Sunnybrook Facial Grading System reliability in subacute stroke patients.

BACKGROUND A lack of evidence on the reliability limits the utility of the Sunnybrook Facial Grading System (SFGS) in individuals affected by stroke both in clinical and research settings. AIM To examine the test-retest reliability and the inter-rater reliability of the SFGS in patients affected by subacute stroke. DESIGN A repeated-assessments design (fifteen days apart) was used to examine the test-retest reliability of the SFGS. SETTING Inpatient rehabilitation unit. POPULATION Thirty-two inpatients with subacute stroke. METHODS Thirty-two inpatients with subacute stroke participated in the test-retest reliability study. All patients were video-recorded in sitting position and uncovered face for about two minutes, in static position to evaluate symmetry at rest, then assessing symmetry of voluntary movement and synkinesis. For the inter-rater reliability, fifteen experts in neurorehabilitation were selected as raters. The facial function was assessed using the Italian version of the SFGS was used as grading system on a recorded video of each patient. The inter-rater reliability was carried out through the ICC, Intraclass Correlation Coefficient, while the intra-rater reliability was assessed comparing for each rater the scores evaluated at T0 and T1 using Wilcoxon signed-rank test. RESULTS Evaluations at T0 and T1 showed a significant correlation with a total ICC (T0 + T1) of 0.901 (95% CI 0.852-0.942; p-value <0.001). The ICC of T0 scores is 0.914 (95% CI 0.869-0.950; p-value <0.001), while the ICC of T1 scores is 0.886 (95% CI 0.829-0.995; p-value <0.001). The scores of the "voluntary movement symmetry" area are correlated with ICC of 0.916 (95% CI 0.835-0.958; p-value <0.001). The scores of the "resting symmetry" area show correlation with ICC of 0.567 (95% CI 0.277-0.762; p-value <0.001). The area of "synkinesis" shows a lower correlation with ICC of 0.175 (95% CI -0.180-0.489; p-value <0.001). The scores attributed to the patients in the two different moments of observation resulted not significantly different from 12 out of 15 raters. CONCLUSIONS The SFGS resulted reliable in its overall score to assess the facial palsy also in patients affected by stroke. CLINICAL REHABILITATION IMPACT The findings of this study provide empirical evidence of reliable properties of the SFGS for assessing central facial palsy in patients with stroke.