Examining Adult Patients’ Success with Discontinuing Long-term Benzodiazepine Use: a Qualitative Study

Abstract

Background

Little is known about patients’ experiences with benzodiazepine (BZD) discontinuation, which is thought to be challenging given the physiological and psychological dependence and accompanying potential for significant withdrawal symptoms. The marked decline in BZD prescribing over the past decade in the US Department of Veterans Affairs healthcare system presents an important opportunity to examine the experience of BZD discontinuation among long-term users.

Objective

Examine the experience of BZD discontinuation among individuals prescribed long-term BZD treatment to identify factors that contributed to successful discontinuation.

Design

Descriptive qualitative analysis of semi-structured interviews conducted between April and December of 2020.

Participants

A total of 21 Veterans who had been prescribed long-term BZD pharmacotherapy (i.e., > 120 days of exposure in a 12-month period) and had their BZD discontinued.

Approach

We conducted semi-structured interviews with Veteran participants to learn about their BZD use and the process of discontinuation, with interviews recorded and transcribed verbatim. Data were deductively and inductively coded and coded text entered into a matrix to identify factors that contributed to successful BZD discontinuation.

Key Results

The mean age of interview participants was 63.0 years (standard deviation 3.9); 94.2% were male and 76.2% were white. Of 21 participants, only 1 had resumed BZD treatment (prescribed by a non-VA clinician). Three main factors influenced success with discontinuation: (1) participants’ attitudes toward BZDs (e.g., risks of long-term use, perceived lack of efficacy, potential for dependence); (2) limited withdrawal symptoms; and (3) effective alternatives, either from their clinician (e.g., medication, psychotherapy) or identified by participants.

Conclusions

BZD discontinuation after long-term use is relatively well tolerated, and participants appreciated reducing their medication exposure, particularly to one associated with physical dependence. These findings may help reduce both patient and clinician anxiety related to BZD discontinuation.

Supplementary Information:

The online version contains supplementary material available at 10.1007/s11606-023-08385-z.

INTRODUCTION

While safety risks associated with benzodiazepines (BZDs), particularly among older adults, have been recognized for decades, prescribing in the USA has remained relatively high. In a nationally representative survey, 12.3% of adults 65 and older reported past-year use of a prescribed BZD.¹ While use among older adults has not increased significantly, it has among middle-aged adults; those 50–64 now use at a rate equivalent to those 65 and older.² In addition, although incident (i.e., new) BZD prescribing among all ages has not increased, Kaufmann et al. found that the proportion of those newly prescribed a BZD who transition to long-term use has grown.³ Ultimately, despite safety concerns, long-term BZD use at all ages remains common.

For patients and clinicians, the prospect of stopping a long-term BZD is concerning given the physiological and psychological dependence that many individuals develop,⁴ meaning the challenges inherent to discontinuing may be heightened for BZDs.^{5, 6} In prior work exploring the attitudes of older adults and their clinicians towards potential BZD discontinuation, one patient said, “I see no reason [to] put myself through hell,” while a clinician said: “do I want to deal with [BZD tapering] that I know is going to be incredibly time-consuming and difficult, and difficult to deal with this patient?”^{7, 8} While the prospect of discontinuation leads to significant anxiety for patients and their clinicians, few studies have examined the process of discontinuation. The limited qualitative literature available describes a harrowing experience,^{9, 10} though this is at odds with other quantitative studies, which do not suggest decompensation.^11–14

Given these challenges, the experience of patients in the US Department of Veterans Affairs (VA) healthcare system—where BZD prescribing was reduced by nearly half from 2013 to 2017—may be informative.¹⁵ This decline in BZD prescribing occurred during two overlapping policy initiatives within the VA system: the Psychotropic Drug Safety Initiative (PDSI), which included a focus on BZD prescribing to older Veterans, and the Opioid Safety Initiative,¹⁶ which included an emphasis on reducing co-prescribing of BZDs with opioids. As part of a larger explanatory mixed-methods study focused on exploring variation in facility-level change in BZD prescribing across the VA,¹⁷ this qualitative study focuses on the experiences of long-term BZD users who discontinued treatment in order to improve our understanding of factors that contributed to successful discontinuation.

