Weighted risk scoring system for predicting peripartum transfusion: development and internal validation

Background Early identification of obstetric patients at risk for transfusion can improve perioperative planning and resource allocation. This study aimed to develop and internally validate a weighted scoring system using routinely available pre-delivery parameters. Methods We retrospectively analyzed deliveries in a tertiary center between [June 2020–June 2025]. Three continuous variables—platelet count, hemoglobin, and prothrombin time—were dichotomized using receiver operating curve characteristic (ROC)–derived thresholds. These, along with placenta previa, emergency cesarean delivery, and known coagulopathy, entered multivariable logistic regression. Model coefficients determined the Weighted Transfusion Risk Score (0–10). Discrimination was assessed by area under the ROC curve (AUC) in derivation (70%) and validation (30%) cohorts, with bootstrap resampling. Results Among 249 women, 69 (27.7%) required transfusion. Six variables remained significant predictors: placenta previa, emergency cesarean delivery, prolonged PT, low hemoglobin, low platelet count, and coagulopathy. The weighted transfusion risk score achieved an AUC of 0.838 (95% CI 0.784–0.892) overall, with 92.8% sensitivity and 57.2% specificity at the optimal cut-off. Validation yielded similar discrimination (AUC 0.871). Conclusion A weighted score based on common clinical and laboratory data effectively stratified transfusion risk and retained performance on internal validation. Incorporation into standardized hemorrhage protocols may enable earlier recognition of high-risk patients, more efficient mobilization of resources, and targeted preventive measures. External validation is needed to confirm applicability.