WP6 Deliverable D6.1: Report on the evaluation of specific congenital anomaly coding in health care databases including a computer algorithm to improve these codes
Authors: Dr Marian K Bakker, Dr Hermien EK de Walle, Dr Gianni Astolfi, Dr Amanda J Neville, Dr Cecilia Martellucci, Dr Elisa Ballardini, Prof Joan Morris
Contributing participants and partners (in Beneficiary numerical order): Joan Morris, Elizabeth Limb, Joachim Tan, Jo Brigden, Hugh Claridge (SGUL), Maria Loane, Joanne Given, Leke Aminkeng (UU), Ester Garne, Stine Kjær Urhøj (RSD), Judith Rankin (UNEW), Amanda Neville, Gianni Astolfi,Cecilia Martellucci, Elisa Ballardini (UNIFE), Ingeborg Barišic (KDB), Anna Pierini, Michele Santoro, Francesca Gorini, Alessio Coi, Silvia Baldacci (CNR-IFC), Marian Bakker, Renée Lutke, Nicole Siemensma Mühlenberg, Hermien de Walle (UMCG), David Tucker, Ieuan Scanlon, (PHW NHS, SU), Clara Cavero Carbonell, Oscar Zurriaga, Laia Barrachina-Bonet, Laura García-Villodre, Ana Ruiz (FISABIO), Sonja Kiuru-Kuhlefelt, Anna Heino, Mika Gissler (THL), Diana Wellesley (Wessex), Jennifer Kurinczuk (Thames Valley), Elizabeth Draper (East Midlands and South Yorkshire).
In this part of the EUROlinkCAT project we compared the diagnosis codes recorded in the hospital databases for CAs to the diagnoses registered in EUROCAT data. Eleven European congenital anomaly registries, which are part of the European Surveillance of Congenital Anomalies network (EUROCAT), successfully linked their records of children with congenital anomalies to regional and national electronic hospital databases. For seventeen specific CAs, we compared what proportion of children from the EUROCAT registries were registered in the hospital databases with the exact CA code (sensitivity). We also compared the proportion of children with a CA code recorded in the hospital data that were registered in EUROCAT data with the same CA code (positive predictive value). Our analyses showed that a high proportion of EUROCAT cases with cleft lip with or without cleft palate, Down syndrome and Hirschsprung’s disease were recorded in the hospital database with the exact
code for these anomalies, but that a low proportion of EUROCAT cases with clubfoot or congenital hydronephrosis were recorded in the hospital databases. We also found that a high proportion of children recorded in the hospital database with gastroschisis or Down syndrome were registered in the EUROCAT data with the same CA code. However, the majority of children for which an ASD code was recorded in the hospital data, were not registered in the EUROCAT data. Results differed per registry and hospital database. Hospital databases are limited in monitoring CAs as the quality of CA coding is variable and information is lacking in hospital databases on terminations of pregnancy, children with CAs that do not require hospitalisations and children with CAs that are treated in specialist centers outside the region of coverage. Limited information on related factors is available in EHCDs. However, we did work on improving an existing algorithm which aims to examine whole electronic hospital databases and identify children who do not have a CA, children who do have a CA and a small group of children whose records need to be examined by a clinician to determine if they have a CA. We provide the complete details of this algorithm to improve the use of hospital databases in the analysis of children with specific anomalies.
Click here for a PDF of the report