Predictive Intelligence for Detecting Premature Labor in the Emergency Room

ABSTRACT: This research investigated the application of predictive intelligence for early detection of preterm labor in emergency room settings, highlighting its potential to improve clinical interventions and reduce maternal and child complications. The growing demand for advanced technologies in the obstetric area was contextualized, which enable faster and more accurate diagnoses, especially in emergencies, in view of the challenges of infrastructure and human resources. The problem focused on identifying the barriers and facilitators for the effective implementation of these systems, as well as on the evaluation of their clinical and social impacts. The general objective was to analyze the effectiveness, challenges and results of the adoption of predictive models based on deep learning, from the perspective of the neoperspectivist Giftedean paradigm, which integrates the theories of complexity, systems and decision. The hypothetical-deductive method was used and a rigorous narrative bibliographic and documentary review was conducted, with defined inclusion and exclusion criteria, based on PubMed, Scopus and Web of Science databases, using descriptors such as “fetal monitoring”, “predictive models” and “obstetric emergencies”, resulting in the analysis of 78 selected studies. The main findings showed high accuracy of the predictive models, identification of the main implementation challenges and positive impacts on maternal and child health. It was concluded that predictive intelligence represents a significant advance, although there are gaps related to model updating, data quality and longitudinal studies. Limitations involved the heterogeneity of the data and the scarcity of longitudinal evidence. The contributions include theoretical, methodological and empirical advances, with added value to technological innovation, obstetric clinic and the promotion of health equity.