OPIOID PRESCRIBING ACROSS NEIGHBOURHOOD INCOME GROUPS FOR ARTHROPLASTY PATIENTS IN NOVA SCOTIA

Pain management is a priority for arthroplasty patients and opioids are frequently an integral component of treatment. In recent years there have been concerns of an opioid epidemic fueled, in part, by an increase in prescriptions for pain. Research in the United States has suggested lower socioeconomic status is associated with an increased likelihood of opioid prescribing including those undergoing orthopaedic surgery. The purpose of this study was to examine opioid use across neighbourhood income level for knee and hip arthroplasty patients. The study cohort was constructed from heath administration data using Canadian Classification of Health Intervention codes to select all primary knee and hip procedures, excluding hip fractures, in Nova Scotia from 2017–2021. The first procedure for each individual over the period was selected. The Drug Information System (DIS) database, which includes all prescriptions (RXs) filled at pharmacies, was used to determine opioid prescribing. Income deciles were derived from 2016 Census dissemination area aggregate data and was linked to the study cohort using residential postal code at the time of surgery. Total morphine milligram equivalency (MME) was calculated to standardize morphine dose across prescriptions. All RXs over the study period were totaled and compared to those only at surgical discharge defined as a RX filled on the day of discharge or the day after. The extremal quotient (EQ), the ratio of the highest to the lowest, was calculated to measure variability and ANOVA tests were used to investigate differences across deciles at a 95% confidence level. There were 7646 and 5181 individuals who had knee or hip arthroplasty over the study period. The highest number of procedures occurred in decile six for both procedures. The mean total MME over the study period was 14,052 for knee patients and 13,105 for hips. MME decreased as neighbourhood income increased with the highest and lowest values occurring in the lowest and highest deciles, respectively. The EQ was 2.2 for knees and 2.8 for hips indicating the level of morphine was over twice as high in the lowest income grouping compared to those in the highest incomes. The ANOVA test showed a statistically significant difference between decile groups (p-value < 0.0001 for knees and 0.029 for hips). The mean MME for RXs at surgical discharge was evenly distributed across income deciles with an ANOVA test failing to show any statistical differences (p-value of 0.745 for knees and 0.441 for hips) Our study indicates that more opioids are prescribed for those arthroplasty patients in lower neighbourhood income areas across the entire continuum of care. At discharge from surgery specifically, however, opioid levels are evenly distributed suggesting prescribing practices outside of perioperative arthroplasty are driving the disparity. Regardless, given the potential harms attributed to opioids, surgical teams should nonetheless be cognizant of differences in opioid use across socioeconomic groups in both pre and post- surgical periods.