METHODS

Study Design and Sample Selection

Using a qualitative descriptive design,^{18, 19} we created comprehensive summaries of factors that contribute to successful BZD discontinuation using the original wording of Veterans’ perceptions and experiences of discontinuing long-term BZD treatment. The study period was October 2015 through June 2017, during which the VA conducted Phase 2 of PDSI, which focused on Veterans ≥ 75 years of age. We selected this time frame for the parent study to capture facility strategies that may account for differences in facility performance in reducing BZD prescribing to older Veterans. In addition, given significant reductions in BZD usage during this time, it presented an opportunity to identify individuals who successfully stopped their BZDs, examine their experiences, and identify factors that contributed to their success. Although this entailed asking participants about experiences that occurred at least three years prior to the interview, studies demonstrate that autobiographical information may be accurately recalled,^{20, 21} and that identifying key life transitions—in this instance, a salient and unusual clinical utilization event that occurred only once (i.e., stopping a medication valued by patients for alleviating significant discomfort)—may enhance memory of past events, although there may be retrospective self-enhancement effects.^22–26

Study participants were identified through a multi-step process. First, as part of the parent study,¹⁷ we used the VA Corporate Data Warehouse to identify all long-term BZD users at the start of October 2015, when PDSI Phase 2 began. We defined “long-term” as those with > 120 days’ supply in the prior 365 days.²⁷ To identify active long-term users (i.e., exclude those who may have already discontinued treatment), we further limited the cohort to those with a BZD prescription fill in the 90 days before October 2015. We excluded patients in an inpatient or long-term care setting during this time, or those who received hospice care. For the parent study, we used facility-level change in BZD prescribing to Veterans ≥ 75 years of age during the study period to identify high- and low-performing facilities, stratified by size (small, medium, and large based on the number of older Veterans receiving care) and geography (US Census region).

To explore BZD discontinuation efforts in more detail, we then conducted semi-structured interviews with clinicians and patients from a subset of high-performing facilities (n = 7). (For context, patients on long-term BZD treatment at the high-performing facilities experienced an average reduction of approximately 0.5-mg lorazepam per year.) Across these facilities, we first identified patients who, from October 2015 to June 2017, appeared to have their BZD prescription discontinued, which we operationalized as a long-term BZD user who exceeded 120 days without exposure to a BZD and who did not receive another prescription through November 2019, when the cohort was derived. Because we learned that most facilities did not limit their BZD-focused strategies to only those ≥ 75, for this phase of the study, we expanded the eligible age range to veterans ≥ 55. We excluded patients with dementia and those who were no longer alive. At each of the seven facilities, we identified possible discontinuers based on prescription fills. To identify individuals for interviews, we grouped potential discontinuers into three age groups (55–64, 65–74, 75 +); if more than 30 discontinuers were identified within a given age group, we randomly sampled 30.

For the n = 594 potential discontinuers identified at the seven facilities as outlined above, study staff then performed chart reviews to confirm that they had been prescribed a BZD and that their BZD had been discontinued under the supervision of their VA clinician. Our goal was to complete three interviews at each facility; chart reviews at each facility were completed on a rolling basis until the desired number of willing participants was reached, with patients recruited as described below.

Recruitment and Data Collection

Patients who met eligibility criteria were sent a letter inviting them to participate in a study interview, which was followed by up to three recruitment calls; 285 patients received at least one call. Our initial target was to recruit three Veterans per facility at seven facilities (n = 21). Our final sample included three individuals from five facilities, and one facility each with n = 2 and n = 4. When we reached our target 21 interviews, we had reached data saturation;²⁸ in other words, interviews provided information redundant to data already collected.

