[An evaluation of the National Patient Register. A study of validity of some abortion diagnoses].

The object of this study was to describe the use of some specific diagnoses for abortion and to study the validity of data in the Danish National Hospital Discharge Register by comparison with data in the discharge records. The study is based on 359 discharge records from 31 hospitals in 1984. Accordance between the two data sources was 92-100% for administrative data (personal identification number, hospital identification, date of hospitalization, and ICD-diagnosis code). Accordance was poorer (31-54%) with the diagnosis in latin and the number code of the diagnosis in the discharge records. This is partly because most of the discharge records include a diagnosis which is difficult to use (ICD-code 644: abortion, not specified as induced or spontaneous) and partly because classification of diagnoses is given low priority in the hospital doctor's job. The article proposes teaching medical students how to use the ICD-classification of diagnoses and surgical procedures. Further, we suggest exclusion of the ICD-code 644 since it is always possible to distinguish a spontaneous from an induced abortion in Denmark. It is proposed that the ICD-code 644 is used for complications following spontaneous and induced abortions (retention, haemorrhage, endometritis etc.).