Federal Guideline Change and its Effects on Opioid-Benzodiazepine Concurrent Use

Although many studies have addressed the adoption of new drugs and procedures, less is known about the process of de-adopting practices that have been shown to cause harm. We used the case of coprescription of opioids and benzodiazepines to study the effect of the CDC opioid prescribing guidelines for patients with chronic pain. Coprescription of opioids and benzodiazepines was found to increase the risk of overdose. We used administrative claims data from the OptumLabs® Data Warehouse (OLDW) –a comprehensive, longitudinal, de-identified claims data repository—to identify beneficiaries with any opioid fills between January 2014 and February 2017. We identified overlapping prescriptions for benzodiazepines, assessing whether the same physician prescribed both drugs. We stratified across three populations: commercial, Medicare Advantage beneficiaries aged 65+ (aged Medicare), and Medicare Advantage beneficiaries aged <65 (disabled Medicare). Opioid use was classified as a long-term episode if it spanned 90+ calendar days and included 120+ days supply or 10+ fills. We assessed both prevalence of overlapping prescriptions and the intensity of the overlap, if any. Intensity was measured as the proportion of opioid days in which benzodiazepines were also available. We used piecewise regression to model the impact of the guidelines on coprescription of opioids and benzodiazepines, allowing both a change in level at the time of the intervention and a change in the trend slope after the intervention. We allowed different effects of the intervention on long-term and short-term episodes of opioid use, as the key population addressed by the guidelines is chronic users of opioids. Other covariates included beneficiary age, sex, race/ethnicity, state of residence, and whether the beneficiary had a cancer diagnosis or hospice claim in the 90 days prior to the opioid episode. We identified 2.8 million beneficiaries contributing 5.3 million months with any opioid use, of which 3% of episodes and 16% of person-months represented long-term use of opioids. We found little impact of the release of the CDC guidelines among beneficiaries with long-term use. Only the disabled Medicare group showed a statistically significant change in trend after the guidelines were released. We observed modest reductions in level at the time of the intervention for the aged Medicare group in both prevalence of coprescription and intensity of coprescription. We found that having the same physician prescribing both opioids and benzodiazepines modestly increased the intensity of coprescription. Contrary to results in the population using opioids long term, we found decreases in both trend and level of prevalence and intensity of coprescription in short-term opioid use episodes. The release of the CDC guidelines had little impact on the prevalence or intensity of coprescription of opioids and benzodiazepines for long-term users of opioids, but modestly reduced coprescription during short-term episodes. Physicians may face substantial difficulty helping complex patients discontinue use of these drugs, both of which can cause withdrawal symptoms. The small but consistent increase in coprescription intensity associated with the same physician prescribing both drugs may indicate that prescription drug monitoring programs will have minimal impact on deimplementation of this practice.