Pharmacist-led sedative-hypnotic deprescribing in team-based primary care practice

Abstract

Background:

Sedative-hypnotic (SH) medications are often used to treat chronic insomnia, with potentially serious long-term side effects. The objective of this study is to evaluate an interprofessional SH deprescribing program within a community team-based, primary care practice, with or without cognitive behavioural therapy for insomnia (CBT-I).

Methods:

Retrospective chart review for patients referred to the team pharmacist for SH deprescribing from February 2016 to June 2019.

Results:

A total of 121 patients were referred for SH deprescribing, with 111 (92%) patients who attempted deprescribing (average age 69, range 29-97 years) and 22 patients who also received CBT-I. Overall, 36 patients (32%) achieved complete abstinence, and another 36 patients (32%) reduced their dosage by ≥50%. For the 36 patients who achieved complete abstinence, 26 (72%) patients remained abstinent at 6 months (9 patients resumed using SH and 1 patient was lost to follow-up). The proportion of patients achieving complete abstinence or reduced dosage of ≥50% (successful tapering) was higher with CBT-I than without CBT-I but did not reach statistical significance (77% vs 62%, p = 0.22). There were also no statistically significant differences detected in the success between those who took a benzodiazepine and those who took a Z-drug (67% vs 61%, p = 0.55) or for those who took SH daily and those who took them intermittently (67% vs 44%, p = 0.09).

Conclusion:

Almost two-thirds of patients participating in our pharmacist-led program were able to stop or taper their SH medications by ≥50%. The role of CBT-I in SH deprescribing remains to be further elucidated. Can Pharm J (Ott) 2021;154:xx-xx.

Knowledge Into Practice.

• About 1 in 10 Canadians are taking sedative-hypnotics, with potential long-term side effects.

• Pharmacists in a team-based, primary care setting can help lead deprescribing efforts.

• The role of cognitive behavioural therapy for insomnia (CBT-I) in sedative-hypnotic deprescribing requires further study.

Mise En Pratique Des Connaissances.

• Environ un Canadien sur 10 prend des sédatifs hypnotiques, avec de possibles effets secondaires à long terme.

• Les pharmaciens, au sein d’une équipe de soins primaires, peuvent aider à diriger les efforts de déprescription.

• Le rôle de la thérapie cognitivo-comportementale pour l’insomnie (TCC-I) dans la déprescription de sédatifs hypnotiques exige une étude plus approfondie.

Background

According to a 2019 report from the Canadian Centre on Substance Use and Addiction, the use of prescription sedative-hypnotic medications (including benzodiazepines and “Z-drugs”) among the Canadian adult general population is about 12% and has remained relatively stable since 2013. ¹ In particular, the report pointed out that usage was highest in older adults over 65 years of age. Sedative-hypnotics are often used to treat chronic insomnia. However, these medications are indicated for short-term relief for up to 2 to 4 weeks only. While the clinical evidence to support their efficacy for chronic insomnia is not strong, there are substantial data on the potential harm of long-term use of sedative-hypnotics, especially in the elderly.^2,3 According to a Canadian clinical review (2013), sedative-hypnotic drugs are associated with nearly 5-fold and 2.6-fold increased risks of adverse cognitive and psychomotor events, respectively, as well as falls and hip fractures, and hospital admission after motor vehicle collisions. ⁴ Decreasing the inappropriate use of sedative-hypnotics for chronic insomnia (“deprescribing”) is therefore an important patient safety initiative.

Two previous deprescribing randomized trials conducted in community pharmacies, the EMPOWER and D-PRESCRIBE trials, have demonstrated that pharmacists could help community-dwelling older patients reduce reliance on sedative-hypnotic medications.^5,6 In those 2 clinical trials, educational materials, including a stepwise tapering protocol, were distributed to patients in the active arm. It was found that consumer-targeted educational intervention resulted in greater discontinuation of prescriptions for sedative-hypnotics at 6 months compared to usual care. However, the generalizability of these findings to other settings requires further research. In addition, relapse rates and concomitant medication use were not reported. The potential effect of cognitive behavioural therapy for insomnia (CBT-I) was also not investigated.

CBT-I is an evidence-based intervention supported by clinical guidelines. It involves cognitive psychoeducation and behavioural techniques like stimulus control and sleep restriction, and it is effective in helping patients improve sleep without medications.^2,3 CBT-I combined with sedative-hypnotic deprescribing could improve success. Baillargeon et al. ⁷ compared the efficacy of benzodiazepine tapering when performed on its own and when combined with CBT-I. They found that offering CBT-I in weekly small group sessions significantly improved the success rate in achieving complete abstinence from benzodiazepine use by the end of an 8-week period, and this persisted at a 12-month follow-up. However, that study had very strict inclusion and exclusion criteria, and participation rate was very low. Of the 344 potential candidates, fewer than 20% entered the study. It is therefore difficult to extrapolate their findings directly to everyday practice with less selected patient populations. A subsequent study found that CBT-I did not help reduce relapse rates at 3, 12 and 24 months after deprescribing in long-term benzodiazepine users. ⁸ Patient self-efficacy and other factors that affect adherence to sedative-hypnotic tapering schedules may also be important to consider.^9,10

In recent years, several Family Health Teams (FHTs) in Ontario have started multidisciplinary sedative-hypnotic deprescribing initiatives in response to the Choosing Wisely Campaign’s call for action. ¹¹ FHTs are team-based, interprofessional primary care organizations.

