The data for this investigation were sourced from the SHED-IT weight loss maintenance trial, which has been published elsewhere [19,20]. Briefly, the study included 209 overweight/obese men from the Hunter Region of Australia who were recruited via a university media release [20]. Eligibility criteria were: (1) age 18-65 years old, (2) body mass index 25-40 kg/m², (3) Internet and mobile phone access, and (4) <5% weight loss in previous 6 months. The current study reports data from the initial weight loss phase of the trial (pre-post design), in which 209 men received the SHED-IT Weight Loss Program for 3 months before being randomized into one of two weight loss maintenance conditions. The study received ethics approval from the University of Newcastle’s Human Research Ethics Committee and was prospectively registered (ACTRN12612000749808).

All men received the SHED-IT program, which was a self-administered eHealth program that included no personalized intervention components. The program consisted of: (1) the SHED-IT Weight Loss Handbook and Weight Loss Log Book for Men, (2) the SHED-IT Weight Loss DVD for Men, (3) self-monitoring tools (ie, tape measure, pedometer), and (4) weekly motivational text messages (standardized). During the program men were advised to self-monitor their physical activity and diet using the freely available CalorieKing Australia website [21] or MyFitnessPal app [22] to create a 2000 kilojoule energy deficit on most days. After receiving the resource pack and an instruction sheet, participants were not provided with additional support during the intervention period.

To increase engagement, most elements were socio-culturally targeted for men, with attention given to both surface-structure components (eg, male-specific pictures and research findings) and deep-structure, value-based components (eg, humor, frank and realistic communication). The program taught men how to lose weight sustainably, without eliminating valued discretionary choices (eg, beer). Extensive details on program development and components are provided elsewhere [20].

Depressive symptoms were measured with the validated 8-item Patient Health Questionnaire (PHQ-8) [23], with men indicating how often in the past two weeks they experienced a range of symptoms associated with major depression (range=0-24). Following established guidelines [24], PHQ-8 scores ≥10 were used to indicate the presence of depression. In a validation study, 88% of people with major depression reported a PHQ-8 score >10 and 88% of people without major depression reported a PHQ-8 score <10 [23]. Weight was measured without shoes on a digital scale to 0.01 kg. In addition to recruitment and retention rates, program satisfaction metrics were collected using a revised process evaluation questionnaire, which was originally developed for use in a previous study [15]. In the current trial, the questionnaire also included new questions relating to engagement with online program components, which were developed specifically for this study.

All analyses were conducted in IBM SPSS Statistics 22 (Armonk, NY; IBM Corp). Changes in depressive symptoms and weight were examined using linear mixed models, and adjusted for the centered baseline score and the following covariates: age, socio-economic status, physical activity (steps/day), energy intake (kilojoules/day), and risky alcohol consumption. Linear mixed models are consistent with an intention-to-treat approach as they model missing responses with a likelihood-based analysis that includes all available data. Effect sizes were represented with Cohen’s d (mean change/standard deviation [SD] of change). Differences between men with and without depression for categorical outcomes were assessed with Chi-square tests.

As reported elsewhere [19], 209 of 236 men who met the study’s eligibility criteria were enrolled after returning consent. The mean weight and age of the sample were 105.7 kg (SD 14.0) and 46.6 years (SD 11.3), respectively. Overall, 36 men (36/209, 17.2%) reported comorbid depression at baseline. Although no incentives/reimbursements were offered for attending assessments, the posttest retention rate was 84.7% (177/209). A greater proportion of men with depression attended the posttest assessment (32/36, 88.9%) compared to those without (145/173, 83.8%), but the difference was not significant (χ²=0.6; degrees of freedom=1; P=.44).

As seen in Figure 1, men with depression reported a substantial decrease in depressive symptoms during the study (adjusted mean difference: -5.5 units, 95% CI -7.2 to -3.8; d=1.0). Consequently, 72.2% (26/36) of these men no longer met the criterion for depression at posttest. A corresponding, albeit smaller, intervention effect on depressive symptoms was also observed in men without depression at baseline (adjusted mean difference: -1.0 units, 95% CI -1.4 to -0.6; d=0.4). Overall, these changes represented a mean decrease in depressive symptoms of 1.8 units (95% CI 1.3 to 2.3; d=0.5) for the sample. The overall intervention effect on weight was -4.7 kg (95% CI -5.2 to -4.2; d=1.3), which did not vary significantly by depression status.

No significant differences were detected between men with and without baseline depression for the program acceptability questions included in the posttest process evaluation (Table 1). Overall, 82.7% (24/29) of the men with depression who completed the process evaluation reported that the program provided them with sufficient support to lose weight, and 93.1% (27/29) indicated they would recommend the program to their friends. The men also reported high levels of agreement that the program resources were enjoyable to read (23/29, 79.3%) and watch (24/29, 82.7%).

Usage rates for the online program components were comparable between men with and without depression (Table 2). Of the 82.8% (24/29) of men with depression who accessed the online components, the median self-reported usage rates were 4 x 10-minute visits/week for the CalorieKing website and 3 x 5-minute visits/week for the MyFitnessPal app, which aligned with program recommendations.