Understanding birthing mode decision making using artificial neural networks

Background. This study examined obstetricians' decisions to perform or not to perform cesarean sections. The aim was to determine whether an artificia l neural network could be constructed to accurately and reliably predict th e birthing mode decisions of expert clinicians and to elucidate which facto rs were most important in deciding the birth mode. Methods. Mothers with si ngleton, live births who were privately insured, nonclinic, non-Medicaid pa tients at a major tertiary care private hospital were included in the study (N = 1508). These mothers were patients of 2 physician groups: a 7-obstetr ician multispecialty group practice and a physician group of 79 independent ly practicing obstetricians affiliated with the some hospital. A feedforwar d, multilayer artificial neural network (ANN) was developed and trained. It was then tested and optimized until the most parsimonious network was iden tified that retained a similar level of predictive power and classification accuracy. The performance of this network was further optimized using the methods of receiver operating characteristic (ROC) analysis and information theory to find the cutoff that maximized the information gain. The perform ance-of the final ANN at this cutoff was measured using sensitivity, specif icity, classification accuracy, area under the ROC curve, and maximum infor mation gain. Results. The final neural network had excellent predictive acc uracy for the birthing mode (classification accuracy = 83.5%; area under th e ROC curve = 0.924; maximum information = 40.4% of a perfect diagnostic te st), Conclusion. This study demonstrated that a properly optimized ANN is a ble to accurately predict the birthing mode decisions of expert clinicians. In addition to previously identified clinical factors (cephalopelvic dispr oportion, maternal medical condition necessitating a cesarean section, arre st of labor, malpresentation of the baby, fetal distress, and failed induct ion), nonclinical factors such as the mothers' views on birthing mode were also found to be important in determining the birthing mode.