A. O'Neil, K. Richardson, J. A. Davis, T. John, S. Mahoney, M. Bryan, R. Fiddes, D. Meyer, D. Saunders, M. L. Chatterton, M. Turner, S. Brown, F. N. Jacka, N. L. Mundell, M. Gojanovic, L. K. Radovic, E. McDonald, M. L. Connolly, G. Murray, N. Thomas, S. Barrand, M. O'Shea, T. Rocks, E. S. George, W. Marx, M. P. Mosharaf, T. Jabeen, A. Ruusunen, R. Iyer, C. Mihalopoulos, M. Berk
Objective: To determine whether a multicomponent lifestyle intervention delivered by accredited dietitians and exercise physiologists was non-inferior to cognitive behavioural therapy (CBT) for reducing depressive symptoms in adults with moderate-to-severe mood disorders. Design: Randomised, parallel-group, single-blind, non-inferiority trial. Setting: Community-based recruitment across Australia with interventions delivered via telehealth. Participants: 358 adults with major depressive disorder or bipolar depression and moderate-to-severe depressive symptoms. Interventions: Lifestyle therapy versus brief structured CBT. Main outcome measures: Blinded interviewer-rated MADRS at 8 weeks with a prespecified non-inferiority margin of -1.6 points. Results: Mean session attendance was 5.11/7 in the lifestyle therapy group and 5.40/7 in the CBT group; 91% of participants attended the initial individual session, and primary outcome data were available for 267 participants (130 lifestyle therapy; 137 CBT). Non-inferiority was supported at 8 weeks in both the intention-to-treat ({beta}=1.76, 95% CI -0.28 to 3.81) and per-protocol ({beta}=1.58, 95% CI -0.56 to 3.73) analyses. Depressive symptoms decreased in both groups. Mean MADRS scores decreased by 8.55 points (95% CI 7.09 to 10.02) with lifestyle therapy and 6.82 points (5.38 to 8.25) with CBT in the intention-to-treat analysis, with similar findings in the per-protocol analysis (8.37 points [6.84 to 9.90] and 6.79 points [5.27 to 8.31], respectively). Conclusions: Lifestyle therapy was non-inferior to CBT and may expand access to evidence-based depression care through a broader allied health workforce.