Measuring coverage and quality of supportive care for inpatient neonatal infections : EN-BIRTH multi-country validation study

Background: An estimated 7 million episodes of severe newborn infections occur annually worldwide, with half a million newborn deaths, most oc-curring in low-and middle-income countries. Whilst injectable antibiotics are necessary to treat the infection, supportive care is also crucial in ending preventable mortality and morbidity. This study uses multi-country data to assess gaps in coverage, quality, and documentation of supportive care, considering implications for measurement. Methods: The EN-BIRTH study was conducted in five hospitals in Bangla-desh, Nepal, and Tanzania (July 2017-July 2018). Newborns with an ad-mission diagnosis of clinically-defined infection (sepsis, meningitis, and/or pneumonia) were included. Researchers extracted data from inpatient case notes and interviews with women (usually the mothers) as the primary fam-ily caretakers after discharge. The interviews were conducted using a struc-tured survey questionnaire. We used descriptive statistics to report coverage of newborn supportive care components such as oxygen use, photothera-py, and appropriate feeding, and we assessed the validity of measurement through survey-reports using a random-effects model to generate pooled estimates. In this study, key supportive care components were assessment and correction of hypoxaemia, hyperbilirubinemia, and hypoglycaemia. Results: Among 1015 neonates who met the inclusion criteria, 89% had an admission clinical diagnosis of sepsis. Major gaps in documentation and care practices related to supportive care varied substantially across the par-ticipating hospitals. The pooled sensitivity was low for the survey-reported oxygen use (47%; 95% confidence interval (CI) = 30%-64%) and moderate for phototherapy (60%; 95% CI = 44%-75%). The pooled specificity was high for both the survey-reported oxygen use (85%; 95% CI = 80%-89%) and phototherapy (91%; 95% CI = 82%-97%). Conclusions: The women's reports during the exit survey consistently un-derestimated the coverage of supportive care components for managing infection. We have observed high variability in the inpatient documents across facilities. A standardised ward register for inpatient small and sick newborn care may capture selected supportive care data. However, tracking the detailed care will require standardised individual-level data sets linked to newborn case notes. We recommend investments in assessing the im-plementation aspects of a standardised inpatient register in resource-poor settings