and the current CLC, certification or specialized training in geriatrics, and whether they also provided care to residents in non-VA settings).

2.2.4. Analytic Approach

During data collection, we held bi-weekly study team meetings, including interviewers and coders. Ongoing discussions allowed us to determine when saturation for family caregivers and for clinicians was reached [37]; that is, when sufficient data was collected to draw conclusions due to similar information obtained and further data collection would not provide additional meaningful insights.

NVivo 12 (Lumivero, Denver, CO) software was used for analyses. Audio recordings of interviews were transcribed verbatim. We conducted thematic analysis using the Editing Approach by Crabtree and Miller [38] via the following steps: 1) avoid preconceptions that may bias the analytic process; 2) identify units of text meaningful to the central research questions; 3) sort units into categories or “codes” that are maintained in a codebook; 4) understand how codes interrelate to gain insight into the study’s research question (i.e., “connecting phase”); and 5) select quotations that best represent the codes and decide the final arrangements of codes. Three study team members served as coders (GK, LS, JR); the coding of clinician (GK, LS) and caregiver (GK, JR) interviews was conducted separately. For each set of interviews, two coders first independently coded three transcripts and met to compare codes and resolve discrepancies. The primary coder (GK) then coded all additional transcripts. Common factors influencing deprescribing decisions were identified and connected across clinician and family caregiver interviews, and emerging themes were shared with the analytic group as a further check on validity.

3. Results

3.1. Sample Characteristics

As shown in table 1, nearly all family caregivers (n=13) were female (12/13), white (12/13) and non-Hispanic (12/13). Seven family caregivers were spouses of the resident, three were adult children, and three reported other relationships, and over half were age 70 years and older (7/13). The clinician sample (n=13; Table 2) consisted of six NPs, three physicians, two pharmacists, and two RNs. Clinicians were mostly female (10/13), six were white, and 12 were non-Hispanic. Most had practiced medicine for over 10 years (6/13). A majority had certification or training in geriatrics (9/13). Two clinicians also saw residents in a non-VA setting.

Table 1.

Self-Reported Characteristics of Family Caregivers

	Family Caregivers (n = 13)
Gender
Female	12
Male	1
Race/Ethnicity
White	12
Black	1
Ethnicity
Non-Hispanic	12
Hispanic	1
Age
<70	5
70–74	6
75+	2
Relationship to Veteran
Spouse	7
Child	3
Sibling	1
Cousin	1
Friend/Neighbor	1
Education
High School/GED	6
Associate Degree	2
Bachelor’s Degree	2
Master’s Degree/Doctorate	3

Table 2.

Self-Reported Characteristics of Community Living Center Clinicians

	Clinicians (n = 13)
Gender
Female	10
Male	3
Race/Ethnicity
White	8
Asian	4
Native Hawaiian and Other Pacific Islanders	1
Ethnicity
Non-Hispanic	12
Hispanic	1
Credentials
Nurse Practitioner	6
Physician	3
Pharmacist	2
Registered Nurse	2
Total Years in Clinical Practice
<10 years	5
10–19 years	5
20+ years	3
Years in Clinical Practice in the VA
<10 years	9
10–19 years	3
20+ years	1
Geriatric Certification/Training
Yes	9
No	4
Sees Residents in Non-VA Setting
Yes	2
No	11

3.2. Key Themes

We identified five key themes related to clinician- and caregiver-perceived influences on deprescribing chronic disease medications in NH residents near the end-of-life. These included: 1) Clinicians and caregivers both prefer to minimize drug burden; 2) Clinical factors strongly influence deprescribing of chronic disease medications, with differences in how clinicians and caregivers weigh specific factors; 3) Caregivers trust and rely on clinicians to make deprescribing decisions; 4) Clinicians perceive caregiver involvement and buy-in as important influences on deprescribing decisions, which require time and effort to obtain; 5) Clinicians perceive conflicting care with other clinicians as a barrier to deprescribing. Below we discuss these key themes in more detail and provide supporting quotations by clinicians and family caregivers in Table 3.

