Psychiatric and racial/ethnic differences in incident and long-term benzodiazepine use: A commercial healthcare claims study

Abstract

Objectives

This study evaluated psychiatric diagnoses and race/ethnicity as predictors of both incident and long-term benzodiazepine use.

Methods

We implemented two designs using commercial healthcare claims from Optum's de-identified Clinformatics® Data Mart Database. We first conducted a case-control study and examined 1,904,608 individuals with a new benzodiazepine prescription (ages 13–64 from 2010 to 2019) and matched controls. We used conditional logistic regression to examine 21 potential indications and other psychiatric diagnoses and race/ethnicity as predictors of filling a benzodiazepine prescription. We then used Cox proportional hazards regression in a cohort study among those with a new benzodiazepine to examine the degree to which psychiatric diagnoses and race/ethnicity predicted transitioning to long-term treatment, defined as six months or more of continuous prescription fills.

Results

All included psychiatric diagnoses were associated with incident benzodiazepine use, and most subsequently predicted long-term treatment. Among the most common and strongest predictors for incident use was having any anxiety disorder (Odds Ratio = 5.71; 95% Confidence Interval [CI], 5.67–5.76). Two years after the initial prescription, 8.0% of BZD recipients had met criteria for long-term treatment at least once. Among the strongest predictors of long-term treatment was severe mental illness [e.g., schizophrenia; (Hazard Ratio = 2.36; 95% CI, 2.27–2.47)]. Individuals from all racial/ethnic minoritized groups were less likely to have both incident and long-term benzodiazepine use compared with White individuals.

Conclusions

These findings highlight that transition to long-term treatment is occasionally occurring and notable among BZD recipients, particularly among those with severe mental illness.