Our team, the North York FHT in Toronto, started such a program in 2016, during which patients were referred by FHT-affiliated family physicians to the team pharmacist for sedative-hypnotic deprescribing. The evidence generated by the EMPOWER and D-PRESCRIBE trials was leveraged and adapted into a pharmacist-led effort embedded in routine, team-based primary care practice. Regular follow-up, monitoring, support and dosage adjustments guided by the pharmacist were built into our process to help better tailor the deprescribing effort according to individual need and response.

CBT-I was also made available as part of the program, to be provided by either a pharmacist or a social worker, either individually or in a small group setting. Patient encounters and medication changes were documented with standardized tools within the patient’s electronic medical records (EMRs).

We recently evaluated our program. Our aim is to provide insights into the effectiveness of sedative-hypnotic deprescribing in our team-based, primary care setting and of resources required.

Methods

This patient safety improvement initiative was conducted at North York FHT. At the time of the project, the FHT was composed of 90 community-based family physicians and 40 interprofessional health care providers (including 3 pharmacists and 8 social workers) practising out of 20 clinical locations and providing care for over 85,000 patients (http://nyfht.com). We followed the SQUIRE 2.0 (the revised Standards for Quality Improvement Reporting Excellence) guidelines when reporting. ¹²

We evaluated our program through retrospective chart audits from the inception of the initiative in February 2016 until June 2019. We reviewed the EMRs of patients within the North York FHT who had been referred by the family physician to the pharmacist for sedative-hypnotic deprescribing. CBT-I was offered and provided by either the pharmacist or the social worker if the patient was having difficulty sleeping at the time of the referral. Patients worked with the pharmacist on deprescribing and were followed and monitored regularly in person or by phone. Deprescribing was guided by the stepwise tapering protocol from the EMPOWER trial and clinical guidelines and adapted as needed to individual need and response.^5,13 Data collected included patient age, gender, sedative-hypnotic use at baseline, indications, dosage, frequency (daily or intermittently as needed), relevant comorbid conditions (such as anxiety/depression, restless leg syndrome), final dosage of sedative-hypnotics at the last follow-up encounter and Insomnia Severity Index (ISI) scores. ¹⁴ We recorded the use of other sedative medications used for sleep (including over-the-counter or natural health products) as a measure of possible unintended consequences. All patient information was kept confidential and secure throughout the study through encryption, and only anonymized patient data were used in the analysis. This retrospective chart review was done for the purpose of program evaluation and was thus considered to be exempt by the North York General Hospital Research Ethics Board, Toronto.

Each of the pharmacists and social workers who provided CBT-I went through a 2-day training course. An internal audit of 52 patients was performed comparing their pre- and post-CBT-I ISI scores. Patients achieved similar average reduction of 8 points in their ISI score post-CBT-I (with lower scores indicating fewer insomnia symptoms), whether working with a pharmacist (35 patients) or a social worker (17 patients), and the improvement was consistent with previously published results. ¹⁵

Statistical analyses

We used descriptive statistics for deprescribing and dosage reduction: percentages of the patient cohort. We used Fisher’s exact test to test for differences in the proportion of patients achieving complete abstinence or dosage reduction of at least 50% in the following subgroups: 1) those who received CBT-I and those who did not, 2) those who took benzodiazepines and those who took Z-drugs and 3) those who took sedative-hypnotics on a daily basis and those who took them on an intermittent, as-needed basis only. The rate of sedative-hypnotic deprescribing was estimated by dividing the percentage of dosage reduction by the time (in weeks) required to complete the process. The sample size required to be able to detect a 25% difference in the proportion of patients achieving complete abstinence or reduced dosage of ≥50% (successful tapering) with and without CBT-I (α level = 5%, β error = 20%) was 104.

Results

Between February 2016 and June 2019, 121 patients were referred by North York FHT family physicians to the team pharmacist for sedative-hypnotic deprescribing. A total of 111 patients (92%) attempted deprescribing. The other 10 patients decided to continue with their current sedative-hypnotic use for the time being after discussion with the team pharmacist but were welcome to contact the pharmacist again should they change their mind. CBT-I was offered to 64 patients (53%) at the discretion of the clinician if the patient was having difficulty sleeping at the time of the referral, but 42 declined. Twenty-two patients (34% of those offered CBT-I) received CBT-I along with deprescribing, and 89 patients attempted deprescribing without CBT-I.