Table 3.

Themes and Supporting Quotations Related to Deprescribing Chronic Disease Medications in CLC Residents Nearing End of Life

Themes	Clinician Quotes	Caregiver Quotes
Key Theme 1: Clinicians and caregivers both prefer to minimize drug burden	“…life expectancy is only 6 months, and I want to decrease the number of medications. And they are at the verge of refusing the medications most of the time, so I want to give them minimal medications and just to have a goal as a comfort.”	“Yeah, because both of us – my husband and I – we’re both of the opinion that the fewer medications he would take would be better for him. We were never ones for a lot of medications.”
Key Theme 2: Clinical factors strongly influence deprescribing of chronic disease medications, with differences in how clinicians and caregivers weigh specific factors	”…if the patient has a prior history of advanced coronary artery disease with medical stent, or if the patient has a history of bypass, a patient is a poorly controlled diabetic patient who we are managing aggressively with oral hyperglycemic medication, I would tend to continue statin, in those patients.” ”So if they’re on the flip side, hypotensive, having persistent low diastolics, having symptoms of dizziness, syncopal events, recurring falls, and the blood pressure’s borderline with that, I would recommend deprescribing….”	“…there were periods that I felt he was too laxed and – what’s the word I’m looking for – he just wasn’t himself, on the ball. And I would say, ‘Could it be the medication?’ But then like I say, when they would decrease it, his activity, he’d get wilder, so we put it back to where it was – don’t increase it. I don’t want him drugged up.”
Key Theme 3: Caregivers trust and rely on clinicians to make deprescribing decisions	“…they come to trust us as the person that is there to advocate for their loved one and to honor them and care for them and all those kinds of things so I really believe they have come to trust that we’re going to do the right thing and that we do what we can do for them to provide the best quality of life and comfort in the last days of life”	“Actually we just trusted the doctors at the…VA Clinic to prescribe what they thought was appropriate, although I had input about his symptoms and saying, ‘This drug is not helping. We need something stronger. We need something different. Or this one causes too many side effects. We have to eliminate that.’ So I had some input there, but the doctors were the ones that made the decisions about what drugs he was going to take…”
Key Theme 4: Clinicians perceive caregiver involvement and buy-in as important influences on deprescribing decisions, which requires time and effort to obtain	“…we stopped something, and then the family comes in and looks at the medications again, says, ‘You know? I think they should continue that.’…if it is really important to them…We would consider treating the family [sic] in that case just to kind of keep peace in the valley.”	“…they included me, every time there was any change of medication. I call them if I had concerns. Dr. [omitted name] gave me his direct line. He really was great so any time I had any questions about how Mom was doing, thought she might benefit from something that they weren’t including, they would hear me out.”
Key Theme 5: Clinicians perceive conflicting care with other clinicians as a barrier to deprescribing	“It’s always a little uncomfortable as a primary care person to go against the recommendation from the specialist. It’s presumed that they know more than you do, so you have to have a really good reason for, in essence like, not ignoring their recommendation, and that’s always very tricky.”	N/A

3.2.1. Theme #1: Clinicians and caregivers prefer to minimize drug burden

Clinicians reported wanting to reduce polypharmacy when possible, citing both safety and comfort of the patient. Tighter treatment targets for diabetes and blood pressure were seen as potentially leading to negative events such as hypoglycemia, orthostatic hypotension, dizziness, and falling. Clinicians working with patients at end-of-life perceived that potential benefits of certain medications may no longer outweigh increased risk of potential adverse drug reactions associated with polypharmacy. In addition, clinicians said many patients preferred to take fewer medications and reported that reducing medication burden would increase patients’ quality of life. Many family caregivers also reported prioritizing patient comfort over prolonging life, and voiced a general preference to limit medications when possible, especially those causing side effects or negatively impacting comfort and/or behavior.

3.2.2. Theme #2: Clinical factors strongly influence deprescribing of chronic disease medications, with differences in how clinicians and caregivers weigh specific factors

Both clinicians and family caregivers frequently discussed the role of various clinical factors in driving decisions about continued intense treatment of chronic diseases versus deprescribing, although there were nuanced differences in the nature of these factors and how they influenced prescribing and deprescribing preferences.

