Clinical Epidemiology (Dec 2015)
Validation of second trimester miscarriages and spontaneous deliveries
Abstract
Kirstine Sneider,1 Jens Langhoff-Roos,2 Iben Blaabjerg Sundtoft,3 Ole Bjarne Christiansen4,5 1Department of Clinical Research, Vendsyssel Hospital, Hjørring, 2Department of Obstetrics, Rigshospitalet, University of Copenhagen, Copenhagen, 3Department of Obstetrics and Gynecology, Aarhus University Hospital, Aarhus, 4Department of Obstetrics and Gynecology, Aalborg University Hospital, Aalborg, 5The Fertility Clinic, University Hospital, Copenhagen, Denmark Objective: To validate the diagnosis of second trimester miscarriages/deliveries (16+0 weeks to 27+6 weeks of gestation) recorded as miscarriages in the Danish National Patient Registry or spontaneous deliveries in the Danish Medical Birth Registry, and asses the validity of risk factors, pregnancy complications, and cerclage by review of medical records. Materials and methods: In a cohort of 2,358 women with a second trimester miscarriage/delivery in first pregnancy and a subsequent delivery during 1997–2012, we reviewed a representative sample of 682 medical records. We searched for clinically important information and calculated positive predictive values of the registry diagnoses stratified by type of registry, as well as sensitivity, specificity, positive predictive value, and kappa coefficients of risk factors, pregnancy complications, and cerclage. Results: Miscarriage/spontaneous delivery in the second trimester was confirmed in 621/682 patients (91.1%). Pregnancy complications in second trimester miscarriages were underreported, resulting in low sensitivities and poor to moderate agreements between records and registries. There was a good agreement (kappa >0.6) between medical records and the registries regarding risk factors and cerclage. The diagnosis of cervical insufficiency had “moderate” kappa values for both miscarriages and deliveries (0.55 and 0.57). Conclusion: Spontaneous second trimester deliveries and miscarriages recorded in the registers were confirmed by medical records in 91%, but register-based information on pregnancy complications need to be improved. We recommend that all pregnancies ending spontaneously beyond the first trimester are included in the national birth registry and described by appropriate variables. Keywords: second trimester, miscarriage, preterm birth, prematurity, registries, reproducibility of results, cerclage, cervical insufficiency
