Logistic regression analysis of factors influencing the effectiveness of intensive sound masking therapy in patients with tinnitus

Objectives To investigate factors influencing the effectiveness of intensive sound masking therapy on tinnitus using logistic regression analysis. Design The study used a retrospective cross-section analysis. Participants 102 patients with tinnitus were recruited at the Sun Yat-sen Memorial Hospital of Sun Yat-sen University, China. Intervention Intensive sound masking therapy was used as an intervention approach for patients with tinnitus. Primary and secondary outcome measures Participants underwent audiological investigations and tinnitus pitch and loudness matching measurements, followed by intensive sound masking therapy. The Tinnitus Handicap Inventory (THI) was used as the outcome measure pre and post treatment. Multivariate logistic regression was performed to investigate the association of demographic and audiological factors with effective therapy. Results According to the THI score changes pre and post sound masking intervention, 51 participants were categorised into an effective group, the remaining 51 participants were placed in a non-effective group. Those in the effective group were significantly younger than those in the non-effective group (P=0.012). Significantly more participants had flat audiogram configurations in the effective group (P=0.04). Multivariable logistic regression analysis showed that age (OR=0.96, 95% CI 0.93 to 0.99, P=0.007), audiometric configuration (P=0.027) and THI score pre treatment (OR=1.04, 95% CI 1.02 to 1.07, P<0.001) were significantly associated with therapeutic effectiveness. Further analysis showed that patients with flat audiometric configurations were 5.45 times more likely to respond to intervention than those with high-frequency steeply sloping audiograms (OR=5.45, 95% CI 1.67 to 17.86, P=0.005). Conclusion Audiometric configuration, age and THI scores appear to be predictive of the effectiveness of sound masking treatment. Gender, tinnitus characteristics and hearing threshold measures do not seem to be related to treatment effectiveness. A further randomised control study is needed to provide evidence of the effectiveness of prognostic factors in tinnitus interventions.