Scale reactivity analysis.

Background
Studies have shown that approximately 30% of patients with Bell’s palsy may experience permanent disfigurement due to muscle weakness affecting facial expressions. The prognosis for Bell’s palsy is often correlated with the degree of impairment in facial nerve function observed in the early stages of the condition. Selecting appropriate assessment tools is essential for the effective treatment and prognosis of patients with Bell’s palsy.
Objective
Examine and compare three scales used to measure Bell’s palsy: the House-Brackmann Grading Scale (HBGS), the Sunnybrook Facial Grading System (SFGS), and the Modified Portmann Scale (MPS).
Methods
A retrospective study involved 97 patients with Bell’s palsy treated at the authors’ institution from 2021 to 2022. The authors accessed these data for research purposes in February 2025. The patient cohort underwent HBGS, SFGS, and MPS to assess facial nerve function before and after treatment. We compared the results of these three evaluation methods. Additionally, we performed correlation analyses and receiver operating characteristic (ROC) analyses of the post-treatment data to examine the relationships between HBGS and SFGS and between HBGS and MPS, using the internationally recognized and most widely utilized HBGS as the standard.
Results
The HBGS demonstrated moderate consistency with both the FSGS and MPS, revealing significant negative correlations (r = −0.876, P < 0.01; r = −0.860, P < 0.01). All three scales exhibited high reactivity and showed no ceiling/floor effect. The optimal cut-off values for the FSGS and MPS were determined to be ≤ 68 and ≤ 16 points, respectively, with AUC of 0.948 (95% CI: 0.910–0.984) and 0.931 (95% CI: 0.883–0.968).
Conclusion
HBGS, SFGS, and MPS are all appropriate tools for assessing Bell’s palsy. The severity of facial nerve palsy can be clinically classified using a transect score of 68 on the SFGS and 16 on the MPS. However, the results should be interpreted with caution, as objective indicators were not utilized as criteria.