Patterns, severity, and outcomes of traumatic brain injury in a regional referral hospital in Kenya: a retrospective cohort study.

Background: Traumatic brain injury (TBI) is a growing public health burden globally, especially in low- and middle-income countries like Kenya. Inadequate prehospital care and limited health care infrastructure in rural areas hinder optimal TBI management. This study aimed to describe the patterns, severity, and short-term outcomes of TBI at a regional referral hospital in western Kenya. Methods: A retrospective cohort study was conducted, reviewing the medical records of 738 TBI patients admitted between January 2021 and December 2023. Data were extracted on demographics, injury mechanisms, Glasgow Coma Scale (GCS) scores, imaging, interventions, and discharge outcomes. Descriptive statistics were used to summarize patient characteristic. Chi-square tests and binary logistic regression were used to identify associations and predictors of mortality. Results: Most patients were males (89%) with a mean age of 30 years. road traffic accidents (RTAs) were the leading cause of injury (61%), followed by assaults (27%) - a figure higher than national and global averages. Mild TBI was most prevalent cases (51.6%). Imaging and neurosurgical intervention rates were low at 34.2% and 2.8%, respectively. At discharge, 59.9% of patients had fully recovered, while the mortality rate was 11.7%. Predictors of mortality included moderate (OR = 8.07) and severe (OR = 61.18) GCS scores and injury from falls (OR = 3.40). Conclusion: The study highlights the high burden of preventable TBI from RTAs and assaults in this setting. The high assault rate, low imaging access, and limited neurosurgical intervention capacity are major concerns. While the findings are limited to admitted patients and short-term outcomes, they underscore the urgent need for targeted road safety initiatives, improved access to specialized care, and a better understanding of the high prevalence of assault-related TBI to inform policy in similar low-resource settings.