Predicting ulcer-free survival using the discriminative value of screening test locations.

Abstract

Current guidelines recommend screening the feet of diabetic subjects with a 10-g monofilament or tuning fork. We investigated which tests and locations on the feet have the best predictive value regarding 1-year ulcer-free survival in diabetic subjects participating in the prospective Rotterdam Diabetic Foot Study. Decision tree analysis was used to predict ulcer-free survival based on responses from individual test locations (monofilaments on 10 sites, vibration sense was tested on both halluces and medial malleoli). Separate trees for patients with and without a history of diabetic foot ulcer (DFU) were developed. Four hundred sixteen subjects (mean [SD] age, 61.8years [12.4]; range, 21.6-90.2) were measured, of whom 24 developed new DFUs. Three tests exhibited discriminative and predictive properties: testing vibration sense on the medial malleolus and monofilament testing on heel and hallux. The decision tree to predict ulcer-free survival in patients with a history of DFU yielded a sensitivity of 87.0%, which was 99.6% for the tree of patients without a history of DFU. The findings of this study aids medical decision making by discriminating between high- and low-risk patients of developing DFU using selective testing on sites with predictive properties.