Assessing the risk of lower extremity amputations using an administrative data-based foot risk index in elderly patients with diabetes

Background: The aims of the present study were to derive a foot risk index using administrative data similar to the International Working Group on the Diabetic Foot clinical risk classification scheme and to evaluate its association with the risk of initial lower extremity (through foot or above) amputation (ILEA). Methods: Merged Veterans Health Administration administrative and Medicare inpatient and outpatient claims data for a historical cohort of Veterans Health Administration users with diabetes were analyzed. Individuals with diabetes, aged ≥67years in 1999, were classified into seven foot risk categories (where 0=no foot risk conditions; and 6=severe foot conditions). The outcome variable was the incidence of an ILEA over the period 2000-2004. Foot risk was derived from 1998 to 1999, and other independent variables were derived from 1999. The associations between foot risk category and risk of ILEA were evaluated using Cox proportional hazards regression models. Results: Of 255534 individuals with diabetes, 54.8% had a documented foot condition in 1999 and 6869 (26.9 per 1000) had an ILEA over the period 2000-2004. We found increased rates of ILEAs as foot risk increased. Based on a Cox regression model containing all independent variables, the hazard ratios for risk of amputation increased progressively from 1.13 (95% CI 1.00-1.28) for the lowest foot risk category to 6.75 (95% CI 6.31-7.23) for the highest foot risk category, compared with individuals without defined foot conditions. Conclusions: An administrative data-derived foot risk category was associated with risk of ILEA in a risk level-dependent pattern.