Background
Insomnia is a common problem affecting one in twenty Australian adults [
1], and up to one in ten people aged 65 years and over [
1‐
3]. The recommended first-line treatment for insomnia management is cognitive behavioural therapy [
4‐
6], followed by short term pharmacological treatment using hypnotics if cognitive behavioural therapy is ineffective [
5]. Several studies have shown that cognitive behavioural therapy is as effective or more effective than hypnotics in terms of time to sleep onset, amount of time awake during sleep and sleep efficiency, and that the effects of hypnotics are lost following cessation of treatment [
7,
8].
Australian guidelines recommend that hypnotics, if required, should only be used short-term (usually no more than 2 weeks), intermittently and at the lowest possible dose [
4,
5]. Prolonged use of hypnotics is not recommended as it may lead to tolerance and dependence [
4]. Despite this, there is evidence that hypnotics are often used continuously for much longer periods than recommended [
9]. In 2006 to 2008, the Australian Bettering the Evaluation and Care of Health (BEACH) survey, which continuously surveyed a rolling cohort of 1000 general practitioners (GPs), reported that 95% of insomnia problems managed at a general practitioner encounter involved a prescription for hypnotics; temazepam (48%) was most commonly prescribed [
10]. A US national study from 1996 to 2001 in adults with insomnia or sleeping difficulty reported that hypnotics alone (without cognitive behavioural therapy) were prescribed during half of all patient visits [
11].
There is limited evidence on the efficacy of hypnotics for insomnia in the older population [
12,
13]. Further, older people are more vulnerable to side effects of hypnotics, with use associated with increased risk of falls and fractures [
14‐
16], confusion and memory impairment [
14,
17,
18], motor vehicle accidents [
14,
19] and incontinence [
20]. In this population, the risks may outweigh any potential benefits [
14]. Hypnotic use in Australia remained stable from 2000 to 2007 and decreased marginally from 2007 to 2011 [
21,
22]. However, 2006 data showed that hypnotic use while relatively low and comparable for the age groups of 30 to 39, 40 to 49 and 50 to 59, increased from the age of 60 years with peak use in the 95 to 99 year age group [
22]. Similarly, the US National Ambulatory Medical Care Survey (NAMCS) data from 1996 to 2001 representing 95 million outpatient visits for insomnia showed that patients aged 65 years and older were 5 times more likely to be prescribed hypnotics compared with those aged 18 to 35 years (OR 4.87, 95% CI 3.64–6.51) [
23].
Discontinuation of hypnotics in older people is feasible [
9,
24,
25] and has been associated with cognitive and psychomotor improvement [
9]. Since most hypnotics are prescribed by general practitioners (GPs) [
22], interventions targeting GPs represent an opportunity to reduce prolonged and inappropriate hypnotic use in older people. However, prior research has shown that general practitioners often think patients expect a hypnotic to be prescribed when they visit a GP with sleep problems. [
26].
The Veterans’ Medicines Advice and Therapeutics Education Services (Veterans’ MATES) [
27] is a national quality improvement program funded by the Australian Government Department of Veterans’ Affairs (DVA) which aims to improve use of medicines and health outcomes for Australian veterans. Almost three quarters of the current veteran population is aged over 65 years (
n = 143,497, 73%) [
28], with associated multimorbidity and multiple medicine use. [
29,
30] To date, the Veterans’ MATES programme has run two interventions aiming to improve the use of hypnotics in older veterans. The first intervention was implemented in March 2009 with a follow-up intervention in June 2012, and both aimed to improve the management of insomnia and reduce hypnotic use among veterans in the primary care setting. The messages in both interventions included the need to reduce hypnotic use and to encourage the use of non-pharmacological options to manage insomnia. This study aimed to evaluate the effectiveness of the Veterans’ MATES insomnia interventions in reducing hypnotic use among older Australians, and to estimate health consequences due to changes in hypnotic use following the interventions. The health consequences included changes in hospital admissions due to falls and hospital admissions for acute confusion.
Discussion
Our evaluation showed that the Veterans’ MATES interventions using patient-specific prescriber feedback, combined with educational information to other health professionals and consumer-focused educational brochures targeting veterans were effective in reducing hypnotic use among Australian veterans. At the start of the study in March 2007, 15% of older Australians were using hypnotics prior to the Veterans’ MATES interventions. Use reduced to 12% of the veteran population in May 2013 following both interventions. While the absolute effect size is small, the absolute effects do equate to health improvements which at a minimum were one less hospital admission for confusion and 7 less hospital admissions due to falls, as well as potential unmeasured improvements in milder forms of cognitive impairment and less serious falls.