Clinicians wanted to have clear justification when deciding to deprescribe medications for their CLC residents, largely based on clinical factors strongly indicating a patient is at end-of-life (e.g., marked declines in cognitive and/or physical function, becoming nonresponsive, entering hospice care). As patients entered end-of-life care, clinicians reported a shift in goals of treatment from curative to comfort. When clinical factors indicated that death may be imminent, clinicians reported high confidence in implementing deprescribing decisions. This was particularly pertinent to statins, which they perceived as lacking benefit closer to end-of-life. Similarly, clinicians felt that CLC residents who had higher levels of functioning did not make good candidates for deprescribing, especially for CLC residents who were still at risk of events that could be prevented by continuing chronic disease medications (e.g., continuing a statin if the patient has cardiovascular risk factors, atrial fibrillation, vascular insufficiency). Some clinicians said they might consider increasing chronic disease medications if the Veteran experienced symptoms of the condition, such as headaches or dizziness in patients with hypertension. Clinicians also reported wanting to maintain or increase chronic disease medications in CLC residents who had recently suffered a potentially preventable event (e.g., hypertension medications among patients who had suffered a myocardial infarction).

Family caregivers also discussed the importance of clinical factors; however, most caregivers were less concerned about life expectancy or clinical aspects of chronic disease management and more concerned about medications impacting behavior or comfort. Many caregivers mentioned mood changes or discomfort in the patient as cues to discussing medication changes with CLC clinicians. However, there were a few caregiver participants who reported a belief that medications should be continued unless there was a compelling reason to change them, such as side effects. Relatedly, a few caregivers reported that they had spent years carefully following a certain medication regimen and being attentive to changes in BP or blood glucose, and that they felt these medications and treatment targets were crucial to the resident’s health.

3.3.3. Theme #3: Caregivers trust and rely on clinicians to make deprescribing decisions

Nearly all family caregivers reported high levels of trust in clinicians about deprescribing decisions, typically believing that CLC clinicians were working in the best interests of the patient and relying on their judgment. Most reported that clinicians were very communicative with caregivers and patients about medication changes. Caregivers reported a desire to stay informed about these medication changes, even if they were going to defer to the clinician’s recommendation anyway.

3.3.4. Theme #4: Clinicians perceive caregiver involvement and buy-in as important influences on deprescribing decisions, which require time and effort to obtain

Clinicians reported family caregiver involvement as a facilitator to deprescribing. Many clinicians described discussions with caregivers in which they perceived a need to discuss the transition to comfort as a primary treatment goal and educate them on the pros and cons of continuing versus deprescribing chronic disease medications in this context. Once they had these discussions, clinicians said that most caregivers agreed to deprescribing as part of prioritizing comfort for the patient.

Clinicians also reported some experiences with caregivers resisting decisions to reduce or discontinue chronic disease medications. Clinicians felt that caregivers sometimes lacked understanding or acceptance of new goals of care that were shifting from curative to comfort. Clinicians reported that caregivers told them they felt like deprescribing would be giving up on the patient, even in cases when death was imminent. Counseling was seen as a way to explain the rationale behind deprescribing, but still some clinicians reported a lack of understanding or continued resistance by caregivers, which ultimately impacted the decision to deprescribe.

3.3.5. Theme #5: Clinicians perceive conflicting care with other clinicians as a barrier to deprescribing

Clinicians reported conflicting opinions and/or care from other clinicians in deprescribing decision-making and implementing deprescribing. Primary care clinicians believed there is a general presumption that the specialist knows more about certain conditions, making it more difficult to go against their recommendations. Furthermore, one clinician described situations in which they had to defend their decisions to deprescribe when a CLC resident under their care was hospitalized. The care team at the hospital wanted to know why these patients were not on more medications to treat hypertension, and the clinician in this study reported having to make tough arguments which he sometimes lost. Such discussions were reported by other clinicians as well, with some reporting institutional pressure, in which they worried about being blamed for deprescribing or “not doing enough.” Family caregivers rarely reported conflicting care from different clinicians, and this did not emerge as a barrier to deprescribing for any caregivers in our study.

