Speaker
Description
Background: Necrotizing Enterocolitis (NEC) is a serious gastrointestinal disease primarily affecting preterm neonates. Despite improvements in neonatal care, NEC continues to contribute significantly to neonatal mortality, particularly in low-resource settings. In Ethiopia, reported NEC-related mortality rates vary widely, from 45% to 89%, reflecting both the severity of the disease and inconsistencies in existing evidence. Moreover, little is known about the predictors of NEC-related mortality in the local context. This underscores the need for a locally derived predictive model to support early risk stratification and guide clinical decision-making, with the ultimate goal of improving survival outcomes among affected neonates.
Methods: A prospective cohort study was conducted among 251 neonates hospitalized with Necrotizing Enterocolitis at Felege-Hiwot Comprehensive Specialized Hospital (FHCSH) and Tibebe Ghion comprehensive Specialized Hospital (TGSH) . Data were analyized using R version 4.4.2 software. A multivariable analysis was performed to identify predictors of mortality, and a simplified nomogram was developed to enhance clinical applicability. Model performance was evaluated using the area under the receiver operating characteristic curve (AUC) and calibration plot. Bootstrapping was used to validate all accuracy measures. A decision curve analysis was used to evaluate the clinical and public health utility of our model.
Results: NEC mortality rate was 51% (95% CI: 45.00-57.34). outborn delivery, lower gestational age, disease onset ≤3 days, delayed first feeding beyond 48 hours of postnatal age, abdominal wall erythema, stage III NEC, severe thrombocytopenia, clinical deterioration within 48 hours of diagnosis, and hospital-acquired infection were Key predictors remained in the reduced model. The AUC of the original model was 0.965 (95% CI: 0.943, 0.982), whereas the nomogram model produced prediction accuracy of an AUC of 0.959 (95% CI: 0.942, 0.982). Our decision curve analysis for the model provides a higher net benefit across ranges of threshold probabilities.
Conclusions: Our model has excellent discrimination and calibration performance. Similarly, the nomogram model has excellent discrimination and calibration ability with an insignificant loss of accuracy from the original. The models can have the potential to improve care and treatment outcomes in the clinical settings.
