Validation of claims data algorithms to identify nonmelanoma skin cancer

Health maintenance organization (HMO) administrative databases have been used as sampling frames for ascertaining nonmelanoma skin cancer (NMSC). However, because of the lack of tumor registry information on these cancers, these ascertainment methods have not been previously validated. NMSC cases arising from patients served by a staff model medical group and diagnosed between 1/1/07 to 12/31/08 were identified from claims data using three ascertainment strategies. These claims-data cases were then compared to NMSC identified using natural language processing (NLP) of electronic pathology reports (EPR), and sensitivity, specificity, positive (PPV) and negative predictive values (NPV) calculated. Comparison of claims data ascertained cases to the NLP demonstrated sensitivities ranging from 48-65% and specificities from 85-98%, with ICD-9-CM ascertainment demonstrating the highest case sensitivity though the lowest specificity. HMO health plan claims data had a higher specificity than all payer claims data. A comparison of EPR and clinic log registry cases showed sensitivity of 98% and specificity of 99%. Validation of administrative data to ascertain NMSC demonstrates respectable sensitivity and specificity though NLP ascertainment was superior. There is a substantial difference in cases identified by NLP compared to claims data suggesting that formal surveillance efforts should be considered.