This study aimed to quantify dispensed opioid prescriptions among primary care practices throughout England and investigate its association with socioeconomic status (SES).
This cross-sectional study used publicly available data in 2015, including practice-level dispensing data and characteristics of registrants from the United Kingdom (UK) National Health Service Digital, and Index of Multiple Deprivation (IMD) data from Department of Communities and Local Government. Practices in England which had opioid prescriptions that could be assigned a defined daily dose (DDD) in the claim-based dispensing database were included. The total amount of dispensed opioid prescriptions (DDD/1000 registrants/day) was calculated for each practice. The association between dispensed opioid prescriptions and IMD was analyzed by multi-level regression and adjusted for registrants’ characteristics and the clustered effect of Clinical Commissioning Groups. Subgroup analysis was conducted for practices in London, Birmingham, Manchester and Newcastle.
Of the 7856 included practices in England, the median and interquartile range (IQR) of prescription opioids dispensed was 36.9 (IQR: 23.1, 52.5) DDD/1000 registrants/day. The median opioid utilization (DDD/1000 registrants/day) amongst practices varied between Manchester (53.1; IQR: 36.8, 71.4), Newcastle (48.9; IQR: 38.8, 60.1), Birmingham (35.3; IQR: 23.1, 49.4) and London (13.9; IQR: 8.1, 18.8). Lower SES, increased prevalence of patients aged more than 65 years, female gender, smoking, obesity and depression were significantly associated with increased dispensed opioid prescriptions. For every decrease in IMD decile (lower SES), there was a significant increase of opioid utilization by 1.0 (95% confidence interval: 0.89, 1.2, P<0.001) DDD/1000 registrants/day.
There was a variation in prescription opioids dispensed among practices from Northern and Eastern England to Southern England. A significant association between increased opioid prescriptions and greater deprivation at a population level was observed. Further longitudinal