Clinical outcomes and mortality risk among inborn and referred newborns admitted to hospitals in Kenya

Referred newborns in this study had nearly three times higher odds of mortality compared to inborn newborns, highlighting the need for improved referral and care processes.

Summary and findings

This study analyzed clinical outcomes and mortality risk among 130,773 newborns admitted to 13 hospitals in Kenya from January 2019 to October 2024. It found that referred newborns had a higher unadjusted mortality rate of 29.0% compared to 11.3% for inborn newborns. Factors associated with mortality included being extremely low birthweight and respiratory distress.

Abstract

<h4>ABSTRACT</h4> <h4>Background</h4> Newborns requiring inpatient care, particularly small and sick newborns (SSNBs), face high risk of mortality. Newborns referred from other facilities may experience worse outcomes than those born and managed within the same hospital ( inborn newborns ). Understanding factors contributing to this disparity in outcomes could support efforts to scale-up care and accelerate progress towards achieving Sustainable Development Goals target 3.2. <h4>Methods</h4> Data on 130,773 newborns admitted to 13 hospitals implementing with NEST360 in Kenya were obtained from the Neonatal Inpatient Dataset, between January 2019–October 2024. We described characteristics and primary diagnoses. Logistic regression was used to evaluate factors associated with mortality. <h4>Results</h4> Among admissions, 114,084 (87.2%) were inborn and 16,689 (12.8%) referred. Referred newborns were more likely to be extremely preterm (6.1% vs 3.1%), have extremely low birthweight (<1,000g) (4.6% vs 2.6%) and present with respiratory distress (26.2% vs 15.0%) and hypoxia (23.2% vs 15.3%) compared to those inborn. Only 59.6% of referred newborns were admitted on first day of life compared to 80.2% inborn newborns. Unadjusted mortality among referred newborns was 29.0% compared to 11.3% in those inborn. Risk factors associated with mortality among referred newborns included being extremely low birthweight (odds ratio [OR] 13.57, 95% CI 11.19–16.44), respiratory distress (OR 4.07, 95% CI 3.77–4.39), and congenital anomalies (OR 1.66, 95% CI 1.41–1.95). Prematurity and intrapartum-related complications were also associated with increased odds of death. In multivariable analysis, being referred remained strongly associated with mortality (adjusted OR [aOR] 2.54, 95% CI 2.39–2.71). <h4>Conclusion</h4> Referred newborns had nearly three times higher odds of mortality compared to those inborn. This may highlight referral selection bias amongst this group and could also be related to inadequate pre-referral stabilisation, unsafe neonatal transportation and admission delays. If successfully implemented, a strong hub-and-spoke approach may improve care at lower levels of care and decongest receiving facilities. Overall, improving quality of care across the continuum of referral process is a cornerstone in strategies to reduce neonatal mortality towards attainment of national and global newborn survival targets. <h4>KEY FINDINGS</h4> <h4>1. WHAT WAS KNOWN?</h4> Neonatal mortality remains high in sub-Saharan Africa and newborns referred from other health facilities may experience poorer outcomes than those born and managed within the same hospital. There is limited evidence on morbidity and mortality outcomes among inborn and referred newborns. This is important to inform specialised newborn care and targeted improvements in referral. <h4>2. WHAT WAS DONE THAT IS NEW?</h4> This study analysed routinely collected clinical data on 130,773 newborns admitted to 13 hospitals implementing with NEST360 in Kenya between 2019 and 2024. Diagnoses outcomes and neonatal characteristics were described and compared between inborn and referred newborns. Factors associated with neonatal mortality were also examined using logistic regression analysis. <h4>3. WHAT WAS FOUND?</h4> Referred newborns had higher unadjusted mortality rate than inborn newborns (29.0% vs 11.3%; p<0.001), with 3 times higher odds of death in univariable logistic regression analysis (OR 3.20, 95% CI 3.08–3.33). Referred newborns were more clinically vulnerable at admission and had higher proportions of extreme prematurity (6.1% vs 3.1%), very preterm birth (14.0% vs 8.6%), and extremely low birthweight (4.6% vs 2.6%). Among both groups, key risk factors associated with mortality included birthweight, gestational age, respiratory distress, hypothermia, and clinical diagnoses. Among referred newborns some of the risk factors associated with mortality included being extremely low birthweight (OR 13.57, 95% CI 11.19–16.44), respiratory distress (OR 4.07, 95% CI 3.77–4.39), congenital anomalies (OR 1.66, 95% CI 1.41–1.95), and intrapartum-related complications (OR 1.35, 95% CI 1.20–1.52). <h4>4. WHAT NEXT?</h4> Strengthen neonatal referral systems through clearer referral criteria, improved pre-referral stabilisation, better neonatal transport, and prompt triage on arrival at receiving hospitals. Routine clinical data should be used to monitor referral processes and outcomes and to guide continuous quality improvement. Further research is needed to capture referral to admission time, transport characteristics, and quality of pre-referral stabilisation. Linking neonatal admission data with maternal records and assessing outcomes beyond hospital discharge would also improve understanding of referral pathways and long-term outcomes.

Elsewhere in the PE 22-28 corpus