Examining non-linearity in the association between age and reported opioid use in different socioeconomic strata: cohort study using Health Survey for England waves from 1997 to 2014

Background Age and socioeconomic status (SES) predict several health-related outcomes, including prescription opioid use. Contrasting findings from previous literature found higher prevalence of opioid use in both people over 65 years old and the working-age population of 35-55 years old. This study aimed to analyse if the association between age and opioid use is non-linear and differs in adults with different SES levels. Methods This cohort study used the Health Survey for England waves 1997-2014 data to investigate the shape of the correlation between reported opioid use and income decile, employment status and educational level. A semiparametric Generalised Additive Model was employed, so that linearity of correlation was not assumed. The shape of the relationship was assessed using the effective degrees of freedom (EDF). Results Positive correlation between age and reported opioid use, more linear in people in the highest income decile (EDF: 1.01, p<0.001) and higher education (EDF: 2.03, p<0.001) was observed. In people on lower income and with lower levels of education, the highes probability of reported opioid use was at around 40-60 years old and slowly decreased after that. Higher income decile and higher levels of education were predictors of a lower probability of reported opioid use (OR: 0.27, 95% CI: 0.21 to 0.36 and OR: 0.48, 95% CI: 0.41 to 0.57, respectively). There was no statistically significant difference in opioid use between employed and unemployed people. Conclusion The relationship between age and the probability of prescribed opioid use varies greatly across different income and educations strata, highlighting different drivers in opioid prescribing across population groups. More research is needed into exploring patterns in opioid use in older people, particularly from disadvantaged socioeconomic backgrounds.