Determinants of severe adverse outcomes among low-birth-weight neonates admitted to a county referral hospital in Kenya: a mixed-methods cross-sectional study

Abstract Background. Low-birth-weight (LBW) neonates carry a disproportionate burden of morbidity and mortality in resource-limited newborn units, yet facility-level evidence on the determinants of severe in-hospital outcomes at county referral level in Kenya is limited. We determined the maternal, neonatal and care-related determinants of severe adverse outcomes among LBW neonates admitted to Kericho County Referral Hospital (KCRH) and used healthcare-provider perspectives to explain the quantitative findings. Methods. We conducted a facility-based, convergent mixed-methods cross-sectional study. Quantitative data were obtained from 169 LBW neonate-mother pairs through structured maternal interviews and clinical-record abstraction; qualitative data came from nine key informant interviews with newborn-unit healthcare workers. A severe adverse outcome was defined as the occurrence of at least one of respiratory distress, sepsis, hypothermia, hypoglycaemia, prolonged neonatal unit stay ([&ge;]7 days) or neonatal death. Associations were examined using bivariate tests and a multivariable binary logistic regression model entering 13 candidate predictors simultaneously, with multicollinearity, calibration and discrimination diagnostics. Interviews were analysed thematically and integrated with the quantitative results in a joint display. Results. A severe adverse outcome occurred in 136 of 169 neonates (80.5%); respiratory distress was the most common single complication (69.8%), and neonates experienced a mean of 2.46 (SD 1.31) adverse outcomes. In the adjusted model (omnibus {chi}{superscript 2}(13) = 57.70, p < 0.001; Nagelkerke R{superscript 2} = 0.46; area under the receiver-operating-characteristic curve = 0.879), three factors independently predicted severe adverse outcome: lower birth weight (adjusted odds ratio [AOR] 0.997 per gram, 95% CI 0.995-0.999, p = 0.012), maternal pregnancy-induced hypertension (AOR 18.49, 95% CI 1.87-182.47, p = 0.013) and warm-chain care (AOR 10.94, 95% CI 1.40-85.41, p = 0.023; a direction consistent with confounding by indication). Providers emphasised the fragile first hour of stabilisation, staffing and workload, warm-chain maintenance, infection prevention, commodity availability and referral coordination. Conclusions. Severe adverse outcomes were near-universal among LBW neonates at KCRH and were driven by neonatal biological vulnerability, maternal hypertensive disease and the readiness of newborn-unit care. Improving outcomes requires early, birth-weight-based risk stratification of the smallest neonates alongside strengthening of antenatal detection of hypertension and reliable, timely newborn-unit care processes.