Table 1 shows the baseline characteristics of the 111 patients who attempted sedative-hypnotic deprescribing. The average age was 69.3, ranging from 29 to 97 years, with an interquartile range of 64 to 78 years.

Table 1.

Baseline characteristics of the 111 patients who attempted sedative-hypnotic deprescribing

Patient characteristics	Categories	Number (%)
Age	≤65	38 (34)
	>65	73 (66)
Sex	Female	78 (70)
	Male	33 (30)
Sedative-hypnotics (SH) (3 patients on both BZ and Z-drugs)	Benzodiazepines (BZ)	67 (60)
	Z-drugs	47 (42)
Number of concurrent SH	1	103 (93)
	2	7 (6)
	>2	1 (1)
Frequency of use	Daily	94 (85)
	Intermittent	17 (15)
Underlying comorbidities	Anxiety	39 (35)
	Depression	28 (25)
	Bipolar disorder	2 (2)
	Restless leg syndrome (RLS)	6 (5)
	Chronic pain	26 (23)
	Substance abuse (alcohol)	4 (4)
Other medications	Antihistamines	5 (5)
	Antidepressants	45 (41)
	Antipsychotics	4 (4)
	Opioids	10 (9)
	Antiseizure medications	13 (12)
	Non-SH RLS medications	4 (4)
	Cannabis (medical)	4 (4)
	Melatonin	10 (9)

For the primary outcome (Table 2), 36 patients (32%) achieved complete abstinence, and another 36 patients (32%) achieved a dosage reduction of at least 50% by the time they completed or dropped out of the program. Patients achieving complete abstinence or a dosage reduction tapered by a mean of 7.3% of their original medication dosage per week. The deprescribing process took an average of 21 weeks, with an interquartile range of 8 to 30 weeks. The proportion of patients who achieved complete abstinence or a dosage reduction of at least 50% was higher with CBT-I than without but did not reach statistical significance (77% vs 62%, p = 0.22).

Table 2.

Achieving complete abstinence or ≥50% dosage reduction in sedative-hypnotic use among patients with or without cognitive behavioural therapy for insomnia (CBT-I)

Sedative-hypnotics	Deprescribing without CBT-I (n = 89)	Deprescribing with CBT-I (n = 22)	Total (n = 111)
Complete abstinence	28	8	36 (32%)
Dosage ↓ by ≥50%	27	9	36 (32%)
Dosage ↓ by <50%	15	3	18 (16%)
No dosage reduction	19	2	21 (19%)
Complete abstinence or dosage reduced by ≥50%	55 (62%)*	17 (77%)*	72 (65%)

^*Nonsignificant difference (p = 0.46, Fisher’s exact test).

As shown in Table 3, there were also no statistically significant differences in success between those who were taking a benzodiazepine and those who were taking a Z-drug (67% vs 61%, p = 0.55), as well as between those taking sedative-hypnotics on a daily basis and those taking them on an intermittent basis only (67% vs 44%, p = 0.09).

Table 3.

Achieving complete abstinence or ≥50% dosage reduction in sedative-hypnotic use among patients taking benzodiazepines vs Z-drugs and those with daily vs intermittent nondaily use

Factors	Total number of patients	Number of patients achieving complete abstinence	Number of patients achieving dosage reduction ≥50%	Number (%) of patients achieving complete abstinence or dosage reduction ≥50%	Fisher’s exact test
Benzodiazepines	64*	20	23	43 (67)	p = 0.55 †
Z-drugs	44*	16	11	27 (61)
Daily administration	95	31	33	64 (67)	p = 0.09 †
Intermittent administration	16	5	2	7 (44)

^*Three patients who were taking both a benzodiazepine and a Z-drug at baseline were excluded.

^†Nonsignificant differences.

Among the 22 patients who underwent CBT-I along with deprescribing, 15 had their ISI scores tracked before and after CBT-I. The average change in ISI score was −5.2 ± 7.8 points, with an improvement in 11 out of the 15 patients (73%). We followed up with the 36 patients who achieved complete abstinence at 6 months. Nine patients (25%) had restarted using sedative-hypnotics, while 1 patient (3%) was lost to follow-up. The other 26 patients (72%) remained abstinent.

Table 4 shows selected additional classes of sedating medications that patients were taking before and after sedative-hypnotic deprescribing. We found little change. An antidepressant was added for 8 of the 111 patients after sedative-hypnotic deprescribing, while 7 other patients stopped taking 1. Five of 111 patients started taking melatonin and 3 other patients discontinued this drug.

Table 4.

