A comparison of remote and bedside assessment of the NIH Stroke Scale in acute stroke patients

Abstract

Telestroke videoconference for conducting the National Institute of Health Stroke Scale (NIHSS) is recommended when direct bedside evaluation by a stroke specialist is not immediately available for hyperacute stroke assessment. However, some NIHSS-telestroke studies inherit systematic bias due to subjectivity of NIHSS administration. Authors aimed to evaluate NIHSS telestroke assessment, while implementing measures to minimize subjectivity bias. Ninety acute stroke patients within 48 hours of onset were assessed by 6 stroke neurologists grouped in 15 pairs. Each pair of physicians assessed 6 patients. Patients were allocated through block randomization to a physician pair and order of bedside or remote assessment. Every patient was assessed once at the bedside and once remotely. Remote examination was performed by a neurologist through high-quality videoconferencing, assisted by a nurse at the patient's bedside. Kappa coefficients and the number of patients with a cumulative difference of ≤3 NIHSS points were calculated to compare bedside and remote measures. Cumulative difference of ≤3 NIHSS points was observed in 85.6% (95% CI 76.6%; 92.1%) cases. Therefore, every fifth remote examination may have been inaccurate. Quadratically weighted κ for total NIHSS score was 0.91 (95% CI 0.87; 0.95). Minimal agreements were for commands (κ=0.46), facial palsy (κ=0.43), and ataxia (κ=0.27). Remote assessments were longer than bedside: 8 minutes (IQR 7; 9) versus 6 (IQR 5; 8), p<0.001. NIHSS-telestroke assessment using high-quality videoconferencing in the acute stroke settings is closely matched with NIHSS-bedside assessment but it's credibility for clinical use needs further evaluation.