Objectives: to evaluate and validate the adoption of an algorithm for the identification of cases of congenital anomalies (CAs) to improve the performance of the Congenital Malformations Registry of Sicily Region (Southern Italy).
Design: an algorithm was used to identify congenital anomalies on a sample of hospital discharge records (SDO) with ICD-9-CM code between 740-759 on any of the diagnoses within the first year of life, together with a sample of healthy births equal to 5% of total births for the same period. The identified cases were evaluated through the clinical record analysis.
Setting and participants: the analysed sample was composed of 4,271 cases identified between June 2013 and December 2014 along with 3,993 SDO without any code of MC (5% of the total volume of births in the same period).
Main outcome measures: positive predictive value (VPP) and negative predictive value (VPN) were computed by means of the comparison between the algorithm outcomes and the clinical record verification.
Results: 4,271 potentially malformed records involving 3,381 subjects born in the Sicilian territory have been identified. Among the hospital discharge records that it was possible to verify, the application of the algorithm led to the exclusion of 924 cases: of these, 62 proved to be false negatives (VPN: 93.3). The valid cases were 1,179, while the cases to be evaluated 617: the comparison between algorithm and clinical record analysis led to a VPP of 91.7 and 72.1, respectively, for valid and to be evaluated.
Conclusions: the tested algorithm proved to be a useful tool for identifying SDO potentially related to congenital anomalies. In the overall sample, the algorithm provided an outcome consistent with the clinical record assessment in 87.4% (2,379) of cases.
Keywords: Algorithm; Congenital anomalies; Congenital malformation register; Healthcare data flows; Hospital discharge records.