Validation of International Classification of Diseases Codes for Dermatologic Conditions

This systematic review of 59 studies evaluated the accuracy of International Classification of Diseases (ICD) codes for identifying dermatologic conditions in US-based datasets. The review found that classification accuracy varied significantly depending on the condition and coding strategy, with the most accurate algorithms often requiring multiple codes, dermatologist involvement, or additional data such as prescriptions. Positive predictive values frequently exceeded 90%, but some common conditions, like seborrheic dermatitis, vitiligo, and rosacea, lacked validated approaches. The study emphasizes the importance of using validated classification algorithms in research to avoid biased results and highlights the need for further research to address gaps in classification for common and understudied conditions.