Accuracy of a Statistical Risk Assessment Model in Predicting Shoulder Dystocia and Brachial Plexus Injury in a High-Risk Population at Einstein Medical Center, Philadelphia [133]

INTRODUCTION: Shoulder dystocia at time of delivery can cause brachial plexus injury in the newborn. Obstetricians have tried to avoid brachial plexus injury by delivering patients at risk through cesarean delivery. A computerized statistical model, the Pericalm Shoulder Screen, has been used to predict these complications and counsel patients on mode of delivery. METHODS: Our primary aim was to determine whether the Shoulder Screen can accurately predict patients at higher risk of having shoulder dystocia with and without brachial plexus injury in a cohort of patients with known diagnoses of shoulder dystocia and brachial plexus injury. We performed a retrospective chart review of women and neonates with these International Classification of Diseases, 9th Revision diagnosis codes from June 1, 2010, to April 30, 2014. The Shoulder Screen was performed twice, using the estimated fetal weight on admission and the neonatal birth weight. A survey questionnaire was sent to mothers of neonates with brachial plexus injury to determine if injury persisted greater than 12 months. RESULTS: One hundred ninety five pairs charts were included on three groups: shoulder dystocia only (n=157), shoulder dystocia and brachial plexus injury (n=30), and brachial plexus injury without shoulder dystocia noted on neonate (n=8). The shoulder screen using estimated fetal weight predicted 10.5 % of brachial plexus injuries and 5.1 % of shoulder dystocia cases. Using the neonatal birth weight, the screen predicted 31.5% of brachial plexus injuries and 23.5% of shoulder dystocia cases. Average estimated fetal weight (3345.04 grams) was significantly lower than average neonatal birth weight (3740.18 grams). Prior vaginal delivery, prior shoulder dystocia, maternal diabetes, and gestational age did not significantly differ amongst groups. Maternal weight gain was significantly higher in the group with shoulder dystocia and brachial plexus injury. Using the neonatal birth weight, 18.4% more patients could have had a cesarean section. Brachial plexus injuries were not associated with shoulder dystocia 26.6% of the time. Only one brachial plexus injury was identified as permanent; the shoulder screen using estimated fetal weight and neonatal weight were low risk for this patient. CONCLUSION: The percentage of patients that the shoulder screen predicts as being at higher risk of having a shoulder dystocia with and without a brachial plexus injury was improved when the neonatal fetal weight was used. However, increasing the number of patients delivered by cesarean would not have prevented permanent brachial plexus injuries. The Pericalm Shoulder Screen should not be used to offer patients prophylactic cesarean deliveries, as shoulder dystocia and brachial plexus injury remain mostly unpredictable in nature.