Childhood Type 1 Diabetes Mapping Using Routinely Collected Hospital Data

Abstract

O-31A5-2 Background/Aims: Childhood onset insulin-dependant (type 1) diabetes mellitus shows spatial trends in incidence at the European and UK regional level. In lieu of a dedicated diabetes register for England, hospital records were used to create a surrogate of a national childhood diabetes register for use in epidemiological studies. This hospital records based proxy register validated well against regional registry data, especially for younger children (aged, 0–9 years) and for more recent years (from 2000). Our aim was to use this surrogate register to map childhood diabetes incidence at the small area level across England, to explore spatial trends in risk, and investigate demographic and social variables that may be associated with diabetes risk. Methods: Incidence of type 1 diabetes in children aged 0–9 years over the period 2000 to 2006 was mapped at the district level to explore spatial patterns in disease risk. Diabetes incidence was explored in relation to population density, urban/rural status, area-level socioeconomic status (Carstairs index 2001), overcrowding, and social class. Results: There was substantial variation in diabetes incidence by district across England. Diabetes incidence rate was 10% higher in rural than urban districts (P = 0.02); 22% higher in districts in the lowest population density quintile compared to the highest quintile (P = 0.01); and 18% higher in districts in the lowest social class quintile (measured as the highest percentage of persons with a semiskilled/unskilled manual worker as household head) compared to the highest social class quintile (P < 0.01). Incidence did not significantly differ with respect to overcrowding or Carstairs index. Conclusion: Incidence of childhood type 1 diabetes shows spatial variation across England. In keeping with previous studies, higher diabetes incidence was associated with rurality and lower population density. The associations between these social and demographic variables and diabetes risk, and the mechanisms by which they operate require further investigation.