Measuring child health inequalities using aggregations of Enumeration Districts

In demonstrating health variations between different areas in a district, it is conventional to use local authority ward-toward variations. In rural districts, because wards have small, heterogeneous populations, this method is less useful. We have investigated alternative ways of showing variations in child health by using different aggregations of Enumeration Districts (ED) in a small, sparsely populated rural area. EDs were aggregated first by a cluster analysis based classification (Super Profiles) and second according to a material deprivation score (the Townsend score). Both methods of aggregation showed similar variations between areas in the proportion of babies with low birthweight, the proportion of teenage mothers, immunization coverage and six-week screening examination coverage. Both methods discriminated better than a straightforward ward-based analysis. Neither method was clearly superior to the other. It is concluded that for both epidemiological research and for health service information purposes, either of these methods of ED aggregation has definite advantages over ward-based analyses in rural areas.