Datapoints: Regional Variation in Benzodiazepine Prescribing for Patients on Opioid Agonist Therapy

P benzodiazepines (BZDs) for patients receiving opioid agonist therapies is controversial (1). Little has been reported on the regional variation of this practice, a possible marker of clinical uncertainty. We examined regional variation in BZD prescribing for opioid agonist therapy patients in the Veterans Health Administration (VHA) for fiscal year 2010. The sample included 5,409 methadone maintenance patients and 5,690 buprenorphine maintenance patients, representing all VHA patients with an opioid use disorder in maintenance therapy. Medication exposure was determined by fill dates and days’ supply, except for methadone maintenance prescriptions, which arenot recorded inVHA administrative data. To be included in the sample, methadone patients had at least two visits to a VHA opioid maintenance clinic and no buprenorphine prescriptions. Concurrent BZD exposure was defined for methadone patients as a BZD prescription and amaintenance clinic visit in the same 30 days; for buprenorphine patients, it was defined as overlapping BZD and buprenorphine prescriptions. Regions were based on the 21 Veterans Integrated Service Networks (VISNs). We excluded VISNs in which fewer than five patients were prescribed BZDs or in which the 95% confidence intervals (CI) around the proportion with concurrent BZD prescriptions were greater than 610% (four and two VISNs, respectively, for the methadone analysis only). A comparison of CIs indicated that ten of 11 VISNs with BZD prescribing of #15% for methadone patients and all 16 VISNs with #25% for buprenorphine patients differed significantly from the one VISN with the greatest BZD prescribing within each opioid agonist type. Considerable variation was found in BZDprescribing between the 21 VISNs included for the buprenorphine and the 15 included for the methadone analyses: range for methadone patients, 6.7%– 24.2%; for buprenorphine patients, 11.0%238.5% (Figure 1). Median BZD prescribing was 13.3% for methadone patients (interquartile range [IQR]5 10.0%217.5% ) and 20.2% for buprenorphine patients (IQR515.3%224.3%). BZD prescribing for patients receiving opioid agonist therapy was not uncommon and varied regionally. This variation in practice may reflect prescriber uncertainty. However, research on the reasons behind it and whether it affects outcomes is warranted.