Abstract:
Background: Globally, an estimated 15 million preterm deliveries are observed annually. Of these, more than two-thirds occurred in Africa and South Asia. Prematurity is the leading cause of neonatal and under-five morbidity and mortality, specifically in low-resource settings, including Ethiopia. Studies showed that the majority of prematurity can be prevented if early interventions are implemented for high-risk pregnancies. Therefore, a prognosis risk score for preterm birth can be developed based on easily available predictors and used by health professionals as a simple clinical tool in their decision-making. Objective: To develop and validate a prognosis risk score for preterm birth among pregnant women who had antenatal care visit at Debre Markos Comprehensive and Specialized Hospital, Ethiopia. Methods: A retrospective follow-up study was conducted from January 1, 2020- August 30, 2022 and a total of 1,132 pregnant women were included in this study. Client charts were selected using a simple random sampling technique. Data were extracted using structured checklist prepared in the Kobo Toolbox application and exported to STATA version 14 and R version 4.2.2 softwares for data management and further analysis. Stepwise backward multivariable analysis was done. A simplified risk prediction model (risk score) was developed based on a binary logistic model, and the model's performance was assessed by discrimination power and calibration. The internal validity of the model was evaluated by bootstrapping, which replicates the sample 2,000 times. Decision Curve Analysis (DCA) was used to determine the clinical impact of the risk score. Result: In this study, the incidence of preterm birth was 10.9% (95% CI: 9.2%, 12.8%). The developed risk score included six predictors that remained in the reduced multivariable logistic regression, including age<20, late initiation of antenatal care, unplanned pregnancy, recent pregnancy complications, hemoglobin<11 mg/dl, and multiparty, for a total score of 17. The discriminatory power of the model was 0.931 (95% CI: 90.3%, 95.5%), and the calibration test was p > 0.05. The optimal cut-off for classifying risks as low or high was 4. At this cut point, the sensitivity is 91.0%, the specificity is 82.1%, and the accuracy is 83.1%. It was internally validated using the bootstrapping method and has an optimism of 0.003. The decision curve analysis revealed the model's net benefit is greater than the treating all or none approach. Conclusion and recommendation: The developed risk-score has excellent discrimination performance and clinical benefit. Hence, it can be applied in clinical practice. We recommend that the researcher externally validate the model and for clinicians use a prognosis risk score
