Classification trees: A possible method for maternity risk grouping

Pregnancy, although being one of the most natural processes in our evolution, still remains subject to numerous complications and potential high risk. Complications at birth, such as the need for a caesarean section or the use of forceps, are not uncommon. An early warning of possible complications would greatly benefit both medical professionals and the expectant mother. Classification tree analysis uses selected independent variables to group pregnant women according to a dependent variable in a way that reduces variation. In this study, data on 3902 births were analysed to create risk groups for a number of complications, including the risk of a non-spontaneous delivery (a complicated birth) and premature delivery. From an overall risk of 23% of a non-spontaneous delivery, the classification tree was able to find statistically significant risk groups ranging from 7% to 65%. The resulting classification rules have been incorporated into a developed database tool to help quantify associated risks and act as an early warning system of possible complications to individual pregnant women.