From April to December 2020, an experienced qualitative methodologist (LT) conducted telephone interviews, and supervised interviews conducted by a research assistant. The interviews lasted 30–60 min and were conducted using a semi-structured interview guide developed by the study team (Appendix 1). Participants were asked about the following: their BZD use; the decision to discontinue their BZD; the process of discontinuation; their experience of BZD discontinuation; and their relationship with their tapering clinician. Interview questions were formulated to be clear and precise to avoid confusion, and included prompts to probe for information, such as the names of the prescribed BZD and the providers involved, to facilitate memory of those events. Interviewers also repeated the time frame of the study to prompt participants to discuss events during that time. Interviews were audio-recorded, transcribed verbatim, and de-identified.

Participants were given a $25 gift card for participating. This protocol was approved by the VA Ann Arbor Institutional Review Board with a waiver of written consent.

Data Analysis

The analysis team (LT, DM, and SK) developed deductive codes from the interview guide, and reviewed transcripts to develop inductively derived codes. The team then independently coded the same transcript, and discussed as a group to identify and correct any misinterpretation of codes. The team repeated this process with subsequent transcripts until codes were applied similarly. Based on these discussions, LT produced a codebook with code definitions and multiple examples illustrating each code. LT then trained two research assistants to code the remaining data using the codebook, and reviewed all coded transcripts for consistency and fidelity to the codebook. Data from reconciled transcripts were inputted into NVivo 12, a qualitative analysis software package, by two research assistants. LT created detailed summaries of each code based on NVivo code reports and entered salient data from the summaries into a matrix. For this analysis, we focused on factors that appeared to influence patient success with stopping, including relationship with clinician, use of alternative treatments, and presence of side effects or symptoms. The results were discussed with the broader research team to further develop findings.

We obtained additional participant clinical data from the VA Corporate Data Warehouse from the 12 months prior to the start of the study period (i.e., October 2014 to September 2015) for descriptive purposes. We determined the presence of select diagnoses that might be in indication for BZD use, including insomnia and conditions for which BZD use among older adults might not be considered potentially inappropriate based on the American Geriatrics Society Beers Criteria.²⁹

RESULTS

The mean age of interview participants was 63.0 years (standard deviation 3.9); 94.2% were male and 76.2% were white (Table 1). Among this group with long-term BZD use, the vast majority did not have a baseline diagnosis that would be an indication for a BZD prescription (e.g., 9.5% had a diagnosis of generalized anxiety disorder and 9.5% had a diagnosis of insomnia). By self-report, the most common indications were anxiety (33.3%), PTSD (28.6%), and insomnia (28.6%). Most (81.0%) of the sample had met the criteria for long-term use (i.e., > 120 days prescribed in a calendar year) for at least five years prior to their discontinuation. Since BZD discontinuation, none of the participants had been restarted by a VA clinician; one sought treatment from a community clinician.

Table 1.

Characteristics of Veterans Who Discontinued Long-term Benzodiazepine Use (N = 21)

	N (%) or mean (SD)
Sociodemographic characteristics
Age, y	63.0 (3.9)
55–64	14 (66.7)
65–74	7 (33.3)
Sex
Male	20 (94.2)
Female	1 (4.8)
Race
White	16 (76.2)
Black	3 (14.3)
Other	2 (9.5)
Missing
Rurality
Rural	4 (19.1)
Urban	17 (81.0)
Baseline clinical diagnosesa
BZD withdrawal	0
Generalized anxiety disorder	2 (9.5)
Insomnia	2 (9.5)
REM sleep behavior disorder	0
Seizure disorder	0
Self-reported reason for BZD prescriptionb
Anxiety	7 (33.3)
PTSD	6 (28.6)
Insomnia	6 (28.6)
Muscle spasm	2 (9.5)
Unsure	2 (9.5)
Restless legs	1 (4.8)

BZD, benzodiazepine; REM, rapid eye movement

^aDerived from clinical encounter diagnoses recorded in electronic health record data

^bReported by participating during interview. Percentages sum to over 100% because some reported multiple reasons for their prescribed BZD

Although we anticipated the quality of participants’ relationships with their BZD-prescribing clinician might be important to the perceived success of the BZD tapering process, this did not appear as a major factor based on our data. During the interviews, participants only mentioned their relationship with their clinician when asked explicitly. Once asked, most were happy with their tapering clinician; if they voiced frustration about discontinuation, this was directed towards the VA system rather than their individual clinician. Even some of those who described their relationship with their tapering clinician in negative terms were satisfied with the BZD discontinuation.

