Clinical Stroke Validation Methods in Under-Resourced Ethiopian Hospital Settings

Abstract

In hospitals with active stroke services, diagnosis, and treatment of stroke patients includes computed tomography (CT), perfusion imaging and angiography. CT scan is traditionally used to assess patients with suspected intracranial haemorrhage. Yet, it is expensive and inaccessible in most under-resourced hospital settings. In these setups, clinical diagnosis of either haemorrhagic or ischemic stroke might turn out to be inaccurate, and as such, stroke validation techniques need to be addressed within the framework of specific hospital setups. Accordingly, we conducted a prospective cross-sectional study to assess the level of accuracy of clinical stroke validation score methods to distinguish stroke subtypes, including aging-related ischemic stroke patients. CT images generated from 140 participants at Mizan-Tepi University Hospital in Ethiopia were interpreted by a neuroradiologist. Data collection and analysis were done using a patient checklist and SPSS. Findings revealed an incidence of haemorrhagic stroke was 50%, ischemic stroke was 48.6% among the studied stroke patients. The sensitivity, specificity, positive predictive value, negative predictive value and the overall accuracy of Siriraj stroke scores to differentiate haemorrhage from ischemic stroke were, 83.9%, 68.6%, 74.6%, 79.5%, and 82%; the Guys stroke scores were 47.8%, 89.7%, 73.3%, 74.5%, and 74.5%; the Greek stroke score were 58.8%, 88.5%, 87%, 62.2%, and 71.6%; while the Besson stroke scores were 35.3%, 88.6%, 75%, 58.5%, and 62.3% respectively. Based on our findings, the Siriraj stroke score showed good sensitivity and fair overall accuracy for haemorrhagic stroke, in low-resource settings. In our stroke validation settings, aging-related cortical atrophy in an 85-year-old female was an impressive finding consistent with the middle cerebral arterial territory of a patient with acute ischemic infarction.