Feasibility of Creating a Diagnosis-Based Diabetic Foot Registry in a Large Health Care Provider

We aimed to assess the validity of diagnostic codes relating to diabetic foot ulcer (DFU) in real-world data seeking to create a national DFU registry. Data were obtained from the diabetes registry of Maccabi Healthcare Service (MHS), a 2.1 million member sick fund in Israel, which included 122,500 patients in 2015. We randomly selected and reviewed ∼400 patient files from each of the following categories: 1) had a diagnostic code of DFU in 2015; 2) had a diagnostic code suggestive of DFU in 2015 including: leg ulcer, amputation, DFU in proximate years or abnormality reported by nurse; 3) high risk patients (peripheral arterial disease (PAD), neuropathy, etc.). The patients' charts were reviewed by study physicians and DFU was validated or refuted. Relying upon diagnostic codes entered by physicians, the positive predictive value was 73.1% (95% CI 67.6-78.2) and the sensitivity was 48.2% (95% CI 45.8-50.7%). We therefore conclude that the diagnostic codes alone cannot be reliably used to create a DFU registry, and a prospective registry is considered. The estimated annual prevalence of DFU in the diabetes registry of MHS in 2015 was 1.3% (95% CI 1.0-1.6%). Clinical features and outcomes of patients with validated diagnostic codes of DFU are shown (Table). In multivariate analysis: age, diabetes duration, socioeconomic status, previous amputation and PAD were independently associated with risk of amputation or death.