Ultimately, we identified three factors that influenced success with the discontinuation process: (1) participant attitude toward BZDs, (2) extent of withdrawal symptoms, and (3) the perceived availability of effective alternatives, either offered by their clinician (e.g., medication, psychotherapy) or identified by the participants. These factors are discussed in more detail below. Please see Table 2 for a summary of these domains with additional representative quotes.

Table 2.

Factors Influencing Benzodiazepine Taper Success

Factor	Description	Representative quotes
Attitude toward BZD	Risks of use • Side effects; e.g., dizziness, feeling sedated, and falling • Possible links to cognitive decline and organ damage	“I noticed that I was getting more wobbly and I ended up having accidents, falling down off of bicycles…and my kids were talking to me about I'm acting strange.” (014-F)
	Perceived lack of efficacy • Concerns that BZDs:   • Address symptoms only   • Mask underlying conditions	“[the BZD] helped some, but it didn’t get rid of the problem, so that’s why I eventually changed [medication].” (008-C)
	Potential for addiction or dependence • Fears of BZD addiction • Preference for taking fewer medications	“If it’s addictive, I don’t want it.” (016-G) “Oh, it’s always better to not be on a medication as far as I’m concerned.” (015-G)
Limited withdrawal symptoms	• Some recurrence of symptoms formerly alleviated by BZDs • Withdrawal symptoms did not induce patients to resume BZD use	“I was willing to go along with [taper] because I knew I wasn’t having any physical reaction…if I had been suffering withdrawal I might’ve asked [provider] to give me something for that, but I really didn’t.” (002-B) “[Taper] was rough…No, it wasn’t withdrawals, it was rough because I was going back in the panic attack, well it’s not panic attacks because that’s different, anxiety attacks.” (017-A)
Effective alternatives	• Patients received substitutes from providers; e.g., medications, psychotherapy, coaching in coping skills, or alternative treatments such as acupuncture • Patients identified other substitutes, which included spiritual practices, cannabis, and alcohol	“I still get aggravated, I mean you know but I don’t think, ‘Oh my God I need a lorazepam’ anymore, I think, ‘Oh, what can I do? Take a walk, talk with people, music therapy,’ I have these tools now and I didn't have them before.” (014-F) “I think [stopping] went well because today, today I don’t need to grab a valium or take a valium whenever it rains or whenever the sun don’t shine…I think [stopping] was a very wise decision and…I’m really, really content with myself now…That’s given me a lot of accomplishment.” (012-F)

Attitude Toward Benzodiazepines

Participants’ attitudes toward BZD discontinuation were shaped by perceptions in three areas: risks of use, lack of effectiveness, and the potential for addiction or dependence.

Risks of Use

Participants reported learning about risks of taking BZDs from visits with health care clinicians, educational materials sent by the VA, or reading their medication prescription information. Experiencing side effects also led participants to believe BZDs were potentially dangerous. For example, a participant said he was experiencing falls and his family told him he was “acting strange.” He then attributed these experiences to his BZD when he received a mailing from his facility about the dangers of long-term BZD use. He told his clinician, “You know what? I don’t want your lorazepam, I don’t want your opiates, I just, let me off,” adding that “it was traumatic to be on that” (study ID 014, site F (014-F)). Two others described side effects of feeling “sedated-like” (008-C) and “a hazy-dazy kind of feeling” (010-D), while others said they believed their BZD affected their “mental stability.”

While some participants initiated their BZD discontinuation, they more commonly agreed to it—at times reluctantly—after a conversation with their clinician. For example, “I wanted to stay on the same medication that was working fine all this time but [clinician’s] suggestion was medically, you know, induced, so I had to follow his instructions” (001-B). Others said they followed their clinicians’ expertise, perceiving their clinicians as working to improve their health. A subset of participants was not convinced by these discussions related to BZD risks. They believed variously that: their clinician was ordered by the VA to stop prescribing BZDs; broad risk assessments did not apply to their individual case; or, because they were taking their BZD intermittently as needed or at low doses, the health risks did not apply to them.

