Pursuant to the concept of a “zip code lottery” of outcomes in diabetic foot complications, we share the excellent work from our colleagues from Glasgow and Amsterdam.
Geospatial mapping and data linkage uncovers variability in outcomes of foot disease according to multiple deprivation: a population cohort study of people with diabetes
Aims/hypothesis Our aim was to investigate the geospatial distribution of diabetic foot ulceration (DFU), lower extremity amputation (LEA) and mortality rates in people with diabetes in small geographical areas with varying levels of multiple deprivation.
Methods We undertook a population cohort study to extract the health records of 112,231 people with diabetes from the Scottish Care Information – Diabetes Collaboration (SCI-Diabetes) database. We linked this to health records to identify death, LEA and DFU events. These events were geospatially mapped using multiple deprivation maps for the geographical area of National Health Service (NHS) Greater Glasgow and Clyde. Tests of spatial autocorrelation and association were conducted to evaluate geograph- ical variation and patterning, and the association between prevalence-adjusted outcome rates and multiple deprivation by quintile. Results Within our health board region, people with diabetes had crude prevalence-adjusted rates for DFU of 4.6% and for LEA of 1.3%, and an incidence rate of mortality preceded by either a DFU or LEA of 10.5 per 10,000 per year. Spatial autocorrelation identified statistically significant hot spot (high prevalence) and cold spot (low prevalence) clusters for all outcomes. Small-area maps effectively displayed near neighbour clustering across the health board geography. Disproportionately high numbers of hot spots within the most deprived quintile for DFU (p < 0.001), LEA (p < 0.001) and mortality (p < 0.001) rates were found. Conversely, a disproportionately higher number of cold spots was found within the least deprived quintile for LEA (p < 0.001). Conclusions/interpretation In people with diabetes, DFU, LEA and mortality rates are associated with multiple deprivation and form geographical neighbourhood clusters.