Prediction Models for Determining the Success of Labor Induction: A Systematic Review [35C]

INTRODUCTION: The purpose of this study was to systematically identify all derived and/or validated clinical prediction models for labor induction containing universally accessible factors and compare their performance to inform the development of a model that could be recommended for clinical practice. METHODS: Four databases were searched from inception to November, 2017 for studies that derived and/or validated clinical prediction models containing antenatal history and cervical examination. Risk-of-bias of included studies was assessed using the Prediction Study Risk of Bias Assessment Tool. In view of anticipated heterogeneity between studies, only descriptive analysis was possible. RESULTS: We identified 16 studies describing 14 prediction models derived between 1966 and 2018. Models varied with regard to participant inclusion, sample size, considered and included variables, endpoint definitions, study design and performance. Of the derived models, six were validated internally and three externally. Performance was most commonly measured using the area under the receiver operator characteristic curve, which ranged from 0.68 to 0.79, 0.67 to 0.77 and 0.61 to 0.73 for derived, internally validated and externally validated models, respectively. Studies' risk-of-bias ranged between studies fulfilling 36% to 86% of eligible items. CONCLUSION: No published model to determine the success of vaginal birth after labor induction can be currently recommended for clinical use. In order to improve performance and uptake, researchers should incorporate attitudes of women and care providers, assess clinical and resource implications and adhere to recommendations made by this systematic review before deriving and validating prediction models for determining the success of labor induction.