Incidence and predictors of stroke mortality among adult patients admitted to the Stroke Unit at Jigjiga University Comprehensive Specialized Hospital

Abstract

Strokes represent a major global health issue, contributing to high morbidity and mortality rates among adults. They are the leading cause of long-term severe impairment in both developed and developing countries and have become a pressing public health concern in Ethiopia. Despite the significant burden of strokes worldwide, there is limited knowledge regarding the clinical profile of stroke in low- and middle-income nations like Ethiopia. Understanding the risk factors for stroke and predicting mortality among affected patients are essential for enhancing management strategies and improving patient outcomes. Therefore, the main objective of this study was to determine incidence and the risk factors associated with strokes mortality among adult patients admitted to the Stroke Unit at Jigjiga University Comprehensive Specialized Hospital. A retrospective cohort study was performed at Jigjiga University Comprehensive Specialized Hospital between September 2020 and April 2023, focusing on the medical histories and laboratory results of 549 patients who experienced strokes. To determine the risk factors associated with strokes mortality several count-regression models were employed, including Poisson, Zero-Inflated Poisson, Negative Binomial, Zero-Inflated Negative Binomial, and generalized negative binomial. This methodology improved the comprehension of the risk factors associated with strokes within the hospital's patient demographic. The number of stroke patient deaths was significantly impacted by the covariates, including hypertension (IRR = 1.33; 95% CI (0.08, 0.49); p-value < 0.01), female stroke patients (IRR = 0.36; 95% CI (− 1.19, − .87); p-value = 0.000), and stroke patients residing in urban areas (IRR = 0.84; 95% CI (0.35, − 0.02); p-value = 0.01). Furthermore, the study identified the Zero-Inflated Negative Binomial model as the most effective for predicting the mortality toll among stroke patients. This finding highlights the utility of the Zero-Inflated Negative Binomial model in forecasting mortality outcomes in this patient population. The findings indicate that a significant majority (85.97%) of patients did not experience stroke-related mortality. Several key factors influencing patient outcomes were identified, including age of patients, gender differences, and place of residence. Lifestyle factors such as alcohol consumption and smoking, along with comorbidities like diabetes and hypertension, further raise mortality risks. Socioeconomic factors, including work status and dietary habits, also contribute to outcomes, while a higher Body Mass Index (BMI) is associated with increased mortality. To reduce stroke mortality rates and enhance patient outcomes, it is essential to address these factors through targeted interventions and improved access to healthcare services.