Gianni Astolfi1, Paolo Ricci2, Elisa Calzolari3, Amanda Neville3, Vanda Pironi2, Michele Santoro4, Fabrizio Bianchi4. 1. Registro indagini sulle malformazioni congenite in Emilia-Romagna (IMER), Dipartimento di scienze biomediche e chirurgico-specialistiche, Università di Ferrara. 2. Registro delle malformazioni congenite, Osservatorio epidemiologico, ASL Mantova. 3. Registro indagini sulle malformazioni congenite in Emilia-Romagna (IMER), Azienda ospedaliera di Ferrara. 4. Unità di epidemiologia ambientale e registri di patologia, Istituto di fisiologia clinica, Consiglio nazionale delle ricerche, Pisa.
Abstract
OBJECTIVES: to evaluate and validate the use of an algorithm designed to identify in hospital discharge records (SDO) cases with congenital malformations (MC) at birth and/or reported in hospitalizations within the first year of life using as gold standard the Congenital malformation Registry of the Local Health Unit of Mantova, Northern Italy, (RMC-MN), which controls all the medical records of infants born to mothers living in the province. DESIGN: an algorithm designed for the identification of malformed cases in the SDO database using two modules, one for identification of cases potentially malformed and one for their validation was used. A comparison of the results with those observed by the RMC-MN was then conducted. SETTING AND PARTICIPANTS: data of the SDO and the RMC-MN for the period 2010-2011 relative to those detected in newborns within the first year of life in the resident population in the province. RESULTS: of 8,042 infants born to mothers residing in the province of Mantova, 7,367 were excluded by the algorithm as malformed with the exception of only one false negative (negative predictive value - NPV: 99.99%); in the remaining 675 cases (8.4%) there was at least one code of congenital malformation. The algorithm has also included 396 cases (4.9%) with isolated minor malformations or diseases considered not malformations, of which 23 were false negatives (NPV: 94.2%). In the remaining 279 cases potentially malformed the algorithm considered as validated 169 cases (60.6%), including 11 false positives (positive predictive value - PPV: 93.5%). In the remaining 110 cases to evaluate, 46 were true positives (PPV: 41.8%). CONCLUSIONS: the proposed instrument has identified correctly SDO in 89.4% of cases registered by the RMC-MN to produce a small number of false positives among the validated cases (6.5%) and effectively exclude inappropriate cases (94.2%). The authors suggest a judicious use of the instrument, which should be led by experts of SDO, clinical and epidemiology of congenital malformations.
OBJECTIVES: to evaluate and validate the use of an algorithm designed to identify in hospital discharge records (SDO) cases with congenital malformations (MC) at birth and/or reported in hospitalizations within the first year of life using as gold standard the Congenital malformation Registry of the Local Health Unit of Mantova, Northern Italy, (RMC-MN), which controls all the medical records of infants born to mothers living in the province. DESIGN: an algorithm designed for the identification of malformed cases in the SDO database using two modules, one for identification of cases potentially malformed and one for their validation was used. A comparison of the results with those observed by the RMC-MN was then conducted. SETTING AND PARTICIPANTS: data of the SDO and the RMC-MN for the period 2010-2011 relative to those detected in newborns within the first year of life in the resident population in the province. RESULTS: of 8,042 infants born to mothers residing in the province of Mantova, 7,367 were excluded by the algorithm as malformed with the exception of only one false negative (negative predictive value - NPV: 99.99%); in the remaining 675 cases (8.4%) there was at least one code of congenital malformation. The algorithm has also included 396 cases (4.9%) with isolated minor malformations or diseases considered not malformations, of which 23 were false negatives (NPV: 94.2%). In the remaining 279 cases potentially malformed the algorithm considered as validated 169 cases (60.6%), including 11 false positives (positive predictive value - PPV: 93.5%). In the remaining 110 cases to evaluate, 46 were true positives (PPV: 41.8%). CONCLUSIONS: the proposed instrument has identified correctly SDO in 89.4% of cases registered by the RMC-MN to produce a small number of false positives among the validated cases (6.5%) and effectively exclude inappropriate cases (94.2%). The authors suggest a judicious use of the instrument, which should be led by experts of SDO, clinical and epidemiology of congenital malformations.
Authors: Michele Santoro; Fabrizio Minichilli; Anna Pierini; Gianni Astolfi; Lucia Bisceglia; Pietro Carbone; Susanna Conti; Gabriella Dardanoni; Ivano Iavarone; Paolo Ricci; Gioacchino Scarano; Fabrizio Bianchi Journal: Int J Environ Res Public Health Date: 2017-03-10 Impact factor: 3.390