Development and Validation of a Nomogram to Identify and Predict High-Distress Psychosocial Phenotypes in Tinnitus Patients: A Latent Profile Analysis.

OBJECTIVE

Evidence is limited on the multidimensional psychosocial heterogeneity among patients with subjective tinnitus, and practical tools for early risk stratification remain largely unexplored. This study aimed to identify distinct psychosocial phenotypes under the biopsychosocial model to pinpoint the high-risk population and to develop and validate a visualized nomogram predicting severe psychological distress for targeted interventions.

METHODS

A total of 534 subjective tinnitus patients were enrolled. Latent profile analysis using seven psychosocial indicators (Self-Rating Anxiety Scale total score, Self-Rating Depression Scale total score, positive life events score, negative life events score, subjective support score, objective support score, and support utilization score) identified distinct patient phenotypes. The "Severe Anxiety Isolated" phenotype was defined as the target "high-distress" group for the prediction model. Clinical characteristics and tinnitus-related disability were compared across groups. Multivariate logistic regression identified independent predictors of the high-distress phenotype to construct a nomogram, validated via ROC, calibration, and decision curve analysis.

RESULTS

Four distinct psychosocial phenotypes were identified via LPA: the Chronic Vulnerable phenotype (18.73%), the Severe Anxiety Isolated phenotype (28.09%, defined as the high-distress high-risk phenotype), the Well Adapted Healthy phenotype (43.82%), and the High Positive Life Events Resilient phenotype (9.18%). The Severe Anxiety Isolated phenotype had significantly higher Tinnitus Handicap Inventory (THI) and Insomnia Severity Index (ISI) scores compared with other phenotypes (all P<0.001). Age group (21-39 years), longer tinnitus duration, concurrent presence of vertigo, bilateral tinnitus, and easily agitated/irritable personality traits were independent risk factors for the high-distress phenotype. The constructed nomogram exhibited good discrimination (AUC=0.756 in the training cohort and 0.737 in the testing cohort), calibration, and clinical utility.

CONCLUSION

Subjective tinnitus patients exhibit profound psychosocial heterogeneity, with the Severe Anxiety Isolated phenotype facing the highest psychological risk. Our developed nomogram accurately predicts this high-risk phenotype using routine outpatient indicators, facilitating early risk stratification and precision biopsychosocial interventions.