Perceived Lack of Efficacy for the Prescribing Indication

Some participants were convinced to taper because they believed a BZD could not solve the underlying issue for which they were prescribed. One stated that “[clonazepam] helped some, but it didn’t get rid of the problem, so that’s why I eventually changed” (008-C). Another came under the care of a new physician who discussed with him “what the medication can do or it can’t do.” He expressed preference for the new physician’s approach, stating that “my new doctor seemed to take more in what we could do to solve issues rather than put you on some medication and hope everything was okay” (003-B).

Potential for Addiction or Dependence

Two participants self-identified as “addicts.” As such, they considered their BZD use to be part of addiction-related dysfunction in their lives, and therefore perceived stopping the BZD as improving their lives. One was further motivated by wanting a better job to “improve [his] stock in life,” and the job required that he stop BZDs. During the discontinuation process, when he experienced withdrawal symptoms, he came to interpret BZD use as another addiction (005-A). The other participant who reported a history of addiction cited his relationship to his family as the impetus for his discontinuation. At his family’s urging, he “confessed” to his PCP that “when the diazepam stopped working, I started using street drugs and I became addicted… and I was seeking treatment, seeking help” (012-F).

Other participants also expressed concerns about addiction or dependence, which led them to support discontinuation. Even though one participant still experienced some anxiety after stopping his BZD, because he perceived the BZD as “an addictive type of medication,” he still felt “it was a good idea, you know, to cut down” (009-C). A number of participants felt that reducing medication dependency overall was a priority: for example, one participant noted “anytime you can reduce…the amount of pharmaceuticals you take…you’re better off for it” (002-B). Another said, “Oh, it’s always better to not be on a medication as far as I’m concerned” (015-G).

Some participants who believed that one can develop BZD addiction sought to distinguish themselves from individuals to whom this might happen. For example, one participant said, “I’m not one of those people, I don’t show up in ERs as you know a frequent flyer” (006-A), while another said, “I realize there was people out there partying on their medication and they could wind up dead…but I didn’t feel that was me” (013-F). Another stated that he initiated the discontinuation conversation with his clinician because “[he] read it was the best thing to do…the addictive nature of the medication…by itself makes you feel somewhat nervous” (009-C).

Limited Withdrawal Symptoms

Just two participants reported experiencing symptoms that they attributed to BZD withdrawal; the vast majority did not recall having experienced withdrawal symptoms. One said “I was willing to go along with it because I knew I wasn't having any physical reaction…If I had been suffering withdrawal, I might've asked [the clinician] to give me something for that” (002-B). Others, while denying withdrawal, did report some symptoms that they perceived to be the return of symptoms that their BZD had been controlling. One participant observed, “I didn’t have a lot of real bad side effects. As far as the anxiety piece, was a little more so than what it was [while taking a BZD]” (008-C). Another participant experienced increased pain from muscle spasms (i.e., the prescribing indication for the BZD), but because she did not have withdrawal symptoms, she did not contact her clinician:

…it was explained, if I had any problems with any kind of withdrawal symptoms, any problems, just to give them a call, and I had no problem. Was I short with my husband? Yes. Do I think it was withdrawal? No, I think I was hurting (006-A).

Effective Alternatives

Some participants reported receiving BZD alternatives from their clinician either during or after their BZD taper was completed. For some, these were alternative medications; for others, this took the form of psychotherapy, acupuncture, and coaching in coping skills. One participant expressed ambivalence about his medication alternative(s): He reported receiving four medications to replace one BZD, describing the VA as “a pill-pushing organization” (004-A).

