What is known and what is new about this study? • Scaling up kangaroo mother care (KMC) has been slow despite the strong evidence base that KMC improves survival for stable babies ≤2000 g weight. Improving data to track coverage is vital to end preventable preterm deaths, the leading cause of under-five mortality. • EN-BIRTH was a large multi-country observational study to assess validity of KMC coverage measurement (n = 840 mother-baby pairs) in exit-survey and routine registers. We observed content and quality of KMC and conducted interviews with health workers and data collectors to explore barriers and enablers to routine register recording.

Survey – what did we find and what does it mean? • Women’s exit survey report after admission to KMC ward/corner had high sensitivity, the first validity testing for measurement.

Register – what did we find and what does it mean? • We found that KMC coverage had high sensitivity in specific KMC registers. Despite the time load for multiple register filling, health workers were motivated if they saw data being used. • KMC coverage measured from KMC specific registers was more accurate than from general registers. • Routine measurement of KMC provided in other wards and for babies re-admitted to KMC wards was not assessed in our study and will be key to consider in the future. • Unnecessary duplication of KMC data elements in multiple documents needs to be streamlined to reduce burden on nurses.

Gap analysis for quality of care and measurement • Observation showed coverage of KMC was not a good proxy for receiving high-quality KMC. • Gaps in quality of care were identified even for initial observation of all KMC position components and baby wearing a hat. • Detailed analyses were conducted in the two Tanzanian hospitals and found large gaps in optimal KMC daily duration/dose and feeding. Focus on supporting care providers for KMC continuity needs to be prioritised to realise the potential of this intervention. • Arrangements for families to support mother-baby pairs during admission was not always available.

What next and research gaps? • Register data for babies admitted to KMC wards have potential for aggregation in routine health information systems (HMIS) to track coverage. More research is needed to assess data flow and quality at different levels of HMIS, including how to capture KMC provided in other newborn wards. • Further research is needed to explore if KMC can still be accurately reported at the typical 2–5 year population-based survey intervals by women who provided or did not provide KMC, and if sample size in household surveys is feasible to capture babies with birthweight ≤2000 g. • Measuring quality of KMC provision and experience of care is less likely to be feasible in routine information systems and further research is needed to identify the best approach. This may include special studies or perhaps routinely tracking selected specific components (e.g. wearing a hat).