THE QUALITY OF DATA IN THE HESSIAN PERINA TAL REGISTRY - COMPARISON WITH ANTENATAL RECORDS AND RESULTS OF INTERVIEWS WITH PREGNANT-WOMEN

In the Federal State of Hesse, Germany, pregnancy and delivery related data are collected in a central perinatal registry called ''Hessische Perinatalerhebung'' (HEPE) since 1981, the objective being to enforce quality assurance and to contribute to a reduction in perinatal morta lity and morbidity. The delivery unit compiles these data based on the antenatal record called ''Mother Passport'' and the actual delivery. In order to check on the quality of data, information on risk factors obtained from the perinatal register was compared to information from antenatal records and from a questionnaire administered to pregnant wo men prior to delivery. The study population comprised 1978 women who d elivered at the maternity unit of the Municipal Hospital of Wiesbaden, the Dr. Horst Schmidt Kliniken, in 1990 and 1991. Comparison of the t hree data sources revealed substantial differences: Almost all risk fa ctors were most frequently mentioned by the mothers themselves, less b y the antenatal record and even less by the perinatal register. These differences were less pronounced for data on previous pregnancies and pre-existing disease and on abnormal positioning. Considerable differe nces were found in relation to pregnancy induced hypertension, intra-u terine growth retardation, incapacity of the cervix, preterm labour an d bleeding. Risk factors related to the social environment and persona l habits like psyche-social stress and smoking were found to be extrem ely under-reported in the antenatal records as well as in the perinata l register. The following factors could be identified as major reasons for the different risk profiles obtained from the various data source s: [1] Data loss through the transfer of data from the antenatal recor d to the record form of the perinatal register or to the computer, [2] selective recording of those risk factors considered by doctors to be particularly relevant, [3] missing definition of some risk factors an d [4] recording of antenatal data for the perinatal register by the de livery unit after knowing the pregnancy outcome. Antenatal data in the perinatal register of Hessen are doubtful and should not be used uncr itically for comparative studies and the assessment of relative risks. The extreme under-reporting of some risk factors may even lead to giv ing some interventions undue priority. In future, regular validity stu dies should become routine. A reduction of the data losses as well as a clear separation of antenatal data and delivery data could be achiev ed if antenatal data would be transferred from the doctor providing an tenatal care direct to the perinatal register.