Participants also sought and initiated their own alternatives. Some said that attending church regularly and other spiritual practices provided support during discontinuation. Several used cannabis to help with pain management or to alleviate anxiety before sleep. According to one participant, “the marijuana was probably the closest to a substitution for anything that might’ve worked” (002-B). These participants said that they used cannabis only occasionally to help them sleep, such as before an important event the next day or when something disturbing happened that day. Some specified that they used cannabidiol rather than tetrahydrocannabinol-containing cannabis products. Other participants preferred alcohol to cannabis, one of whom mentioned that he took care to avoid drinking alcohol and taking his painkiller simultaneously.

For the participants who considered themselves to be “addicts,” they did not want a prescribed BZD alternative. The first contrasted his life prior to stopping as “literally floundering…I had no sense of job, no sense of purpose, no sense of where I was going in life, no close personal relationships.” After stopping, he acquired a better job and described himself as “somewhat stable” (005-A). The second expressed pride in no longer depending on his BZD, and that he is “really, really content with myself now” (012-F).

The participants who did not perceive that they had received a satisfactory alternative were more likely to be dissatisfied with their BZD discontinuation. While one participant received a satisfactory substitute medication for the BZD she was prescribed for sleep, she did not find the substitute for the diazepam she received for muscle spasms to be effective. She said, “it’s got to be four years now and I’m still ticked off about the diazepam, how’s that?” (006-A). These two participants believed they should have been allowed to continue using their BZD, as they did not take the medication in a way that they felt created safety risks. One participant who was not happy with his substitute medications felt the VA’s initiative to reduce BZD prescribing was not motivated by concern for patient welfare:

Seemed like all they done was just jumped on the bandwagon of you know, with not prescribing all that stuff, and using us for an example and it didn’t help us none (007-C).

Some participants asked their tapering provider to resume their BZD, but were refused. None in our sample attempted to switch to other providers within the VA willing to prescribe BZDs. One participant sought a BZD prescription outside the VA. No other participants sought BZDs from non-VA providers. Reasons given included having limited access to non-VA providers, concern about PDMP monitoring/random drug testing at their VA facility, and ethical objections to “going behind [tapering provider’s] back” (002-B). Participants also expressed discomfort at perceived illicit means of obtaining BZDs such as “scor[ing] drugs on the street” (018-D) or using a family member’s medication.

Another participant did not feel his current medications addressed his anxiety. He reported that he was not offered alternatives (e.g., counseling), and continued to suffer from symptoms: “I just know I stayed stressed out a lot since I’m not on anything like [clonazepam]…I don’t understand it, it helps you a lot and then they take it away from you” (011-E). A third participant, while dissatisfied with discontinuing his clonazepam, refused a substitute medication out of fear of experiencing negative side effects again: “let me make this clear…[after having] what I call negative effects from the other meds they had me on, I hesitate to try any of the others.” He also refused individual and group therapy, citing high clinician turnover and privacy concerns (010-D).

DISCUSSION

This study helps characterize the experience of adult patients in the Veterans Health Administration prescribed long-term BZD therapy who then had their BZD discontinued. The growth in BZD use that is long-term³⁰ means that efforts to limit BZD prescribing cannot only focus on limiting new prescriptions, but requires discontinuing therapy among prevalent users for whom use may no longer be appropriate. While there is significant interest in limiting BZD prescribing,³¹ little is known about the actual experience of discontinuation from the patient’s perspective.

There are several notable aspects of our findings. First, BZD discontinuation among these Veterans was relatively well tolerated. Only two of the 21 patients interviewed made mention of withdrawal symptoms. This is in stark contrast to other qualitative studies describing the experience of withdrawal, where “coming off even a little is HELL”¹⁰ or “metaphors about being trapped, stuck, or imprisoned are used to describe withdrawal.”⁹ It is possible that the people who struggled the most with withdrawal resumed their BZD therapy, and were therefore excluded from consideration for our sample. We cannot conclude that the discontinuation experience would be well tolerated by all individuals who are prescribed BZD; on the other hand, at least for some of those who have been prescribed long-term therapy, the process of discontinuation is possible and relatively well tolerated.

