Benzodiazepine usage and patient preference for alternative therapies: A descriptive study

Abstract Background and aims The prevalence of chronic benzodiazepine use in primary care settings remains high despite clear evidence of adverse health outcomes resulting from long‐term use and the availability of effective alternative behavioural therapies. Eliciting factors influencing past or current usage experience of benzodiazepine users and their future behavioural intention regarding discontinuation or alternative behavioural therapy adoption could be useful in developing informed strategies facilitating successful benzodiazepine withdrawal in long‐term users. The aim of this study was to identify patient factors influencing their current long‐term benzodiazepine use, past withdrawal attempt, and future intention to trial safer alternative behavioural therapies. Additionally, the study also aimed to explore patients' preference for information sources on behavioural therapies. Methods Point of purchase surveys were conducted with patients obtaining benzodiazepines from selected pharmacies across New South Wales (NSW), Australia. Survey items included the Beliefs about Medicines Questionnaire (BMQ‐specific), questions about patient's sociodemographic characteristics, as well as their views about long‐term benzodiazepine use and behavioural therapies. Results Seventy‐five patients were recruited from 12 pharmacies across New South Wales (NSW). The surveys were conducted from November 2016 to July 2017. The mean (±SD) age of the participants was 54.3 (±16.7) with a range of 23 to 86 years, and 67% of the participants had been using the benzodiazepine for at least 1 year. Lower‐education levels, stronger beliefs about the necessity of use, and lower concerns about ongoing benzodiazepine use were significantly associated with prolonged use. Sixty‐four percent of the participants were not interested in behavioural therapies, and there was a significant relationship between the participants' future preference for behavioural therapies and their concerns about the potential adverse effects of benzodiazepines. A majority of the participants rated general practitioners (GPs) as their first choice and pharmacists as the second choice for discussing behavioural therapies. Conclusions Specific individual sociodemographic characteristics of benzodiazepine users and their medication‐related beliefs influence their current benzodiazepine usage and future intention to trial behavioural therapies as an alternative to their benzodiazepines. Based on the reported preferences of benzodiazepine users in this study, developing and evaluating GP‐pharmacist collaborative services to improve the uptake of behavioural therapies as an alternative to benzodiazepines can be recommended.

