Predicting Age-Related Hearing Loss in Community-Dwelling Older Adults: Multicenter Retrospective Cohort Study

Age-related hearing loss (ARHL) is associated with severe negative outcomes, including social isolation, depression, and cognitive decline. Despite this, routine ARHL screening is often neglected in primary care due to low awareness, resource limitations, and inefficiencies. A practical risk assessment tool could effectively address this gap.

This study aims to develop and validate a user-friendly nomogram for identifying older adults at high risk of ARHL in community settings, thereby facilitating targeted screening and timely interventions.

This multicenter retrospective cohort study included 34,983 older adults from 3 primary health care centers in Beijing (January 2020 to October 2023). Data from center A (n=18,707) were used for model development, with external validation performed on cohorts from center B (n=11,008) and center C (n=5268). Least absolute shrinkage and selection operator and logistic regression identified the final predictors. Model performance was evaluated using discrimination, calibration, and decision curve analysis, leading to the development of an online nomogram.

In the training cohort (center A), 1177 participants (6.3%) had hearing loss. Six key predictors were identified: age, education, exercise frequency, physical function, dietary habits, and hypertension. The multivariate logistic regression model demonstrated good discrimination in internal validation (area under the curve [AUC] 0.806, 95% CI 0.782-0.831; sensitivity 0.774; specificity 0.820). External validation confirmed its generalizability (AUC 0.720, 95% CI 0.670-0.771 and AUC 0.747, 95% CI 0.712-0.782). Decision curve analysis highlighted a substantial clinical net benefit. A user-friendly online prediction web page was also developed.

We successfully developed and validated a dynamic, online nomogram for predicting ARHL in older adults. Comprising 6 readily available predictors, this model shows potential as a practical, online tool for proactive risk identification in primary care. However, further validation in larger and more diverse populations is essential to confirm its generalizability and real-world clinical utility.