In contrast to other qualitative studies of discontinuation, our findings are more consistent with quantitative studies. For example, in a cluster randomized controlled trial of long-term BZD discontinuation in primary care, Vicens et al. found no increase in depression or anxiety, sleep dissatisfaction, or alcohol consumption among discontinuers.³² Similarly, Vikander et al. demonstrated that, among a group of long-term BZD users who discontinued, the overall burden of psychopathology decreased over the year following discontinuation.¹¹

Second, among the 21 individuals we identified as discontinuers by June 2017, just one participant had resumed BZD treatment (from a community clinician) by the time of our interviews in 2020. The appropriate amount of BZD prescribing almost certainly is not no prescribing; for some individuals, treatment with a BZD is appropriate, and thus continuation through non-VA means is not entirely unexpected.³³ Fortunately, from our prior work examining declines in BZD prescribing among older Veterans, it does not appear that VA patients are shifting BZD prescriptions to Medicare-reimbursed sources in substantial numbers.³⁴ It is reassuring that just one individual sought a prescription from a community source, yet any amount of dual-system medication use is potentially concerning given the associated harms for Veterans.^{35, 36} This lack of resumption of BZD therapy may also be consistent with our finding that relatively few study participants had a clinical diagnosis of common indications for BZD therapy. This is consistent with prior work suggesting that BZD prescribing is often done absent an appropriate indication.³⁷ It may be that the prescription was started during a period of transient distress that, years later, had resolved.

The primary limitation of this work is that our interviews were with a sample of patients who did, with one exception, successfully complete BZD discontinuation. Therefore, it does not reflect the experience of those who may have attempted discontinuation and did not succeed. In addition, we were asking patients to describe a process that they experienced at least three years prior; they may have provided different responses in interviews during or immediately after the discontinuation process. Our population was primarily male and comprised adults aged 55–74, so may generalize less well to groups that are more female or other ages. Finally, non-VA clinicians may have fewer nonpharmacological alternatives to which they can refer patients, such as psychotherapy.

The field of deprescribing—“the systematic process of identifying and discontinuing drugs in instances in which existing or potential harms outweigh existing or potential benefits”³⁸—has grown rapidly since this work was proposed and conducted, but we believe our findings may inform future work. The authors of a recent commentary on the nexus between deprescribing and implementation science suggest a framework such as COM-B³⁹ could be used to consider the capabilities, opportunity, and motivation at play.⁴⁰ Our findings relate to clinicians’ perceived self-efficacy (i.e., capabilities) by demonstrating that BZD discontinuation is, in fact, possible. Participants’ generally successful experiences also support patient motivation—again, by presenting successful examples of discontinuation—which is frequently unaddressed in deprescribing interventions. On the other hand, BZD discontinuation approaches may generalize less well to other types of deprescribing, as this was a top-down initiative intended to stop potentially harmful care, rather than a scenario of shared decision-making to address low-value care.

In conclusion, these interview participants shared useful information that may help inform BZD discontinuation initiatives. First, individuals who have been prescribed long-term therapy are still receptive to education related to potential harms, suggesting the potential of patient-facing educational interventions to promote discontinuation.^{14, 41} Second, these successful discontinuers recalled few symptoms of withdrawal. While this was not a study to compare taper pace with discontinuation rates, it may still suggest clinicians proceed with tapers slowly to minimize the potential for physiological withdrawal. Finally, working with patients to identify potential alternatives is also important, but this does not necessarily have to mean an alternative prescription medication. Ultimately, these results should be reassuring to both clinicians and patients alike and potentially help alleviate the significant amount of anxiety related to BZD discontinuation.

Supplementary Information

Below is the link to the electronic supplementary material.

Funding

This work was supported by Merit Award Number I01HX002340 from the U.S. Department of Veterans Affairs Health Services R&D (HSRD) Service. Dr. Krein is also supported by an HSRD research career scientist award (11–222).

Data Availability

The data underlying this study are not publicly available due to privacy concerns, but a limited dataset may be available from the corresponding author pursuant to a Data Use Agreement.

Declarations:

Conflict of Interest:

The authors declare that they do not have a conflict of interest.