Presentation at the Global Maternal Newborn Health Conference, October 19, 2015
Background: Most preterm births and deaths occur in homes or first-level facilities in low income countries, where gestational age is uncertain. Early identification of premature infants in these settings may improve care-seeking and delivery of interventions to prevent neonatal morbidity and mortality. The objective of this study was to determine the accuracy of different signs of newborn maturity assessed by frontline community health workers (CHWs) to identify the preterm infant.
Methodology: We conducted a validation study in a prospective birth cohort in rural Sylhet, Bangladesh. Early pregnancy ultrasonography was performed for gold standard gestational age dating, and newborns were visited at home by a CHW at <72 hours of life. Neonatal assessment included physical/neuromuscular signs and anthropometrics. Sensitivity and specificity of individual signs and different algorithms were calculated.
Results: In the live birth cohort (n=727), the mean gold standard gestational age was 38.7(±2.7) weeks and birthweight was 2786g (± 427). The prevalence of preterm birth was 14.0%, low birth weight (LBW<2500g) 16.2%, birth weight <2000g 2.5%. By Ballard scoring, the mean gestational age was 39.4(± 2.1) weeks and preterm birth rate 11.8%. However the sensitivity of the Ballard for detecting preterm births was low (17%). A simple 5 sign assessment (ear recoil, foot creases, breast bud, skin opacity, and foot length) had sensitivity of 75% but low specificity (28%). Foot length <75mm had fair sensitivity for identifying preterm, LBW (75%), and <2000g (83%) babies but lower specificity for all conditions (35-40%). However, the identification of babies <2000g was reasonable with a foot length <74mm (sensitivity 82%, specificity 44%) and head circumference <32.5 cm (sensitivity 89%, specificity 68%).
Conclusions: Clinical neonatal assessment of gestational age is challenging in low-resource settings with high rates of fetal growth restriction. Foot length and head circumference may reasonably identify the high-risk “<2000 gm infants” in settings where birth weight is unavailable.