Scottish Census Research Articles

O-24 Total duct excision is still required if breast cancers are not to be missed

Material deprivation contributes to inequalities in health; areas of high deprivation have higher rates of ill-health. How deprivation is measured has a great impact on its explanatory power with respect to health. We compare previous deprivation measures used in Scotland and proposes a new deprivation measure using the 2001 and 2011 Scottish census data. We calculate the relative index of inequality (RII) for self-reported health and mortality. While across all age groups different deprivation measures provide similar results, the assessment of health inequalities among those aged 20–29 differs markedly according to the deprivation measure. In 2011 the RII for long-term health problem for men aged 20–24 was only 0.71 (95% CI 0.60–0.83) using the Carstairs score, but 1.10 (0.99–1.21) for the new score and 1.13 (1.03–1.24) for the income domain of Scottish Index of Multiple Deprivation (SIMD). The RII for mortality in that age group was 1.25 (0.89–1.58) for the Carstairs score, 1.69 (1.35–2.02) for the new measure and 1.76 (1.43–2.08) for SIMD. The results suggest that researchers and policy makers should consider the suitability of deprivation measures for different social groups.

Research Article: Estimating local variations in land use statistics

The Scottish agricultural census provides land use statistics as summaries for parish areas. We investigated the dis-aggregation of these parish summaries using the Land Capability for Agriculture and the 1988 Land Cover of Scotland as supporting data sets. It is unlikely that allocation rules to implement the dis-aggregation should be identical across all parishes. Equally, rules for individual parishes are indeterminate. The 891 parishes were classified into nine classes, then each class was regionalised, creating 91 regions overall. Allocation rules were estimated independently for each region and class. Statistical testing identified greater variations in the rules than is expected by random allocation of regions to classes.