Prediction of preterm birth: past, present, and future approaches to an ongoing challenge

Abstract

Preterm birth is the leading cause of neonatal morbidity and mortality worldwide and contributes to substantial long-term health and economic burdens. Despite decades of research, overall rates remain largely unchanged, highlighting the urgent need for more effective predictive and preventative strategies. Traditional approaches, including risk-factor scoring, cervical length measurement, and fetal fibronectin testing, provide limited predictive value.

Given the limitations of traditional approaches, emerging technologies provide new opportunities to elucidate mechanisms and improve prediction. Proteomic analyses have identified pathways, such as inflammation and angiogenesis, while metabolomics has revealed small-molecule alterations reflecting biochemical and microbial processes. Genetic investigations highlight complex contributions from both maternal and fetal genomes. Artificial intelligence and machine learning are being applied to integrate multi-omics data with clinical variables with early studies suggesting improved predictive accuracy compared with conventional models.

Despite these advancements, significant challenges remain. Many prediction studies are constrained by heterogeneous definitions of preterm birth, small sample sizes, and lack of validation across diverse populations. Beyond research limitations, system-level factors such as social determinants of health, environmental exposure, and inequities in access to prenatal care contribute to disparities in both risk and outcomes. These realities underscore the need for predictive tools that are not only scientifically robust but also applicable across diverse populations and care settings. Although clinical translation of novel approaches remains limited, continued innovation, longitudinal research, and commitment to equity will be essential to achieving meaningful improvements in the prediction of preterm birth. Understanding the pathophysiology and applying proven interventions is essential to improving perinatal health outcomes.