usage and future intention to trial behavioural therapies as an alternative to their benzodiazepines. Based on the reported preferences of benzodiazepine users in this study, developing and evaluating GP-pharmacist collaborative services to improve the uptake of behavioural therapies as an alternative to benzodiazepines can be recommended. KEYWORDS behavioural therapies, beliefs, benzodiazepine, chronic use, patient factor, primary care, withdraw Benzodiazepines are widely prescribed for their hypnotic, 1 anxiolytic, 2 muscle relaxant, and antiepileptic indications. 3 Despite their common use, concerns remain over their long-term safety.
Pharmacoepidemiological data indicate that prolonged use of benzodiazepines is associated with an increased risk of falling, 4,5 therapeutic dose dependence, 6 and an increased incidence of dementia. 7 Benzodiazepine exposure is also associated with an increased risk of physical disability. 8 Recent cohort studies suggest an increased risk of exacerbations in benzodiazepine users with asthma and a higher likelihood of episodes of pneumonia and related mortality. 9,10 A recent systemic review also highlighted an overall increased risk of all-cause mortality in benzodiazepine users. 11 Inappropriate use of benzodiazepines in Australia has been recently reported to lead to high costs of managing adverse events in residential aged care facilities. 12 In light of the well-established side effect profile of benzodiazepines, prescribing guidelines do not recommend their prolonged use.
The Royal Australian College of General Practitioners (RACGP) recommends that benzodiazepines should not be prescribed for longer than 4 weeks. 13 For patients requiring ongoing treatment, behavioural therapies such as cognitive behavioural therapy offer a promising alternative to the benzodiazepine for several conditions. For example, mounting evidence suggests that behavioural treatments produce comparable efficacy with benzodiazepines and have a longer-lasting effect in patients with insomnia. [14][15][16] Similarly, research evidence highlights better or equivalent efficacy profiles for behavioural therapies over benzodiazepines for managing anxiety disorder, panic disorder, and dysthymia. 17 The benefits of de-prescribing benzodiazepines have been highlighted in several studies. For example, in elderly nursing home residents, benzodiazepine withdrawal significantly improved memory and cognitive functioning compared with those who continued to take benzodiazepines; withdrawal did not give rise to anxiety, agitation, or sleeplessness. 18 Health care utilization and hospital admission costs resulting from traffic accidents and falls attributable to benzodiazepine use can be reduced through successful discontinuation 19 ; thus, benzodiazepine discontinuation can also have an economic benefit.
Despite the prescribing guidelines, the listed adverse effects of prolonged benzodiazepine use and robust evidence supporting behavioural therapies over benzodiazepines, the long-term use of benzodiazepines remains high. While the pattern of benzodiazepine prescribing varies globally, prolonged use of benzodiazepines is a com- Although there has been a modest decline in the volume of benzodiazepine prescriptions dispensed annually, overall, there is still a high level of long-term benzodiazepine use in Australia. [23][24][25] For example, in the case of insomnia, almost 80% of patient presentations result in a prescription for benzodiazepine derivatives (such as temazepam, oxazepam, diazepam, and nitrazepam), and this figure has remained stable over the past 10 years. 26,27 Both patient-and prescriber-related factors appear to be linked with prolonged benzodiazepine use. Prescribers' perceptions and attitudes have been reported to be a key factor leading to prolonged benzodiazepine use. 28 Recently, an Australian study highlighted that physicians often believe that patients taking benzodiazepines are unlikely to be willing to withdraw their medication and, thus, renew prescriptions without offering discontinuation or withdrawal plans. 29 Individual patient factors can also affect the length of benzodiazepine use; data indicate that being older, lonely, less educated, as well as having a poorer mental health profile and lower perceived general health status are associated with prolonged benzodiazepine use. 30,31 Other patient-related factors that can influence ongoing use of benzodiazepines include beliefs and attitudes of the users towards their medication. 32 This has been proposed based on the well-known Health Belief Model (HBM), which suggests that patients' beliefs about their health issues, perceived benefits of and barriers to action, and self-efficacy explain engagement in health-promoting behaviours while the actual action is triggered by cues. 33  around the necessity of using benzodiazepines and a lower level of concern about long-term use were associated with resistance to benzodiazepine discontinuation. 35 Similarly, another qualitative study reported that individuals with higher concerns about the risk of taking benzodiazepines were more likely to attempt ceasing the medication of their own volition and were often more interested in trying behavioural interventions. 36  behaviours require a prior intention, and past behaviour may contribute to this future intention. 37 In most research reporting trials of benzodiazepine weaning off, the patients' willingness to withdraw benzodiazepines and intention or readiness to adopt alternative therapies are assumed but not explored. [38][39][40] It is not surprising that in many of these reported trials, benzodiazepine discontinuation rates remain less than 65%, with low-response and high-dropout rates. [38][39][40] Patients' willingness to stop the medication and try behavioural therapies might be a key factor that affects successful discontinuation. BMQ-specific was also used to assess participants' personal views about their benzodiazepines ( Figure 1). 34 Figure 2 outlines the diagrammatic representation and flow of questions used in the survey.

This study was approved by The University of Sydney Human
Research Ethics Committee (HREC: 2014/1020).

| Participants
Pharmacies throughout the Australian state of New South Wales

| Analysis
Completed survey questionnaires were assigned a code before initiat-   Table 1.

| Past benzodiazepine use behaviours (withdrawal attempts)
In the sample, 47% (n = 35) of the participants had attempted to withdraw the benzodiazepine, and in these cases, sudden cessation of the benzodiazepine was the most commonly reported strategy applied to withdraw the medication. The reported reasons for attempting benzodiazepine withdrawal included concerns about health (49%), fear of  Using benzodiazepines for at least 1 year 50 (67) Reason for benzodiazepine use Using benzodiazepine for sleep disorders 44 (59) Using benzodiazepine for psychiatric disorders 23 (31) Using benzodiazepine for pain 5 (7) Using benzodiazepine for other problems 3 (4) developing dependence (40%), doctors' suggestion (37%), and other reasons (17%). Details about the reasons and strategies for withdrawing the benzodiazepine are presented in Table 2.
In a binary logistic regression analysis, none of the variables evaluated (age, sex, level of education, benzodiazepines being used, length of action of the benzodiazepine, and reason for taking the benzodiazepine) were significantly associated with a benzodiazepine withdrawal attempt. There was no significant correlation between the BMQ themes and past withdrawal attempts.  Table 3.

| Current long-term benzodiazepine use
We observed no significant correlation between participants' age, sex, benzodiazepines being used, duration of action of the benzodiazepine, or their reason for taking the benzodiazepine with current longterm benzodiazepine use.

