Longitudinal Analysis of Opioid Analgesic Dose and Diabetes Quality of Care Measures.

Abstract

To examine associations of opioid analgesic dose with quality of care for diabetes mellitus. Longitudinal statewide cohort. Subjects with diabetes filled one or more prescriptions for Schedule II/III opioids for noncancer pain in Blue Cross Blue Shield of Texas from 2008 through 2012. Opioid dose and outcomes were assessed in 6-month intervals after first filled prescription. Two morphine equivalent dose measures were daily dose and quartiles of total dose from all filled prescriptions. In fixed effects models adjusted for clinical and treatment variables, associations of opioid measures were examined for five outcomes: hemoglobin A1c (HbA1c) test, low density lipoprotein cholesterol (LDL) test, any hospitalization, any diabetes-related preventable hospitalization, and any emergency department (ED) visit. All daily and total opioid doses were associated (P < 0.05) with poorer outcomes for all five measures. For HbA1c testing, adjusted odds ratios (AORs) were reduced by 19% for high daily dose (≥100 mg) and highest quartile total dose (>900 mg), respectively, vs no opioids but >900 mg total dose had the lowest AOR for LDL testing (0.74 [CI 0.68, 0.80]). The AORs of any hospitalization or diabetes-related hospitalization were, respectively, 8.19 (CI 7.21, 9.30) and 2.76 (CI 2.19, 3.48) for >900 mg total dose but only 6.22 (CI 4.94, 7.83) and 2.16 (CI 1.34, 3.48) for >100 mg daily dose. Both opioid measures had nonmonotonic associations with ED use. Daily opioid dose but especially total dose of opioids was strongly associated with poorer diabetes quality of care in a statewide cohort.