Identifying psychological and clinical risk factors for moderate-to-severe tinnitus in older patients with hearing loss: a multivariable prediction model

Abstract

Objective: To develop and validate a clinical prediction model for moderate-to-severe tinnitus (THI ≥ 38) in patients with hearing loss and to identify the key psychological and clinical factors associated with its risk.

Methods: This retrospective single-centre study included 301 patients with hearing loss who visited Qinghai University Affiliated Hospital between August 2024 and May 2025. The cohort was randomly divided into a training set (n = 210) and a validation set (n = 91) in a 7:3 ratio. Moderate-to-severe tinnitus served as the outcome of interest. Psychological and clinical risk factors were initially screened using univariate logistic regression, and variables with p < 0.05 were subsequently included in a multivariable logistic regression model.

Results: The final multivariable model identified five independent psychological and clinical risk factors for moderate-to-severe tinnitus: older age (OR = 2.415), hypertension (OR = 2.120), poor sleep quality (OR = 2.821), anxiety (OR = 1.967), and severe hearing loss (OR = 3.452). The model demonstrated good discriminative performance, with an AUC of 0.734 in the training set and 0.760 in the validation set.

Conclusion: In patients with hearing loss, psychological and clinical risk factors-including poor sleep quality, anxiety, hypertension, and severe hearing loss-were significantly associated with moderate-to-severe tinnitus. These findings underscore the need for integrated management strategies that address both psychological and clinical components of tinnitus risk.

Keywords: anxiety; hearing loss; nomogram; poor sleep quality; tinnitus.

Figure 1

Performance evaluation of the prediction model for moderate-to-severe tinnitus. AUC, area under the receiver operating characteristic curve; OR, odds ratio; CI, confidence interval. (A) Receiver operating characteristic (ROC) curves for the training and validation sets, showing good discriminative ability in both sets. (B) Calibration curves based on 1,000 bootstrap resampling iterations, indicating good agreement between predicted probabilities and observed outcomes. Model calibration was further supported by the Hosmer–Lemeshow goodness-of-fit test. (C) Decision curve analysis (DCA) showing a positive net clinical benefit across a range of threshold probabilities in both sets. (D) Risk stratification based on tertiles of predicted probabilities in both the training and validation sets, showing a progressive increase in the observed incidence of moderate-to-severe tinnitus across low-, medium-, and high-risk groups.