Objective: To assess the relationship between the loudness of tinnitus and insomnia via a mediation analysis.
Design: Retrospective cross-sectional.
Study sample: 417 consecutive patients seeking treatment for tinnitus in an Audiology Department in the UK.
Results: Mediation analysis showed that the relationship between tinnitus loudness measured using the visual analogue scale (VAS) and insomnia measured using the insomnia severity index was fully mediated via depression measured using the Patient Health Questionnaire, tinnitus handicap measured using the Tinnitus Handicap Inventory and tinnitus annoyance measured using the VAS. The regression coefficients for the indirect effects of tinnitus loudness on insomnia were: via depression b = 0.53 (95% confidence interval, CI: 0.35-0.71); via the VAS for tinnitus annoyance, b = 0.33 (95% CI: -0.004-0.66); and via tinnitus handicap, b = 0.38 (95% CI: 0.16-0.6). The coefficient for the total indirect effect was b = 1.23 (95% CI: 0.89-1.58). The coefficient for the direct effect of tinnitus loudness on insomnia was b = 0.11 (95% CI: -0.27-0.51), a non-significant effect.
Conclusions: Insomnia is not directly related to tinnitus loudness. Depression, tinnitus handicap and tinnitus annoyance mediate the relationship between tinnitus loudness and insomnia.
Keywords: Tinnitus; depression; insomnia; mediation analysis; tinnitus loudness.