4. Discussion

Considering perspectives of both clinicians and family caregivers, we identified five themes regarding key factors that influence the implementation of deprescribing decisions for diabetes, hypertension, and lipid control in VA NH residents who are potentially overtreated for these conditions near the end-of-life. The results of this study have several important implications for the design of interventions to support appropriate deprescribing and reduced drug burden in NH residents who are unlikely to benefit from continued intense treatment of these chronic conditions given their life expectancy and high adverse drug reaction risk.

We found that both VA NH clinicians and family caregivers in the current study prioritized comfort and generally wanted the patient to be on as few medications as necessary, as has been suggested in previous qualitative studies [39,40]; however, when considering decisions to reduce medications for diabetes, hypertension, and lipid control specifically, clinicians and caregivers differed in how they weighed clinical factors. Clinicians often cited reduced benefits and increased potential for adverse drug reactions associated with using diabetes, hypertension, and lipid control medications to pursue tight disease control targets at end-of-life. This finding suggests a high level of knowledge of and agreement with geriatric palliative care deprescribing frameworks [41,42] and guidelines recommending deprescribed medication regimens for diabetes [15,16,43,44], hypertension [21], and lipid control [26,28,29] for NH residents with LLE. Clinicians also discussed greater inclination to deprescribe these medications when there were obvious clinical factors indicating imminence of death. They were also less inclined to deprescribe when patients maintained some degree of functional independence or demonstrated greater need for continued intense treatment (e.g., recent condition-related events or cardiovascular risk factors). This is consistent with previous quantitative findings [4,17,22] examining predictors of statin discontinuation and diabetes and hypertension deprescribing in NH residents with AD or a high likelihood of death within the next six months. These studies showed lower rates of deprescribing in residents lacking explicit documentation of their end-of-life status at admission, as well as those lacking other obvious clinical markers of poor prognosis (e.g., recent weight loss) and those with greater chronic condition severity. Guidelines to avoid overtreatment of diabetes, hypertension, and lipid control pertain to a much broader group of NH residents beyond those for whom death is imminent. Therefore, our findings reinforce previous assertions that clinicians may benefit from decision support interventions to assist with the identification of potentially overtreated residents who have LLE but are not already documented as such at admission [45].

Our findings that clinicians may be more hesitant to deprescribe medications for residents with LLE who have more severe diabetes and cardiovascular disease is also consistent with prior findings. Earlier studies found that clinicians report challenges in applying evidence about benefits and risks of preventive medications in these complex patients with multimorbidity [46]. It is essential that we continue to build the evidence base to inform stronger deprescribing guidelines that incorporate tailored, evidence-based recommendations for subgroups of patients, such as for multimorbid NH residents with varying levels of underlying disease severity. With more robust evidence and specific recommendations regarding how to balance potential benefits of deprescribing with the risks of exacerbating chronic conditions, clinicians would be able to make deprescribing decisions with more confidence. Family caregivers expressed that medications with obvious negative effects on patients’ mood, cognitive function, and/or comfort were more salient targets for deprescribing than the specific chronic disease medications that were the focus of this study. Some caregivers reported they saw no reason to make changes to lipid, diabetes, or hypertension agents that the patient had been taking for years without major adverse drug reactions, a perspective that has been shared by caregivers and patients in other qualitative studies [47–49]. This may suggest that some family caregivers may be less focused on the burdens or risks associated with continuing medications, which may not always align with the goals of prioritizing patient comfort at end-of-life. As clinicians in this study reported, helping caregivers see how deprescribing chronic disease medications are aligned with pursuing patient comfort may require both considerable time and skill. Accordingly, efforts may be needed to encourage greater uptake of conversation tools designed to help clinicians discuss patient-specific goals of care in relation to chronic disease deprescribing decisions. Tools such as those developed by the VA National Center for Ethics in Healthcare [50] are designed to help clinicians approach deprescribing conversations in a way that encourages patient and caregiver trust and understanding [51]. Patients and caregivers may be more receptive to these conversations when the rationale emphasizes the potential positive impacts of deprescribing, such as reduced pill burden and avoidance of side effects and overtreatment [52]. In addition, patient- and caregiver-facing interventions to promote conversations about deprescribing may also be helpful. Since varying self-efficacy and limited knowledge of medications are barriers for older adults to initiating conversations about deprescribing [53], providing patients and family caregivers with educational materials may encourage them to start these conversations with clinicians [54].

