Chlorhexidine for facility-based umbilical cord care: EN-BIRTH multi-country validation study

Abstract

Background: Umbilical cord hygiene prevents sepsis, a leading cause of neonatal mortality. The World Health Organization recommends 7.1% chlorhexidine digluconate (CHX) application to the umbilicus after home birth in high mortality contexts. In Bangladesh and Nepal, national policies recommend CHX use for all facility births. Population-based household surveys include optional questions on CHX use, but indicator validation studies are lacking. The Every Newborn Birth Indicators Research Tracking in Hospitals (EN-BIRTH) was an observational study assessing measurement validity for maternal and newborn indicators. This paper reports results regarding CHX.

Methods: The EN-BIRTH study (July 2017-July 2018) included three public hospitals in Bangladesh and Nepal where CHX cord application is routine. Clinical-observers collected tablet-based, time-stamped data regarding cord care during admission to labour and delivery wards as the gold standard to assess accuracy of women's report at exit survey, and of routine-register data. We calculated validity ratios and individual-level validation metrics; analysed coverage, quality and measurement gaps. We conducted qualitative interviews to assess barriers and enablers to routine register-recording.

Results: Umbilical cord care was observed for 12,379 live births. Observer-assessed CHX coverage was very high at 89.3-99.4% in all 3 hospitals, although slightly lower after caesarean births in Azimpur (86.8%), Bangladesh. Exit survey-reported coverage (0.4-45.9%) underestimated the observed coverage with substantial "don't know" responses (55.5-79.4%). Survey-reported validity ratios were all poor (0.01 to 0.38). Register-recorded coverage in the specific column in Bangladesh was underestimated by 0.2% in Kushtia but overestimated by 9.0% in Azimpur. Register-recorded validity ratios were good (0.9 to 1.1) in Bangladesh, and poor (0.8) in Nepal. The non-specific register column in Pokhara, Nepal substantially underestimated coverage (20.7%).

Conclusions: Exit survey-report highly underestimated observed CHX coverage in all three hospitals. Routine register-recorded coverage was closer to observer-assessed coverage than survey reports in all hospitals, including for caesarean births, and was more accurately captured in hospitals with a specific register column. Inclusion of CHX cord care into registers, and tallied into health management information system platforms, is justified in countries with national policies for facility-based use, but requires implementation research to assess register design and data flow within health information systems.

Keywords: 7.1% chlorhexidine; Birth; Coverage; Health management systems; Hospital records; Neonatal sepsis; Newborn; Survey; Umbilical cord care; Validity.

Fig. 1

Chlorhexidine validation design, EN-BIRTH study. EN-BIRTH: Every Newborn Birth Indicators Research Tracking in Hospitals; HMIS: Health Management Information Systems; DHIS2: District Health Information Software 2; DHS: Demographic and Health Surveys; MICS: Multiple Indicator Cluster Surveys. 7.1% Chlorhexidine solution applied to the umbilicus

Fig. 2

Flow diagram for cord Chlorhexidine application in Bangladesh and Nepal, EN-BIRTH study. 7.1% Chlorhexidine solution applied to the umbilicus

Fig. 3

Coverage rates for Chlorhexidine cord application measured by observation, register and exit-survey, EN-BIRTH study (n = 12,379). Register-recorded (n = 11,002 live births) and exit survey-reported (n = 11,827 live births), split by three hospitals. BD: Bangladesh; NP: Nepal. 7.1% Chlorhexidine solution applied to the umbilicus

Fig. 4

Facility register design and completion approaches for Chlorhexidine application by site, EN-BIRTH study (n = 12,379). n = 12,379 observed live births, n = 10,772 register extracted live births. BD: Bangladesh; NP: Nepal. 7.1% Chlorhexidine solution applied to the umbilicus. In Bangladesh, the registers were revised to a standardised national EmONC register (Additional file 3), neither original facility register had any column for CHX documentation. Completeness calculations were “not possible” for Bangladesh registers as in this design, left blank also meant that the intervention/practice was not done. Reference: Cut-off ranges adapted from WHO Data Quality Review, Module 2 “Desk review of data quality” [36]

Fig. 5

Heat map of validity ratios for chlorhexidine cord application, EN-BIRTH study. BD: Bangladesh; NP: Nepal. Using cut off ranges adapted from WHO Data Quality Review, Module 2 “Desk review of data quality” [36]. Survey-reported to observed and register-recorded to observed. Observation n = 12,379 live births, register-recorded n = 10,002 live births and exit survey-reported n = 11,827 women with live births. 7.1% Chlorhexidine solution applied to the umbilicus

Fig. 6

Gap analysis for Chlorhexidine cord application coverage and quality, EN-BIRTH study (n = 12,379). Register-records n = 11,002 live births, and exit survey-report n = 11,827 women with live births. ‘Right time’ < 1 h was used here as the observation period is only during admission to labour and delivery wards. The current WHO recommendations advise that Chlorhexidine application should be completed within the first week of life [6]. 7.1% Chlorhexidine solution applied to the umbilicus