Factors affecting antenatal corticosteroid use in low- and middle-income countries: Facility characteristics, structural readiness, and past performance of CEmONC signal functions

Antenatal corticosteroids (ACS) utilization is disproportionately limited in low- and middle-income countries where most global preterm newborns who could benefit from this intervention are born. Understanding the factors affecting ACS use is crucial for improving its uptake. This study aimed to investigate facility-level factors associated with ACS use in low-resource countries. We used data from ten Service Provision Assessment surveys across nine countries. We restricted the sample to facilities that provided delivery services. Our primary outcome was recent ACS use, defined as having administered ACS within the past three months before the survey. We conducted mixed-effect log binomial regressions, with country as a fixed effect and sub-national regions as random intercepts, to explore the association between recent ACS use and facility characteristics, injectable corticosteroids and ultrasound availability, facility structural readiness, and past performance of nine Comprehensive Emergency Obstetric and Newborn Care (CEmONC) signal functions. A sensitivity analysis excluding the Afghanistan data was performed since the country adopted a different sampling strategy that mainly sampled urban facilities. This study included 6183 facilities from nine countries. Across eight countries with nationally representative data, only 22.7% (median, range 4.0% to 27.4%) of facilities that provided delivery services had used ACS recently, compared to 59.4% in the Afghanistan survey. Urban facilities had a 21% higher likelihood of recent ACS use (95% CI 6%–38%) than rural facilities. Corticosteroid availability was associated with a 14% higher likelihood of recent ACS use (95% CI 1%–29%). Facilities in the highest readiness tertile were more likely to have recent ACS use than those in the lowest (RR 1.91, 95% CI 1.58–2.30). Each CEmONC signal function, except assisted vaginal deliveries, was significantly associated with recent ACS use, with neonatal resuscitation having the largest effect (RR 2.62, 95% CI 1.93–3.55). In conclusion, facilities that had performed CEmONC services were more likely to administer ACS, highlighting the importance of provider knowledge, skills, and competence in managing obstetric and newborn emergencies for effective ACS provision.