Replication and extension of the simplified modified Rankin Scale in 150 Chinese stroke patients

Background: Recently, a simplified modified Rankin Scale (mRS) questionnaire (smRSq) showed good reliability but has not been tested for its validity by its original creators. Our study aimed to test its reliability and validity in Chinese stroke patients. Methods: Randomly chosen paired raters scored the smRSq, the conventional mRS, and the NIH Stroke Scale (NIHSS) face-to-face in 150 hospitalized stroke patients. Inter-rater reliability and concurrent validity were assessed for this translated questionnaire. Results: For inter-rater reliability of the smRSq, the overall agreement among the raters was 84%, the κ was 0.79 (95% CI 0.72-0.87), and the κw was 0.91 (95% CI 0.88-0.94). For inter-rater reliability of the mRS, the overall agreement among the raters was 81%, the κ was 0.75 (95% CI 0.67-0.83), and the κw was 0.88 (95% CI 0.84-0.92). The agreement between the mRS and smRSq was 71%, κ = 0.63 (95% CI 0.54-0.71), and κw = 0.83 (95% CI 0.79-0.88). The correlation between the NIHSS and the smRSq (concurrent validity) was moderate (Spearman's correlation coefficient 0.70, p < 0.0001). Conclusions: Our results confirm the value of the smRSq in the assessment of stroke functional outcome in China. As this is a novel stroke tool, further validations are needed.