Abstract
<title>Abstract</title> <p>Background. Neonatal intensive care units (NICUs) routinely host two clinically and operationally distinct cohorts: neonates with congenital anomalies whose course is dominated by neurosurgical decision-making, and neonates with acute acquired neurological insults. Whether these two trajectories translate into measurably different lengths of hospital stay (LOS) in a low-resource Southern African setting remains poorly characterised. Objective. To compare LOS between the congenital/neurosurgical and the acute acquired neurological macrogroups among neonates admitted with neurological or neurosurgical conditions to a tertiary NICU in Luanda, Angola. Methods. A retrospective observational study was conducted on a cleaned administrative database of 1,728 NICU episodes (after deduplication and terminology normalisation of 1,774 original records). Of 536 neurological/neurosurgical cases, 153 formed the congenital/neurosurgical macrogroup and 383 the acute acquired neurological macrogroup. LOS was compared with the Mann–Whitney U test; anticonvulsant use was compared with Pearson's χ²; heterogeneity across the seven primary diagnostic groups was assessed with the Kruskal–Wallis test. Reporting followed the STROBE framework [1]. Results. LOS was substantially longer in the congenital/neurosurgical macrogroup (n = 137 with LOS data; median 26.0 days; Q1–Q3 17.0–38.0; mean 29.54 ± 18.68) than in the acute acquired neurological macrogroup (n = 341 with LOS data; median 6.0 days; Q1–Q3 4.0–13.0; mean 9.57 ± 8.77; Mann–Whitney U = 40,352.5; p = 1.20 × 10⁻³⁵). Anticonvulsants were used in 3.27% (5/153) of the congenital/neurosurgical macrogroup versus 19.06% (73/383) of the acute acquired macrogroup (χ² = 20.68; p = 5.44 × 10⁻⁶). Heterogeneity in LOS across the seven primary diagnostic groups was confirmed (Kruskal–Wallis H = 160.04; p = 5.81 × 10⁻³²). Conclusion. Congenital neurosurgical admissions consumed markedly more inpatient days than acute neurological admissions, with an effect that is statistically robust and clinically plausible. The distinction has direct implications for bed planning, operating-room prioritisation, transfer pathways and the organisation of neonatology–neurosurgery interfaces in Angola and comparable Sub-Saharan African settings. Causal inference, mortality and definitive functional outcomes cannot be drawn from this design.</p>