An important lesson learned from conducting the Veterans’ MATES program is that interventions with very specific messages (e.g. reduce hypnotic use) were more effective than that interventions with more generic adverse event messages (e.g. reduce potentially inappropriate medicine use in elderly) [
27]. In addition, repeated messages over time are needed to sustain improvement in use of medicines. Interventions need to be repeated to increase the persuasiveness of the messages [
36]. Building on the success of the first intervention, the second intervention repeated the key messages and informed GPs of their patients who were still using hypnotics via the patient-specific prescriber feedback. The combined effect elicited sustained practice change with a further reduction in hypnotic use following the second intervention.
A successful national level intervention should always be underpinned by behavioural theories. The success of the Veterans’ MATES interventions is underpinned by the social cognitive theory [
32] and the Precede-Procede health promotion framework [
33], the latter which emphasises strong stakeholder engagement. During both interventions, targeted GPs were mailed patient-specific prescriber feedback, which comprised patient’s relevant medicines, a prompt for GP review (notes identifying potential problems) and calls to action as a tool to support cognitive engagement with the materials. The patient-specific prescriber feedback helped GPs to easily identify patients using hypnotics who were at increased risk of adverse events. In addition, tailored educational brochures with recommendations on management techniques and advice to patients were provided to assist GPs in resolving the potential medication-related problems. Veterans were mailed a consumer-focused educational brochure explaining the benefits of non-pharmacological options in the long-term. Engagement of other relevant stakeholders including pharmacists and directors of nursing in residential aged care facilities at a national level maximised the efficiency of outreach to support behaviour and practice change.
Mismatch between GPs’ and patients’ perceptions or expectations often hinders effective management of insomnia [
26]. Patients often try to resolve sleep problems themselves by using various unproven alternative treatments which are ultimately ineffective [
37]. By the time they consult GPs, patients may have the same notion that behavioural and cognitive approaches recommended by GPs will not be effective [
37]. Patients may also feel that GPs cannot help them with insomnia as they think that GPs have other priorities [
26]. Thus, during both interventions, veterans using hypnotics were mailed educational materials emphasizing that their GPs can recommend effective non-drug treatments and asking the veterans to make an appointment with their GPs. In contrast, GPs often assume that patients expect a hypnotic [
26], or that patients already on hypnotics are resistant to stopping the hypnotics [
37]. Based on veterans’ survey response in the first intervention, the second intervention highlighted to the GPs the veterans’ willingness to try non-pharmacological interventions and, for patients already taking hypnotics, their willingness to reduce the amount of hypnotics they were using.
A 2013 systematic review which included eight qualitative studies on GPs experience, one which was conducted in Australia [
38], revealed that the decision to prescribe or withdraw benzodiazepines in primary care patients can be complex, uncomfortable and demanding due to reasons including time-constraints, their perception of patient expectations and the doctor-patient relationship [
39]. The wish to maintain a good doctor-patient relationship and to remain competitive in the healthcare market often involve succumbing to patients’ demand for medicines [
40,
41], which in this context meant prescribing hypnotics. As patients may often have attempted many different, ineffective, alternatives prior to consulting GPs [
37], many would expect a prescription for hypnotics by the time they consult their GPs. To address these issues, patient empowerment and involvement in withdrawing hypnotics was accounted for in constructing the framework of our intervention. The tailored educational brochures sent to veterans described the non-drug options for insomnia, the harms of hypnotic use and the potential to stop hypnotics. A 2017 Cochrane systematic review reported that use of decision aids such as brochures is effective in improving patients’ knowledge of treatment options, patient-clinician communication, and patients’ ability to participate in shared decision-making [
42]. The review also showed that patients were more likely to choose the more conservative treatment option when they were better informed about the benefits and harms of treatment [
42]. It was anticipated that increasing patient understanding and awareness on the risks of hypnotics would make it easier for GPs to communicate the rationale for stopping hypnotics.
