Kangaroo mother care among hospitalised neonates: evaluation of the validity of duration measurement methods compared to observation linked to the OMWaNA trial in Uganda.

BACKGROUND: Studies evaluating the impact of kangaroo mother care (KMC) on neonatal mortality and morbidity often rely on healthcare worker records or caregiver reports to measure intervention duration. However, the accuracy of these methods remains uncertain. We examined the validity of different methods of KMC duration measurement amongst neonates ≤ 2000 g in Uganda. METHODS: This observational study was embedded within the OMWaNA trial, which examined the impact of KMC on neonatal mortality before clinical stability. An independent observer (considered the gold standard) monitored neonates every 2 h to confirm KMC position, using an Android tablet-based application adapted from the EN-BIRTH study. The gold standard was compared to routine healthcare workers' charting and caregiver diary reports of KMC. RESULTS: Among 222 caregiver-newborn pairs, 219 initiated KMC. The mean daily KMC duration recorded by the gold standard was 8·4 h (SD 3·5). Healthcare workers reported an average of 8·5 h (SD 4·0), while caregivers reported 10·4 h (SD 3·8). The mean difference was 0·2 h less for healthcare workers (95% CI -0·3 to 0·6) and 1·7 h more for caregivers (-2·1 to -1·3) compared to the gold standard. Agreement rates for individual KMC episodes were 55·2% (95% CI 54·4-55·9) for healthcare workers and 58·2% (57·2-59·0) for caregivers. Participants with a helper (substitute KMC provider) had longer daily duration compared to those without (mean difference 1·89 h [0·89 - 2·84]; p < 0·001). CONCLUSION: Healthcare worker records provide a reasonably accurate estimate of KMC duration at the population level, supporting the integration of KMC indicators into national health information systems to facilitate monitoring and evaluation. The presence of a helper increases KMC duration, underscoring the need for research to identify strategies to increase family involvement.