Interobserver variation in the classification of diseases

: The aim of this study was to describe the variation within the observer in classification of diseases, with special regard to pregnancy, birth and maternity. Four observers A, B, C and D classified 150 random cases of pregnancy, birth and maternity on two different occasions with six months interval. The rate of agreement within the observer between the two different occasions was evaluated. The average agreement rate was 86% (71-92%) concerning diagnoses of pregnancy. 90% (84-95%) concerning maternity and 65% (53-75%) concerning delivery. The diagnostic spectrum of the observers and their criteria for the use of diagnoses were variable. The amount of variation correlated to the number of diagnoses and poorly defined diagnostic criteria. In conclusion, great variation was found within the observer in classifying diseases, especially concerning delivery, due to increased numbers of diagnoses and poorly defined diagnostic criteria. The setting-up of branch specialist committees is recommended, if we in the future want to use a public diagnostic register as a tool for research. These committees must achieve consensus concerning the need for classification within each specialty, define criteria for diagnosis and communicate recommendations and education in the system of classification.