Development and evaluation of a clinical model to predict low-frequency hearing prognosis in Menière’s disease: a multicenter retrospective cohort study
Year: 2026
Type: article
Study design: Observational Study
Abstract: Patients with Menière’s disease (MD) typically present with fluctuating low-frequency sensorineural hearing loss. The progression of hearing loss varies among individuals, making it challenging to predict prognosis and select appropriate treatment strategies. A method to estimate future hearing levels in patients with MD could support personalized treatment decisions; however, such an approach has not yet been established. Accordingly, we aimed to develop and validate a model to predict low-frequency hearing levels 1–2 years after the initial visit using readily available baseline patient data. We retrospectively collected electronic health record data from 124 patients diagnosed with definite MD according to the Bárány Society criteria. The development cohort ( n = 92) was recruited from two university hospitals, whereas the external validation cohort ( n = 32) was obtained from one general hospital. The outcome was the low-frequency hearing level measured 1–2 years after the initial visit. We developed the predictive model using multiple linear regression analysis, selecting predictors based on the smallest root mean square error (RMSE) during K-fold cross-validation. Model performance was evaluated using RMSE and the coefficient of determination (R²). The model incorporating age and low- and high-frequency hearing levels at the initial visit showed the highest predictive accuracy. By inputting the values of each predictor into the model, low-frequency hearing levels 1–2 years after the initial visit were predicted. In internal validation, RMSE ranged from 9.1 to 11.5 dB, with R² values of 0.64–0.78, and 71% of predictions were within the measurement error range of < 10 dB. In external validation, RMSE was 11.2 dB, R² was 0.60, and 63% of predictions were within 10 dB of the measured values. A simple model for MD that can reasonably predict low-frequency hearing levels 1–2 years after the initial visit was developed using three predictors readily obtained in routine practice. This model may assist clinicians and patients in making informed treatment decisions. Not applicable.
Language: English
Subfield: Neurology
Field: Neuroscience
Domain: Life Sciences
Open Access status: gold
Estimated APC paid: $2,990
OpenAlex ID: W7147666704
