GEOGRAPHICAL MAPPING OF CHILDHOOD DIABETES IN THE NORTHERN ENGLISH COUNTY OF YORKSHIRE

This study aimed to highlight geographical differences in childhood Ty pe 1 (insulin dependent) diabetes mellitus (IDDM) by mapping incidence at 3 different geographical scales, within the northern English count y of Yorkshire. Mapping techniques are applied to incident cases from a population-based regional register of childhood IDDM. The Yorkshire Children's Diabetes Register (YCDR) comprises 1310 children (0-14 year s) diagnosed with IDDM from 1978-1990. Age standardized incidence rate s (cases/100000/year) are given for administrative county, standardize d incidence ratios (SIRs) are mapped by local government district and electoral ward. Heterogeneity between areas was assessed by a chi(2) t est. At county level, incidence is 25 % higher in Humberside (16.82 pe r 100000 per year) compared to West Yorkshire (12.21 per 100000 per ye ar) (p < 0.001). SIRs for the 22 districts display significant heterog eneity (p < 0.001) with deficits in the urban areas of Kirklees (70 95 % Cl 57-85) and Bradford (81 95% Cl 68-95) and significant excesses in the rural districts of Ryedale (147 95% Cl 106-198), Beverley (149 95 % Cl 113-193), Holderness (166 95% Cl 112-237) and Boothferry (186 95% Cl 134-250). At the smallest scale, variation between wards is signif icant (p < 0.001) and low incidence in urban areas is notable. Geograp hical variation in the incidence of childhood IDDM may provide strong clues to its aetiology. Within Yorkshire, geographical distribution sh ows significant heterogeneity at three different scales. The magnitude of the variation is not explained by ethnic or genetic differences in the population and underlines the important influence of environmenta l factors in disease aetiology.