Concomitant medication use before and after sedative-hypnotic deprescribing

Medication classes	Added after sedative-hypnotic deprescribing attempt (among 111 patients)	Removed after sedative-hypnotic deprescribing attempt (among 111 patients)	Net changes (among 111 patients)
Antihistamines	2	2	0 (0%)
Antidepressants	8	7	1 (0.9%)
Antipsychotics	1	0	1 (0.9%)
Opioids	2	4	−2 (–1.8%)
Antiseizure medications	0	3	−3 (–2.7%)
Nonbenzodiazepine medications for restless leg syndrome	1	1	0 (0%)
Cannabis	2	0	2 (1.8%)
Melatonin	5	3	2 (1.8%)
Other natural products	2	0	2 (1.8%)

A total of 89 patients who attempted sedative deprescribing without CBT-I had an average of 5.9 encounters (interquartile range [IQR]: 3 to 8) between the clinician and the patient during the process, with 1.9 encounters (IQR: 1 to 3) in person and 4 encounters (IQR: 1 to 5) by phone. Twenty-two patients who received CBT-I along with deprescribing had an average of 10.8 encounters (IQR: 5 to 14) in total, 6.3 encounters (IQR: 3-8) in person and 4.5 encounters (IQR: 1 to 6) by phone. Patients were followed either in person or by phone about once every 2 weeks on average.

Discussion

We found that a pharmacist-led, team-based sedative-hypnotic deprescribing program was associated with abstinence or a dose reduction of 50% or more in almost two-thirds of participants. Most patients referred were willing to try reducing their sedative-hypnotic use after discussing it with the pharmacist. The pace of dosage tapering was consistent with clinical guideline recommendations. ¹³ Sedative-hypnotic deprescribing did not seem to lead to an increase in the use of other sedating medications.

In contrast to the results from previous trials,^7,10 abstinence and dosage reduction rates did not differ statistically significantly between those who received CBT-I and those who did not. While this is a retrospective observational study with a limited sample size, a possible explanation for the observed discrepancy was that eligibility in those trials included impaired daytime functioning or marked distress, as well as sleep difficulties. Those patients might therefore have a higher degree of interest and engagement in the process of concurrent CBT-I coupled with sedative-hypnotic deprescribing. Our program included less stringent criteria for patients referred by their family physician. CBT-I was offered at the discretion of the clinician and would only proceed if a patient agreed to participate. In fact, only about one-third (22 out of 64) of patients who were offered CBT-I agreed to it. Patients who chose to participate in CBT-I might have been more concerned about their sleep. They might therefore be more reluctant to reduce their sleep medications, contributing to the observation that CBT-I did not improve deprescribing success.

Z-drugs have been perceived as having fewer withdrawal effects upon discontinuation. ¹⁶ In our program, those on Z-drugs did not have better success compared to those on benzodiazepines. We also observed no statistically significant difference between patients taking sedative-hypnotics on a routine daily basis and those on intermittent, as-needed medications, highlighting the difficulties in stopping these drugs. The EMPOWER and D-PRESCRIBE trials have previously reported that other factors, including age greater than 80 years, sex, education level, health and frailty status, duration of sedative-hypnotic use (less than or more than 5 years), indication for use, dose, previous attempt to taper and concomitant polypharmacy (10 drugs or more per day), had no significant effect on sedative-hypnotic discontinuation rates.^5,6

Patients were referred by family physicians after a brief discussion of deprescribing. These patients may have been more ready to stop the medications than unselected patients on sedative-hypnotics. ¹⁷ We did not measure stage of change of patients at the time of referral or family physician practice patterns (awareness of the initiative and proportion of qualifying patients who were offered participation in the program), which were limitations. However, this represents real-world primary care practice with a new interprofessional program implemented in a routine primary care setting; patients were less selected than would be expected in clinical trials.

This was a retrospective chart audit rather than a randomized trial, and we had a limited number of patients from a single FHT. The findings are considered hypothesis-generating only. We do not make definitive claims of effectiveness or of generalizability to other settings.

In conclusion, this program was associated with tapering or discontinuation of sedative-hypnotic medications for over 60% of patients participating. Future research in routine primary care settings could be done to further examine the effects of sedative-hypnotic deprescribing on sleep, mood and quality of life, as well as to identify which patients may benefit more from the addition of CBT-I. In addition, ways to reach out to the nonresponders and those who restarted using sedative-hypnotics after discontinuing need to be explored further. Moreover, a more proactive approach to identify patients who are taking sedative-hypnotics for chronic insomnia, such as performing systematic searches through electronic databases, would help efforts to reach out to those who may benefit from sedative-hypnotic deprescribing. This applies to both the traditional community pharmacy practice and team-based primary care settings. Further studies would help provide more comprehensive evidence on effectiveness of methods to improve the care of patients suffering from chronic insomnia and taking sedative-hypnotic medications. ■