Preterm labor and birth continue to pose a significant challenge to physicians in the obstetrics and neonatal fields. Until specific and effective therapeutic treatments are developed to prevent preterm labor, the best means of reducing preterm birth rate is early detection and diagnosis. However, current approaches to predict preterm labor have had variable success in the clinical setting. In this review, we discuss several limitations of using biomarkers from biological samples to predict preterm labor. In addition, we propose strategies for improving our ability to predict preterm labor, as well as directing therapies that are best suited to the underlying cause of preterm labor.