TY - JOUR
T1 - Simplified models to assess newborn gestational age in low-middle income countries
T2 - Findings from a multicountry, prospective cohort study
AU - Aftab, Fahad
AU - Ahmed, Parvez
AU - Ahmed, Salahuddin
AU - Ali, Said Mohammed
AU - Bahl, Rajiv
AU - Banda, Bowen
AU - Baqui, Abdullah H.
AU - Akonkwa, Corneille Bashagaluke
AU - Begum, Nazma
AU - Deb, Saikat
AU - Dhingra, Pratibha
AU - Dhingra, Usha
AU - Dutta, Arup
AU - Edmond, Karen
AU - Grogan, Caroline
AU - Hamer, Davidson H.
AU - Herlihy, Julie
AU - Hurt, Lisa
AU - Hussain, Atiya
AU - Ilyas, Muhammad
AU - Jehan, Fyezah
AU - Kapasa, Monica Lulu
AU - Karim, Muhammad
AU - Kausar, Farzana
AU - Kirkwood, Betty R.
AU - Lee, Anne C.C.
AU - Manu, Alexander
AU - Mehmood, Usma
AU - Mitra, Dipak
AU - Mohammed, Mohammed
AU - Mweene, Fern
AU - Nadeem, Naila
AU - Nisar, Muhammad Imran
AU - Paul, Rina
AU - Rahman, Mahmoodur
AU - Rahman, Sayedur
AU - Sajid, Muhammad
AU - Sazawal, Sunil
AU - Semrau, Katherine E.
AU - Shannon, Caitlin
AU - Straszak-Suri, Marina
AU - Suleiman, Atifa
AU - Uddin, Mohammad J.
AU - Wilbur, Jayson
AU - Wylie, Blair
AU - Yoshida, Sachiyo
N1 - Funding Information:
Competing interests The WHO and study sites received funding from the Bill and Melinda Gates Foundation (BMGF) to conduct this study. The statistician performing the machine learning analysis (JW) is an employee of Metrum Research Group which received funding from the Bill and Melinda Gates Foundation. BB, CBA, CG, DHH, JH, LH, MK, FM, KS, MSS and JW report research grants from the BMGF during the conduct of the study. ACL reported research grants from the NICHD and BMGF, and does consultancy to WHO.
Funding Information:
Funding The study was funded by the Bill and Melinda Gates Foundation.
Publisher Copyright:
© 2021 BMJ Publishing Group. All rights reserved.
PY - 2021/9/13
Y1 - 2021/9/13
N2 - Introduction Preterm birth is the leading cause of child mortality. This study aimed to develop and validate programmatically feasible and accurate approaches to estimate newborn gestational age (GA) in low resource settings. Methods The WHO Alliance for Maternal and Newborn Health Improvement (AMANHI) study recruited pregnant women from population-based cohorts in five countries (Bangladesh, Ghana, Pakistan, Tanzania and Zambia). Women <20 weeks gestation by ultrasound-based dating were enrolled. Research staff assessed newborns for: (1) anthropometry, (2) neuromuscular/physical signs and (3) feeding maturity. Machine-learning techniques were used to construct ensemble models. Diagnostic accuracy was assessed by areas under the receiver operating curve (AUC) and Bland-Altman analysis. Results 7428 liveborn infants were included (n=536 preterm, <37 weeks). The Ballard examination was biased compared with ultrasound dating (mean difference: +9 days) with 95% limits of agreement (LOA) -15.3 to 33.6 days (precision ±24.5 days). A model including 10 newborn characteristics (birth weight, head circumference, chest circumference, foot length, breast bud diameter, breast development, plantar creases, skin texture, ankle dorsiflexion and infant sex) estimated GA with no bias, 95% LOA ±17.3 days and an AUC=0.88 for classifying the preterm infant. A model that included last menstrual period (LMP) with the 10 characteristics had 95% LOA ±15.7 days and high diagnostic accuracy (AUC 0.91). An alternative simpler model including birth weight and LMP had 95% LOA of ±16.7 and an AUC of 0.88. Conclusion The best machine-learning model (10 neonatal characteristics and LMP) estimated GA within ±15.7 days of early ultrasound dating. Simpler models performed reasonably well with marginal increases in prediction error. These models hold promise for newborn GA estimation when ultrasound dating is unavailable.
AB - Introduction Preterm birth is the leading cause of child mortality. This study aimed to develop and validate programmatically feasible and accurate approaches to estimate newborn gestational age (GA) in low resource settings. Methods The WHO Alliance for Maternal and Newborn Health Improvement (AMANHI) study recruited pregnant women from population-based cohorts in five countries (Bangladesh, Ghana, Pakistan, Tanzania and Zambia). Women <20 weeks gestation by ultrasound-based dating were enrolled. Research staff assessed newborns for: (1) anthropometry, (2) neuromuscular/physical signs and (3) feeding maturity. Machine-learning techniques were used to construct ensemble models. Diagnostic accuracy was assessed by areas under the receiver operating curve (AUC) and Bland-Altman analysis. Results 7428 liveborn infants were included (n=536 preterm, <37 weeks). The Ballard examination was biased compared with ultrasound dating (mean difference: +9 days) with 95% limits of agreement (LOA) -15.3 to 33.6 days (precision ±24.5 days). A model including 10 newborn characteristics (birth weight, head circumference, chest circumference, foot length, breast bud diameter, breast development, plantar creases, skin texture, ankle dorsiflexion and infant sex) estimated GA with no bias, 95% LOA ±17.3 days and an AUC=0.88 for classifying the preterm infant. A model that included last menstrual period (LMP) with the 10 characteristics had 95% LOA ±15.7 days and high diagnostic accuracy (AUC 0.91). An alternative simpler model including birth weight and LMP had 95% LOA of ±16.7 and an AUC of 0.88. Conclusion The best machine-learning model (10 neonatal characteristics and LMP) estimated GA within ±15.7 days of early ultrasound dating. Simpler models performed reasonably well with marginal increases in prediction error. These models hold promise for newborn GA estimation when ultrasound dating is unavailable.
KW - child health
KW - obstetrics
UR - http://www.scopus.com/inward/record.url?scp=85115090202&partnerID=8YFLogxK
U2 - 10.1136/bmjgh-2021-005688
DO - 10.1136/bmjgh-2021-005688
M3 - Article
AN - SCOPUS:85115090202
SN - 2059-7908
VL - 6
JO - BMJ Global Health
JF - BMJ Global Health
IS - 9
M1 - e005688
ER -
