Abstract Background/Aims Social risk factors such as poverty and financial strain are strongly linked to poor health. Their measurement is crucial for efficiently guiding services and resources that improve patient and population health. The absence of individual socioeconomic measures in electronic health records (EHRs) has led to the routine use of area-based measures, such as the Index of Multiple Deprivation (IMD), as proxies for individual measures. However, inaccuracies can arise when using an area-based measure as a surrogate for individual-level poverty. The impact of these substitutions on individuals reporting chronic pain (CP), one of the leading causes of disability and loss of productivity globally, is unclear. This study aims to explore whether individual-level measures (e.g. financial adequacy, food poverty, living conditions) better predict CP and high-impact chronic pain (HICP) outcomes than area-level information. We further assessed the social determinant factors that are associated with these conditions in primary care consulters. Methods Data from the MIDAS population survey of adults aged 35 and over in North Staffordshire were analysed. The IMD, an area measure of deprivation, was used as an area-level indicator, including its seven domains. Individual indicators include income adequacy, education and other social risk measures. We used a multilevel model with individuals clustered within general practices and assessed the added discriminating value of each indicator using the area under the curve (AUC). The independent association of these social determinants of health (SDoH) factors with CP and HICP were further investigated. Results The prevalence of CP and HICP in the population were 43.27% and 22.49%, respectively. Individual indicators such as financial adequacy, education, and loneliness were stronger predictors of CP and HICP than IMD and its seven domains in a baseline model, adjusting for age, sex and ethnicity. However, these added discriminating values were not statistically significant when BMI, physical activity and depression and anxiety were adjusted for in these models. In models assessing the impact of the combined measures, there was no statistical evidence of contextual area effect. However, individuals experiencing poor housing quality, transport poverty, loneliness, financial strain and no educational qualification had higher odds of chronic pain compared to their counterparts. These associations were even stronger in those experiencing HICP, indicating that those experiencing worse outcomes have worse social circumstances. Conclusion The study reinforced the need for individual social risk measures in health data registry and electronic health records to complement the IMD that are readily available. These will not only aid in discerning the mechanistic pathways through which SDoH impact health but also in identifying those that will benefit the most from interventions addressing health equity. Disclosure M. Abdullateef: Grants/research support; ESRC CASE studentship (ES/P000665/1). E. Parry: Grants/research support; Nuffield Foundation’s Oliver Bird Fund and Versus Arthritis (OBF/43990). R. Wilkie: Grants/research support; Nuffield Foundation’s Oliver Bird Fund and Versus Arthritis (OBF/43990). D. Yu: Grants/research support; Nuffield Foundation’s Oliver Bird Fund and Versus Arthritis (OBF/43990).
Abdullateef et al. (Wed,) studied this question.