Clinical Evaluations Have Low Sensitivity for Identifying Preterm Infants in a Clinical Trial in a Limited Resource Setting

Abstract

Preterm birth is a primary outcome of interest in maternal vaccination trials but determination of gestational age is challenging in limited-resource settings. This study compares the New Ballard Score and fundal height measurements with the current standard of early ultrasound for sensitivity of predicting preterm birth. A trial of maternal influenza vaccination was conducted in Bamako, Mali. The New Ballard Score and fundal height were collected on 4038 infants born in the trial, ultrasound data were available for 1893 of those infants. New Ballard Score and fundal height were compared, consecutively, to all ultrasound results, early ultrasound results from the first trimester, and the date of last menstrual period for estimation of gestational age. Sensitivity of the New Ballard Score for identifying preterm infants was 0.33 compared with early ultrasound and 0.1 compared with the last menstrual period based estimates of gestational age. Sensitivity of low birth weight alone was 0.43 compared with early ultrasound. New Ballard Score estimated gestational age within 1 week of ultrasound more frequently than fundal height (53% compared with 7.6%, respectively) yet New Ballard Score identified few infants as preterm (1.8% vs 5.8% by early ultrasound), and was biased toward categorizing low birth weight infants and infants requiring hospitalization as preterm. New Ballard Score is not an ideal measure for identifying preterm births in low-resource settings. Despite the time and cost of training required for correct measurement of New Ballard Score, measurement of low birth weight alone performed better than New Ballard Score for identifying preterm infants.