| Future benzodiazepine use behaviours
In the sample, about two thirds (n = 48) of the participants were not willing to consider behavioural therapies as a substitution for their benzodiazepines (only 27/75 participants were interested in behavioural therapies). In these cases, participants' lack of confidence about the efficacy of behavioural therapies and their lack of time to try behavioural therapies were the main two reasons for not considering behavioural therapies. The reported reasons for not considering behavioural therapies are highlighted in Table 4.
We observed no statistically significant association between past benzodiazepine use behaviours (withdrawal attempts) and the future intentions of trialling behavioural therapies (Table 3).
There was a weak (r s [75] = 0.297, p = 0.010) relationship between the BMQ concerns theme and future preference for behavioural therapies (Table 3). Interestingly, binary logistic regression revealed that participants using benzodiazepines for sleep disorders were more likely to be in the "interested in behavioural therapies"   Dependency on sleeping pill 15 (31) Participants' perception that behavioural therapies take longer time to produce effect 11 (23) Participants' perception that seeing a psychologist is costly 9 (19) Other reason 5 (10) Reasons for not preferring behavioural therapies were derived by having participants choose from multiple choice options presented to them with the questions. The dependent variable in this analysis is future preference for behavioural therapies coded as 0 = not interested in behavioural therapies and 1 = interested in behavioural therapies (target group). Abbreviations: CI, confidence interval; OR: odds ratio. their ability to be able to cease use. Therefore, based on theories such as those around planned behaviour 52,53 as well as our study results, counselling directed at current use and future intention to use is perhaps a more judicious use of clinical time when attempting deprescribing of benzodiazepines, rather than focussing on past withdrawal attempts.
The observed positive relationship of the necessity theme and the negative relationship of the concerns theme with actual long-term benzodiazepine use (current behaviour) highlights that the decision to continue benzodiazepines reflected the balance between these two opposing belief sets. However, we observed that future intentions around continued benzodiazepine use (eg, voiced preference for behavioural therapies in the future) were only related with the con- Introducing other health care professionals co-located within GP practice centres to support GPs for educating patients about behavioural therapies may help to overcome some of these issues. 29 In a qualitative Australian study, GPs acknowledged the role of other health professionals in facilitating successful benzodiazepine cessation. 44 In the current study, following GPs, pharmacists were the second preferred source of information for behavioural therapies. Thus, a collaborative approach with pharmacists could be an option where they can support GPs in withdrawing benzodiazepines and providing behavioural therapies. 78

| Strengths and limitations
The small sample size may be a limitation of this study. For moderate strength correlations (eg, 0.30-0.40), a sample of 75 may be sufficient to demonstrate significance at a 0.05 significance level with 75% power. 84 However, the study may have been underpowered to demonstrate lower-strength correlation at the same significance level and power. There is a possibility of selection bias, as individuals who have successfully withdrawn their benzodiazepines were not recruited for this study. Further, the questions used in this study to explore participants' willingness to withdraw and preferences for behavioural therapies were customised for this study. However, the questionnaire was developed by reviewing the previous literature and was facevalidated by psychology, sleep, and pharmacy practice researchers.
Given this part of the questionnaire measured a set of diverse issues (past behaviour of withdrawing, actual daily use of benzodiazepine, and future intentions to try alternatives to benzodiazepine use), internal consistency was not measured, as behaviour and intention constructs are in themselves quite different and these questions also had different response items. Past behaviours appeared less important based on our data, and current behaviours are verifiable through pharmacy prescription records. Therefore, it may be suggested that a fuller set of items with comparable response sets around intended future behaviours with respect to benzodiazepines should be constructed and tested psychometrically. This would be useful for future research and is certainly a limitation in our study. The questionnaire did not ask explicitly about participants' readiness or current intention to withdraw from the benzodiazepine. Exploring participants' belief was limited to BMQ-specific rather than using the more comprehensive exploration of belief sets, for example, using variables included in models such as the HBM. Lastly, there is a possibility for response bias (eg, strong beliefs about the necessity of benzodiazepines, low concerns, and social desirability).

| CONCLUSION
Specific characteristics of benzodiazepine users and their beliefs about taking the benzodiazepine can inform the provision of individualised interventions by GPs to help switch patients currently on long-term benzodiazepines to alternative behavioural therapies. This study highlights the significance of informing patients about the balance between the necessity of use versus the concerns that long-term use of benzodiazepines raises. Given that GPs are very time pressured, introducing practice pharmacists within general practices could be time efficient and enhance GPs' capacity for providing behavioural therapies, and this area of collaborative care in de-prescribing unwarranted use of high-risk medications warrants future research.