Validation of ICD 10 on congenital anomalies in the state of Penang

Background: Database on hospital records like dischargedata, birth and death certificates are widely used forepidemiological and research studies. However there area very few validation studies on these data. The aim ofthis study was to validate and assess the accuracy of theICD 10 database on congenital anomalies in the state ofPenang. This study was carried out for three years, from2002 to 2004.Methods: The list of cases coded under the general coding“Q” was extracted and approximately 30% of cases wererandomly selected from the list. Medical records for theselected cases were checked and discrepancies for thediagnoses between the medical records and the ICD 10data base were recorded for three years. Verification wasdone for basic demographic variables and the coding ofthe diseases. Discrepancies, sensitivity and specificitywere calculated.Results: The ICD 10 database for congenital anomaliesare classified into two types: Type 1 and Type 2.Discrepancies on demographic information were foundamong the age of patients (babies with congenitalanomalies). In Type 1, there was a discrepancy of about0.02 % to 0.05% probability that a congenital anomalycase can be recorded as non congenital anomaly inthe ICD 10. In Type 2 there was a discrepancy that anon-congenital anomaly was classified as congenitalanomaly and this ranged from 26.7% to 50.0%. Thesensitivity ranged from 96.85% to 97.98%, thus it canbe concluded the ICD 10 database is highly sensitivewhile the specificity ranged from 50.00% to 78.57 %. Inother words the ICD 10 is not accurate when classifyingthe non- congenital anomaly cases. A fair percentage ofnon-congenital anomaly cases were classified as CA inthe ICD 10 database.Conclusion: Even though hospital databases areused as a baseline data for a number of research andepidemiological studies it cannot be used at facevalue. Validation of these data is necessary before anyconclusions can be drawn or intervention measures areundertaken