Sex moderates diet quality differences in integrated collaborative care for comorbid obesity and depression: Post-hoc analysis of the RAINBOW RCT

Objectives: To investigate (1) whether an evidence-based behavioral weight loss intervention was associated with improved diet quality in adults with obesity and depression compared with usual care; and (2) whether the associations were modified by sex. Design: In the RAINBOW RCT, 409 participants were randomized in 1:1 ratio to receive a 12-month intervention integrating a Diabetes Prevention Program-based behavioral weight loss treatment with problem-solving therapy for depression (n = 204) or usual care (n = 205). Participants completed 24-h dietary recalls at baseline, 6, 12, 18, and 24 months. Dietary outcomes included the DASH score as the composite measure of diet quality and its components. Between-group differences in dietary outcomes were examined among all participants and by sex, using repeated-measures mixed-effects linear models. Results: Intervention group had mean age 50.9 (SD 12.2) years and 71% women, while control had 51.0 (11.9) years and 70% women. Changes in DASH scores did not differ between the intervention and usual care control groups through 24 months in both females and males. Compared with controls, males in the intervention group had decreased nut, seed, and legume intake at 6 (mean difference, -1.1; 95% CI, -1.9, -0.3 servings/day; P = 0.01; Pajd = 0.73) and 12 months (-1.0; -2.0, -0.0 servings/day; P = 0.048; Pajd = 0.73). Compared with controls, females in the intervention group had decreased fruit and vegetable intake at 18 (-1.8; -2.9, -0.6 servings/day; P = 0.002; Pajd = 0.08) and 24 months (-1.1; -2.2, -0.1 servings/day; P = 0.03; Pajd = 0.25), and whole grain intake at 24 months (-0.5; -0.9, -0.1 servings/day; P = 0.03; Pajd = 0.25), but increased percent calories from fat at 24 months (3.6; 0.6, 6.5; P = 0.02; Pajd = 0.25). Conclusion: Diet quality not only did not improve in an effective behavioral weight loss intervention but deteriorated in females, in particular. These post-hoc findings warrant confirmation and may suggest sex-tailored behavior change techniques specifically targeting diet quality are needed in behavioral weight loss interventions aside from caloric reductions. Trial registration: ClinicalTrials.gov#NCT02246413 (https://clinicaltrials.gov/ct2/show/NCT02246413). (c) 2024 The Author(s). Published by Elsevier Masson SAS on behalf of SERDI Publisher. This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).