OP65 Developing a new small-area measure of deprivation using scottish census data

Abstract

Background Area-level measures of material deprivation are important for understanding health inequalities and used by governments to target funding to the most disadvantaged communities. The effectiveness of a measure in achieving these goals depends on how well it reflects actual deprivation and on its validity for any particular social group. We compare health inequalities using common small-area deprivation measures for all ages and for specific age groups to those using a new score. Methods Deprivation is measured using the Carstairs score and its components (male unemployment, overcrowding, no car ownership and low social class), the income domain of the Scottish Index of Multiple Deprivation, and a new score comprising employment, educational qualifications, socioeconomic classification (NS-SeC) and social renting. We use datazone (average population 800) level data from the 2001 and 2011 Scottish censuses for deprivation variables and self-reported health (evaluations of general health and presence of long-term illness). Age-standardised mortality rates (ASMR) are calculated for 2000–2002 and 2010–2012, and the health measures for the census years using the 2013 European Standard Population. We assess the relationship between self-reported health and individual components of the deprivation scores using correlation coefficients. Mortality rates are calculated for population-weighted deprivation quintiles for all ages and for different age groups. We then compare the relative index of inequality (RII) for different ages and deprivation measures. Results The different measures of deprivation are significantly associated with self-reported health for all ages, but correlation coefficients vary substantially. For most components of deprivation the correlation coefficients with long-term health problems and bad general health range from 0.75–0.85. But for overcrowding the coefficients were respectively 0.67 and 0.68 in 2001 and 0.53 and 0.55 in 2011. The RII for ASMR across all ages and for older age groups (50–54 and older) are fairly similar for the different deprivation measures, but substantial differences in RII are evident at younger ages. In 2001 the RII for men aged 20–24 is 0.59 (95% CI 0.32–0.89) when using no car ownership, while with no qualifications as the deprivation measure it is 1.84 (CI 1.58–2.09). Conclusion Across all ages the different deprivation measures mostly give very similar results, but substantial variation exists for younger age groups. Disagreements in RII for all cause mortality are large among those aged 20–24, meaning that some conventional measures (no car ownership) may significantly underestimate inequalities in health. The new area-based deprivation measure captures health inequalities among younger people more effectively.