Older people are more vulnerable to the side effects of hypnotics due to multimorbidity and multiple medicines that they are taking [
29,
30]. Insomnia management using hypnotics becomes particularly problematic when concomitant medicines not indicated for insomnia treatment also have sedative properties. Many studies have attempted to evaluate the feasibility of withdrawing or reducing hypnotic use in older people and the subsequent improvements in clinical outcomes [
25,
43]. A 2017 systematic review of randomised controlled trials and non-randomised studies evaluated interventions to deprescribe benzodiazepines and benzodiazepine related drugs (Z-drugs) in adults aged 65 years and above [
25]. The included studies were published between January 1995 and July 2015 and involved between 14 and 259 patients aged 65 years and above. One study included a GP-targeted intervention while other studies used pharmacological substitution with melatonin, or mixed approaches such as patient education with tapering advice or temporary pharmacological substitution with psychological support. Different interventions had variable discontinuation rates of between 27 and 80%. Sustainability of these interventions was unknown due to the short follow-up ranging from six weeks to one year. If measured, these intervention trials reported improvements in clinical outcomes using surrogate endpoints such as questionnaire scores, hand grip strength and balance [
25,
43], without reporting important and clinically relevant health consequences such as hospitalisations for falls or confusion. While randomised and non-randomised trials to withdraw or reduce use of hypnotics showed promising results, such trials are expensive, resource-intensive and time-consuming to conduct. Furthermore, these studies were able to target only a small number of patients. The small sample sizes and short follow-up also meant the studies lacked the capacity to detect relatively less common but clinically relevant outcomes such as hospitalisations for falls or confusion. In contrast, our study assessed the effectiveness of reducing hypnotic use among older people at a national level. Our results showed that the interventions achieved sustained reduction in hypnotic use, as demonstrated by the monthly decrease rate in hypnotic use which persisted up to 12 months after each intervention.
An important implication of the Veteran’s MATES programme is the health and economic consequences associated with the reduction in hypnotic use following both interventions. We have previously demonstrated that hypnotic use was associated with increased risk of hospitalisation for falls [
15], and hospitalisation for acute confusion [
18] in the same population. A previous meta-analysis has also shown an increased risk of fractures with hypnotic use [
44]. Using the risk estimates calculated using the Australian veterans’ data [
15,
18], the same population that was targeted by our intervention, we were able to estimate the impact of reduction in hypnotic use on clinically important patient health outcomes. The cumulative reduction in patient-months of hypnotic use was estimated to lead to a reduction in the number of hospitalisations for falls and hospitalisations for confusion. It is important to remember that the health consequences that we have estimated are likely to be an underestimate, as many people have falls or experience confusion that does not result in hospital admission, but which may still have a substantial impact on their health or quality of life. The costs associated with conducting the program are offset by the cost-savings associated with improved use of medicines and associated improvements in health outcomes for veterans, arising from the Veterans’ MATES program.
Our study has several strengths. We used data from a large sample of older Australians and tracked real-time use of hypnotics at the national level. Our results are likely to be generalisable to all older Australians as they had similar rates of medicine use and medical services when compared with veterans targeted in this intervention [
45,
46]. While the results on impact of cumulative reduction in hypnotic use on health outcomes should be regarded as an estimate, it is likely a conservative estimate of improvements in health outcomes.
One of the limitations of this study was that we do not know the number of targeted stakeholders who actually read the information. We only had information on the number of letters returned to sender, which was below 1%. We were unable to evaluate the usefulness of tailored educational brochures in influencing veterans’ willingness to stop hypnotic use or initiate conversations with GPs to withdraw their hypnotics. Similarly, we had no information on the extent to which pharmacists and directors of nursing in residential aged care facilities affected trends in hypnotic use, although the intervention targeted these groups to assist with reinforcement of key messages. Measuring changes in GP prescribing as a result of our intervention is difficult. This is because veterans in Australia are able to visit any GP and do not always visit the same GP. This means that the medicines that they are taking may be prescribed by different GPs. We were unable to determine how many veterans were referred to psychologists to receive cognitive behavioural therapy, and whether there were enough psychologists available to provide cognitive behavioural therapy for the patients referred to them in a timely manner. Patients awaiting cognitive behavioural therapy from a qualified psychologist are unlikely to have been able to reduce their use of hypnotics while waiting for this service and this may have influenced the results of our study.
Acknowledgements
The research was funded by the Australian Government Department of Veterans’ Affairs as part of the delivery of the Veterans’ Medicines Advice and Therapeutics Education Services (Veterans’ MATES) program. Veterans’ MATES is provided by the University of South Australia, Quality Use of Medicines and Pharmacy Research Centre, in association with Discipline of General Practice, The University of Adelaide; Discipline of Public Health, The University of Adelaide; Repatriation General Hospital, Daw Park; NPS – Better choices, Better health; Australian Medicines Handbook; and the Drug and Therapeutics Information Service.