OBJECTIVES: To evaluate the potential of clinical, demographic and utilization factors in predicting treatment costs for diabetic foot ulcers. METHODS: Retrospective analysis of administrative outpatient pharmacy and medical claims from 1/96 12/98 was conducted using Protocare Sciences Managed Care Database. Diabetic patients were identified using diagnoses and medication claims. Diagnosis of chronic skin ulcer of lower limbs defined foot ulcer. Descriptive analysis of demographic, clinical, and prior medical resource use information was performed, followed by multivariate regression to estimate the potential of these variables in explaining foot ulcer treatment charges. RESULTS: The prevalence of diagnosed foot ulcers was 1.4% (2,881 of the 212,563 identified diabetics). Patients with foot ulcers were older (mean age 64.9 years vs. 60.7 years) and more likely to be male (59.6% vs. 50%) than the non-foot ulcer diabetic population. Diabetes-related complications and comorbidities were common including skin infections, peripheral vascular disorders, and neurologic problems. During three months following onset of foot ulcer, related care represented 40% of the $2,596 average monthly charges per patient. Using regression, demographic, clinical and prior resource utilization information explained one-fifth of the ulcer-related charges (R-square = 0.22). Ulcer severity at onset, previous cellulitis or skin abscess, and pentoxifylline use were the strongest contributors. CONCLUSIONS: Although claims data cannot fully explain variations in foot ulcer treatment costs, clinical factors, such as markers for ulcer severity and active treatment of peripheral vascular disease are predictors of treatment costs. The correlation between previous skin problems and treatment costs was negative, suggesting that heightened awareness through past experience may result in lower treatment costs.