Outcomes for low-birth-weight (LBW) neonates reflect both biological factors and the timeliness of the care pathway. The Three Delays Model—Delay 1: decision to seek care; Delay 2: reaching a facility; Delay 3: receiving adequate care after arrival—provides a validated framework to locate where the pathway fails. This study applied the model to characterise care-pathway barriers affecting LBW neonates admitted to a county referral hospital in Kenya and to relate delay indicators to severe adverse outcomes.
The investigation was a facility-based mixed-methods cross-sectional study conducted at the newborn unit of Kericho County Referral Hospital. The quantitative component enrolled 169 LBW neonate–mother pairs admitted to the unit. The qualitative component comprised nine key-informant interviews with providers. The authors stated that denominators for delay indicators were defined explicitly and that quantitative and qualitative data were integrated in a delay-structured joint matrix.
Delay indicators were constructed for each of the three delays with explicit denominators. Descriptive statistics were used to summarise each indicator. Associations between delay indicators and a composite severe adverse outcome were examined using chi-square or Fisher exact tests. The authors emphasised descriptive testing rather than multivariable modelling.
Nine provider interviews were analysed thematically. Interview codes were mapped directly to the three delays: decision-making and caregiver recognition (Delay 1), transport and referral coordination (Delay 2), and arrival-phase care processes such as stabilisation, supplies, staffing, and monitoring (Delay 3). Qualitative results were used to contextualise and explain quantitative findings within the delays framework.
A severe adverse outcome occurred in 136 of 169 neonates (80.5%). Several pathway barriers clustered before hospital arrival:
After arrival, active care began within 30 minutes for 66.3% of neonates.
In bivariate analyses, referral/outborn status was associated with higher crude odds of a severe adverse outcome (crude OR 2.25, 95% CI 1.01–5.00; p = 0.043). Essential drug/feed shortage was also associated with higher odds of severe outcome (OR 2.26, 95% CI 1.04–4.90; p = 0.036).
The authors report paradoxical inverse associations whereby decision delay, transport problems, and any pathway delay were each associated with a lower proportion of severe outcomes (all p < 0.01). They interpreted these inverse associations as likely reflecting confounding by indication and reverse causation: the sickest neonates were prioritised for rapid transfer and expedited care, biasing crude associations and making delays appear protective in descriptive comparisons.
Provider narratives, coded to the Three Delays, described multiple barriers:
These provider accounts corroborated quantitative indicators that pathway barriers were concentrated before arrival and highlighted the fragility of early in-hospital care.
The combined quantitative and qualitative analysis supports the interpretation that vulnerability for the smallest neonates accumulates along the prehospital pathway. Referral/outborn status should be seen not simply as a transport category but as a marker of compounded vulnerability resulting from delayed decisions, transport constraints, incomplete pre-referral stabilisation, and receiving-facility response gaps. The observed crude associations and paradoxical inverse relations are plausibly explained by selection and confounding rather than by beneficial effects of delay.
Based on the findings, the authors emphasise targeted interventions focused on shortening specific, identifiable delays rather than attempting to reproduce a broad determinants model. Priority areas identified include:
These measures aim to reduce accumulated vulnerability and lower the risk of severe adverse outcomes among LBW neonates.
The study reported ethical approvals: Kabarak University Research Ethics Committee (approval number KUREC-170226), licensing by the National Commission for Science, Technology and Innovation (NACOSTI), and administrative authorisation from Kericho County Referral Hospital. Written informed consent was obtained from mothers/guardians for the neonatal record component and from providers for interviews. Data were de-identified prior to analysis. The authors stated that data are accessible on the OSF platform at the provided DOI.
The manuscript is posted as a medRxiv preprint and has not been peer reviewed; the authors and medRxiv noted that findings should not be used to guide clinical practice without further evaluation.
The study used descriptive association testing and did not present multivariable modelling to adjust for potential confounders. Where inverse or paradoxical associations were observed, the authors cautioned against interpreting these as protective and attributed them to confounding by indication and reverse causation. Specific numerical details, confidence intervals, and p-values reported above are those provided in the source report.