Even when clinicians are confident in their decision to deprescribe, actually implementing this decision is seen as an extra step—one that opens them up to potential blame if a negative outcome occurs. Therefore, the option to leave a medication regimen unchanged was seen by clinicians as simpler and less risky [55], even if not in the patient’s best interest. Furthermore, patients at end of life may see multiple clinicians across different settings. This situation leads to hesitancy of some clinicians to deprescribe medications initiated by specialists, a sentiment that has been echoed in other research [49,56,57]. Many clinician participants reported conflicting opinions and/or care from other clinicians, so that they had to defend their deprescribing decisions. Furthermore, some deprescribing was reversed during transitions to other healthcare settings. This emphasizes the importance of targeting clinician-focused deprescribing interventions to the entire healthcare team beyond just those practicing within the NH [58–60]. Encouraging collaboration and communication across the clinician team to develop a “united front” regarding patient goals of care may help ensure that deprescribing decisions are implemented and maintained even after NH discharge. Consistent communication from all involved clinicians also assures patients and family caregivers that these changes are in the best interests of the patient [61].

4.1. Limitations

Several limitations of this research should be noted. Interviews were conducted early during the COVID-19 pandemic, when family caregivers were limited or prohibited from in-person visits to the NH and clinicians faced numerous disruptions to normal NH care processes. This may have impacted the quality and quantity of opportunities for clinicians, family caregivers, and patients to discuss deprescribing, and introduced additional barriers to deprescribing [62]. Furthermore, the additional workload faced by many VA NH clinicians during the pandemic may have made otherwise potentially interested and eligible participants feel they did not have the time to participate in this study. As our study was limited to VA NHs, results may not generalize to non-VA NHs or other care settings. The family caregiver sample was largely female and non-Hispanic white race/ethnicity, further limiting generalizability. Caregivers were contacted several weeks after the death of the VA patient rather than contemporaneously with deprescribing decisions, and responses may have been affected by inaccurate recall and/or bereavement. Lastly, although we interviewed family caregivers of Veterans with coronary artery disease, stroke/transient ischemic attack, diabetes, and/or hypertension who recently died in the NH, we did not formally ascertain that the Veterans met criteria for diagnosis with life-limiting illness at the time of their death.

4.2. Conclusion

Findings suggest that VA NH clinicians and family caregivers both generally prioritized avoiding unnecessary medications during end-of-life care. However, additional efforts may be needed to translate these common goals of care into decisions to deprescribe chronic disease medications with limited benefits and increased risks near the end-of-life. Encouraging clinician communication and education for family caregivers may help to align deprescribing decisions with end-of-life goals of care. In addition, fostering collaboration among multiple clinicians may be essential to facilitating deprescribing in the VA NH setting.

Key Points:

We found that both Veterans Affairs (VA) Nursing Home (NH) clinicians and family caregivers in the current study prioritized comfort and generally wanted the patient to be on as few medications as necessary. However, additional efforts may be needed to translate these common goals of care into decisions to deprescribe chronic disease medications which may have reduced benefits and increased risks near the end-of-life.

Encouraging clinician communication and education for family caregivers may help to align deprescribing decisions with end-of-life goals of care.

Fostering collaboration among multiple clinicians may be essential to facilitating deprescribing in the VA nursing home setting.

Data Availability

To protect interview participant privacy, transcript data cannot be openly shared.