W J Hueston1. 1. Eau Claire Family Practice Residency Program, University of Wisconsin.
Abstract
OBJECTIVE: To explore the potential use of a risk score based on clinical risk factors related to cesarean delivery. METHODS: From a multivariate analysis of a large data set, clinical risk factors associated with nonelective cesarean delivery were used to develop a risk scoring system based on a summation of logistic coefficients corresponding to individual clinical risk factors. After development and evaluation of the system at one site with 1523 women, the model was back-tested for external validity against a total of 5994 deliveries at four unrelated sites. RESULTS: The scoring system showed a clear cutoff point of maximal sensitivity and specificity that corresponded to an overall value of 3, which appeared to be consistent among the sites studied. Using this cutoff value, cesarean delivery could be predicted with a sensitivity of 81% and a specificity of 57%. However, the positive predictive values based on this rating system were low. Negative predictive values were much higher, indicating that the scoring system was more useful in identifying women for whom cesarean delivery was unlikely. CONCLUSION: A simple risk scoring system for cesarean delivery based on individual clinical risk factors does not appear to be useful in predicting women at high risk for cesarean delivery. The cesarean risk score could have some value in identifying women at low risk for cesarean and for evaluating the use of resources for large populations of patients.
OBJECTIVE: To explore the potential use of a risk score based on clinical risk factors related to cesarean delivery. METHODS: From a multivariate analysis of a large data set, clinical risk factors associated with nonelective cesarean delivery were used to develop a risk scoring system based on a summation of logistic coefficients corresponding to individual clinical risk factors. After development and evaluation of the system at one site with 1523 women, the model was back-tested for external validity against a total of 5994 deliveries at four unrelated sites. RESULTS: The scoring system showed a clear cutoff point of maximal sensitivity and specificity that corresponded to an overall value of 3, which appeared to be consistent among the sites studied. Using this cutoff value, cesarean delivery could be predicted with a sensitivity of 81% and a specificity of 57%. However, the positive predictive values based on this rating system were low. Negative predictive values were much higher, indicating that the scoring system was more useful in identifying women for whom cesarean delivery was unlikely. CONCLUSION: A simple risk scoring system for cesarean delivery based on individual clinical risk factors does not appear to be useful in predicting women at high risk for cesarean delivery. The cesarean risk score could have some value in identifying women at low risk for cesarean and for evaluating the use of resources for large populations of patients.
Authors: James M Nicholson; Aaron B Caughey; Morghan H Stenson; Peter Cronholm; Lisa Kellar; Ian Bennett; Katie Margo; Joseph Stratton Journal: Am J Obstet Gynecol Date: 2009-03 Impact factor: 8.661