Characterisation of children's asthma status by ICD-9 code and criteria-based medical record review

Objective: To characterise the relationship between ICD code-based (i.e. physician diagnosis-based) and criteria-based asthma ascertainment. Methods: We compared identification of children with asthma between criteria-based medical record review for asthma ascertainment and an ICD-9 code-based approach. We determined the agreement rate and validity index of ICD code-based asthma ascertainment using asthma status by medical record review as a gold standard. Results: Of the 115 study subjects, the agreement between medical record review and ICD-9 coding was 81.6% with a kappa value of 0.28 (P<0.0001). Sensitivity, specificity, positive and negative predictive values for ICD-9 code were 24.0%, 97.8%, 75.0%, and 82.0%, respectively, using criteria for asthma by medical record review as gold standard. Conclusions: ICD code-based asthma ascertainment appears to under-identify children with asthma compared to criteria-based medical record review. ICD codes may be useful for etiologic research but may not be suitable for surveillance of asthma epidemiology.