Patient- and Health Facility-Level Determinants of Stroke Fatality Across 15 Large Hospitals in Ghana & Nigeria

Background: Every 6 minutes, an indigenous African succumbs to stroke. Patient- and system-level contributors to early stroke mortality in this region are yet to be delineated.Purpose: To identify and quantify the contributions of patient- and system-level determinants of in-patient stroke fatality across 15 large hospitals in Ghana and Nigeria. Methods: The Stroke Investigative Research and Educational Network (SIREN) is a multicenter study involving 15 sites in Ghana and Nigeria. Cases include adults ≥18 years with clinical and radiological evidence of an acute stroke. Data on 23-items within the domains of stroke services and resources availability at each study site were collected and analyzed as system-level factors. A host of demographic, clinical, laboratory and treatment variables of cases were analyzed as patient-level factors. A generalized linear mixed model including both patient-level and system-level covariates were fitted. Results are presented as adjusted odds ratio (aOR) with respective 95% confidence intervals (95%CI). Sensitivity analyses by stroke type (ischemic vs intracerebral hemorrhage) are also presented.Results: Overall 814 (21.8%) of the 3,739 stroke patients admitted died on discharge; 18.1% for ischemic stroke and 30.6% for intracerebral hemorrhage. The variability in the odds of stroke fatality that could be attributed to the system-level factors across study sites assessed using model intracluster correlation coefficient was substantial at 20.4% (above a 5% threshold). The aOR (95% CI) of five patient-level factors associated with stroke fatality were: National Institute of Health Stroke Score (NIHSS) score 1.61 (1.31-1.99) for each 5 units rise, elevated intracranial pressure 3.96 (2.72-5.78), aspiration pneumonia 2.76 (1.78-4.27), use of anticoagulants 3.22 (1.42-7.30) and presence of cardiac arrhythmias 3.04 (1.71-5.45). Seven factors were independently associated with ischemic stroke mortality namely low vegetable intake, larger infarct volume >30 cm3, higher NIHSS score, use of anticoagulants, raised ICP, aspiration pneumonia, while use of anti-platelet therapy was protective. For intracerebral hemorrhage, NIHSS score, raised ICP, aspiration pneumonia and presence of arrhythmia were associated with mortality. Stroke units were available in only 5 out of 15 centers and did not independently predict case fatality in our models.Conclusion: A mix of patient attributes and system-level barriers identified in these 15 hospitals provide potential actionable points to address the high stroke fatality across hospitals in Africa.Funding Information: The SIREN (Stroke Investigative Research and Education Networks) and SIBS Genomics studies were funded by the National Institutes of Health grant U54 HG007479 and R01NS107900 under the H3Africa initiative.Declaration of Interests: None.Ethics Approval Statement: Ethical approval was obtained from all study sites and informed consent was obtained from all participants.