Opioid prescribing and health outcomes in opioid-naive patients: Analysis of a statewide health information exchange

Widespread use of prescription opioids is associated with adverse outcomes. To identify factors associated with adverse health outcomes and health care use using a statewide health information exchange. This is a retrospective cohort study using the Indiana Network for Patient Care. Adult opioid-naive patients who received an opioid prescription between January 2012 and December 2017 were included. The outcomes included (1) a composite outcome of any combination of opioid abuse, dependence, or overdose, (2) all-cause mortality, and (3) health care use. Independent variables included opioid dosage, dispensed amount, days supply, concurrent use of short-acting (SA) and long-acting (LA) opioids, and concurrent use with benzodiazepine or gabapentinoids. Additional variables included patients’ age, sex, race, modified Charlson Comorbidity Index score, mental health conditions, and medications for opioid use disorders. Factors associated with composite outcome and mortality were identified using Cox proportional hazards and reported as adjusted hazard ratio (aHR) and 95% CI. Factors associated with health care use were identified using Poisson regression and reported as adjusted incidence rate ratio (aIRR) and 95% CI. 1,328,287 opioid prescriptions were identified for 341,722 patients. Opioid-related factors associated with the composite outcome, mortality, and hospitalizations, respectively, included opioid dosage (aHR 1.003 [95% CI 1.001–1.006]; aHR not applicable; aIRR 1.07 [1.06–1.08]), opioid days supply (aHR 1.03 [1.02–1.03]; aHR 1.009 [1.005–1.014]; aIRR 0.94 [0.92–0.96]), concurrent SA/LA opioids (aHR 2.12 [1.78–2.54]; aHR 1.40 [1.14–1.70]; aIRR 1.40 [1.37–1.42]), and use of benzodiazepines/gabapentinoids (aHR 1.68 [1.38–2.04]; aHR 1.23 [1.01–1.51]; aIRR 1.25 [1.23–1.27]). Many factors are associated with poor health outcomes, especially concurrent use of SA and LA opioids and overlapping prescriptions of opioids with benzodiazepines or gabapentinoids. Identification of factors associated with adverse outcomes may help identify patients at risk for poor outcomes and could inform possible interventions.