Diagnostic Accuracy of the Cincinnati Prehospital Stroke Scale in an Urban Emergency Department in Ghana.

Background The Cincinnati Prehospital Stroke Scale (CPSS) is a quick and easy-to-use tool that has been validated for early stroke identification, particularly in prehospital settings. Its utilization in low-resource settings, where access to imaging may be limited or costly, has not been thoroughly explored. This study sought to evaluate the diagnostic accuracy of the CPSS in a large urban emergency department (ED) in Ghana, which has a high volume of stroke admissions. Methods A prospective cross-sectional study was conducted in the adult ED at Komfo Anokye Teaching Hospital (KATH), Ghana's second largest hospital. Patients, 18 years or older, who were referred or self-reported to KATH ED with clinical features suggestive of stroke, were identified by trained triage nurses upon arrival. Following informed consent, the nurses collected demographic information and also administered the CPSS. The components of the CPSS, which include facial droop, arm drift, and speech impairment, were each assigned a score of 1 if present and 0 if absent. All responses, including the total CPSS score, were documented on pre-designed paper forms and later entered into an electronic platform. A total score of ≥1 indicated a possible stroke. Patients were then followed throughout their hospital stay until they received brain computed tomography (CT) imaging, and the results interpreted by a specialist radiologist to confirm the diagnosis of stroke. Results A total of 110 individuals met the inclusion criteria and gave consent to participate. These included 51 males (46.4%) and 59 females (53.6%), with ages ranging from 33 to 92 years (median: 54 years, IQR: 48-67). The majority of patients (89, 89.9%) were referred. Eleven patients (10%) were unable to undergo CT imaging, while 99 patients (90%) completed CT imaging, with 86 (78.2%) having a confirmed stroke diagnosis. Upon reviewing the CPSS scores, 18 patients had incomplete entries and were excluded. The remaining 92 were analyzed. Among them, 35 patients (38.0%) scored 1 for face droop, 72 (78.3%) scored 1 for arm drift, and 72 (78.3%) scored 1 for speech impairment. Eighty-seven patients (94.6%) had a total CPSS score of ≥1. Sixty-nine patients (75.0%) had both a CPSS score of ≥1 and a confirmed stroke on CT scan. The CPSS had a sensitivity of 88.5% (CI: 69.8-97.6), a specificity of 25.5% (CI: 15.8-38.0), a positive predictive value (PPV) of 31.9% (CI: 21.4-44.0), and a negative predictive value (NPV) of 85.0% (CI: 62.1-96.8). Individually, face droop had the highest sensitivity (82.9%, CI: 66.4-93.4). Conclusion The CPSS, administered by trained triage nurses, showed a high sensitivity but low specificity in identifying patients with stroke at the KATH ED. It can serve as a valuable prehospital screening tool in low-resource settings. Future studies should explore its feasibility and adaptation across a wider group of prehospital personnel, including emergency medical technicians (EMTs) in Ghana.