Factors associated with early opioid dispensing compared with NSAID and muscle relaxant dispensing after a work-related low back injury

Abstract

ObjectivesThe objective of this historical cohort study was to determine the claimant and prescriber factors associated with receiving opioids at first postinjury dispense compared with non-steroidal anti-inflammatory drugs (NSAIDs) and skeletal muscle relaxants (SMRs) in a sample of workers’ compensation claimants with low back pain (LBP) claims between 1998 and 2009 in British Columbia, Canada.MethodsAdministrative workers’ compensation, prescription and healthcare data were linked. The association between claimant factors (sociodemographics, occupation, diagnosis, comorbidities, pre-injury prescriptions and healthcare) and prescriber factors (sex, birth year, specialty) with drug class(es) at first dispense (opioids vs NSAIDs/SMRs) was examined with multilevel multinomial logistic regression.ResultsIncreasing days supplied with opioids in the previous year was associated with increased odds of receiving opioids only (1–14 days OR 1.62, 95% CI 1.51 to 1.75; ≥15 days OR 5.12, 95% CI 4.65 to 5.64) and opioids with NSAIDs/SMRs (1–14 days OR 1.49, 95% CI 1.39 to 1.60; ≥15 days OR 2.82, 95% CI 2.56 to 3.12). Other significant claimant factors included: pre-injury dispenses for NSAIDs, SMRs, antidepressants, anticonvulsants and sedative-hypnotics/anxiolytics; International Statistical Classification of Diseases and Related Health Problems, 9th Revision diagnosis; various pre-existing comorbidities; prior physician visits and hospitalisations; and year of injury, age, sex, health authority and occupation. Prescribers accounted for 25%–36% of the variability in the drug class(es) received, but prescriber sex, specialty and birth year did not explain observed between-prescriber variation.ConclusionsDuring this period in the opioid crisis, early postinjury dispensing was multifactorial, with several claimant factors associated with receiving opioids at first prescription. Prescriber variation in drug class choice appears particularly important, but was not explained by basic prescriber characteristics.