Validation of in-hospital diagnosis codes in one French hospital and out-hospital algorithm to identify skin ulcers in healthcare databases in France.

Abstract

Purpose: This study has two main objectives: 1/ to validate the International Classification of Diseases, 10th revision (ICD-10) diagnostic codes of skin ulcer in one French hospital using medical charts; 2/ to validate an out-hospital algorithm against ICD-10 codes using a healthcare database.

Methods: We first validated in-hospital ICD-10 codes for pressure, diabetic and vascular skin ulcers using the Grenoble University Hospital medical charts. Secondly, we assessed the validity of an out-hospital algorithm using dressing reimbursements, medical exams and comorbidities to identify skin ulcers using the French "échantillon généraliste des benéficiaires" database. We then compared the type of skin ulcers in patients hospitalized 1 year around the out-hospital skin ulcer identification date. We calculated specificity, sensitivity, positive predictive value (PPV) and negative predictive value (NPV).

Results: The performances of ICD-10 codes for identifying patients with vascular, diabetic and pressure ulcers were all superior to 70%. The out-hospital identification of skin ulcers selected very different patients, younger and with less comorbidities than those hospitalized for skin ulcers. In patients hospitalized 1 year before or after the first dispensation of wound dressings, the concordance with ICD-10 codes was modest. Indeed, patients are wrongly classified as pressure ulcers, vascular ulcers and diabetic foot ulcers in respectively 27.7%, 52.0% and 48.8% of skin ulcers.

Conclusion: We found that performances of the in-hospital identification of pressure, vascular and diabetic foot ulcers were high allowing to use them to conduct observational studies in healthcare databases. However, outpatient identification retrieved heterogeneous performance, we therefore advise researchers using the latter to perform a sensitivity analysis restricted to hospitalized patients.