Prescription duration of benzodiazepines: a retrospective cohort study of prescription duration in primary care.

BACKGROUND AND OBJECTIVES: Benzodiazepines are approved for and used in the treatment of anxiety, insomnia, seizures, and alcohol withdrawal. However, benzodiazepine prescriptions often do not align with regulatory approved (i.e. "on-label") indications, or duration, which can result in serious adverse drug events, especially among older people. Benzodiazepine labels universally warn against long-term use due to known adverse effects, dependence, and withdrawal symptoms. It is therefore important to understand factors associated with durations longer than regulators approved (i.e. "off-label" duration). This can be challenging due to a lack of data on prescribed duration of treatment and treatment indication. The aim of this study is to identify risk factors for long-term benzodiazepine use. METHODS: Data from the MOXXI (Medical Office of the XXIst century) electronic health record system in Quebec Canada was used to collect benzodiazepine prescription indication and duration, which was then integrated with the public pharmacare system to determine dispensation in the period between December 17th, 2002 and December 31st, 2023. The entire dataset was used. Each drug indication and duration was retrospectively classified as either on-label or off-label, and short-duration (≤30 days) or long-duration (>30 days) according to the Health Canada drug database. To assess determinants of long-term use, a multi-level approach was used, with the patient-drug indication pair as the unit of analysis. Duration of prescription was evaluated as the outcome. RESULTS: 69.2% of benzodiazepines were prescribed long-term. The odds of being prescribed a benzodiazepine long-term increased with age (1.01 (1.01-1.01) per year) and was more probable for male patients (1.13 (1.04-1.22)) than female. Patients taking concomitant medication are more likely to be prescribed benzodiazepines long-term (1.05 (1.04-1.07)) per additional medication. Male physicians were more likely to prescribe benzodiazepines long-term (1.29 (1.19-1.40)), and there was a significant relationship between institution of medical school training and long-term benzodiazepine prescribing behavior. Off-label benzodiazepines were less likely to be prescribed long-term compared to on-label benzodiazepines (0.81 (0.76-0.87)). CONCLUSION: Using 21 years of indication-linked e-prescribing and pharmacare data, we found benzodiazepines were rarely limited to short-term use. Older age and an increasing number of concomitant medication were associated with long-term use. These findings underscore the need for prescriber education and highlight e-prescribing systems' value for monitoring and improving prescribing behavior.