During the active treatment, participants will interact with the Lumen virtual coach (for depression management) and watch the GLB videos (for weight loss). Using the Amazon Alexa app on their study iPads, participants will complete 4 weekly Lumen PST sessions initially, 2 bi-weekly sessions, and then 3 monthly sessions during the active intervention. Beginning in week 6, participants will receive automated messages with instructions to view one GLB video a week for 11 weeks.

The SCL-20 is a valid, reliable measure of depression severity. It has been used in numerous depression treatment trials in primary care and community settings, making it particularly useful for cross-study comparisons and data synthesis in meta-analyses. The SCL-20 had good internal consistency (Cronbach's alpha=0.859) and correlated well with the General Health Questionnaire (r=0.78), a psychiatric screening test.

Dichotomous variable based on if participants reaching 5% weight loss from baseline.

Dichotomous variable based on if participants achieving 50% decline in Depression Symptom Checklist (SCL-20) scores from baseline.

Dichotomous variable based on if participants achieving Depression Symptom Checklist (SCL-20) scores <0.5.

Anxiety will be measured with the Generalized Anxiety Disorder Scale (GAD-7). GAD-7 is a valid and reliable 7-question scale for screening generalized anxiety disorder and strongly associated with multiple domains of functional impairment. It has excellent internal consistency (Cronbach's alpha=0.92) and good test-retest reliability (intraclass correlation=0.83). GAD-7 correlated well with the Beck Anxiety Inventory (r=0.72) and the anxiety subscale of the Symptom Checklist-90 (r=0.74).

Sleep disturbance will be assessed on the 8-item PROMIS sleep disturbance (SD) scale. SD measures symptoms of insomnia. Each item is scaled from 1 (not at all) to 5 (very much). Correlation between the SD short form and the Pittsburgh Sleep Quality Index (PSQI, hypothesized to measure similar attributes) is 0.83, similar to the correlations between the SD full banks and PSQI (0.85).

Sleep impairment will be assessed on the 8-item PROMIS sleep-related impairment (SRI) scale. SRI measures symptoms of daytime sleepiness. Each item is scaled from 1 (not at all) to 5 (very much). Correlation between the SRI short form and the Epworth Sleepiness Scale (ESS, hypothesized to measure a related but slightly different construct, the propensity to doze during activities) is 0.46, similar to the correlations between SRI full banks and the ESS (0.45).

The Work and Social Adjustment Scale (WSAS) is a validated questionnaire that measures impaired functioning. Patients rate the extent to which their depression symptoms impair work, home management, leisure activities, and relationships with others on a 0 (no impairment at all) to 8 (very severe impairment) scale. Cronbach's alpha ranged from 0.70-0.94. Test-retest correlation was 0.73.

The PROMIS Cognitive Abilities and Cognitive Concerns Scales each contain 8 items. The Cognitive Abilities items positively worded and the Cognitive Concerns items are worded negatively. Items on both subscales use a 5-point rating from "not at all" to "very much." Items are summed to create a total score for each subscale. Cronbach's α coefficients were 0.94 for both scales. The test-retest correlation coefficients were 0.83 for Cognitive Concerns and 0.80 for Cognitive Abilities over a 2-month period. A total score was calculated by summing item scores with reverse coding the Concerns subscale items. Higher total scores indicate better cognitive function.

The Obesity-Related Problem Scale specifically measures the impact of obesity on psychosocial functioning. The 8-item scale has internal reliability (Cronbach's a) above 0.90. The scale correlates strongly with theoretically related constructs and weight loss. The scale among the obese was significantly related to poor mood measured with the Mood Adjective Check List (MACL) (P<0.0001) and anxiety and depression symptoms measured with the Hospital Anxiety and Depression Scale (HAD) (P<0.0001). Change in scale over time was strongly correlated with weight loss (P<0.0001).

The 8-item Short-Form Health Survey (SF-8) is an 8-item version of the SF-36 that measures overall health-related quality of life. Physical and mental health composite scores are computed using the scores of 8 questions and range from 0 to 100, with 0 indicating the lowest level of health and 100 indicating the highest level of health. Correlation between the SF-8 and the SF-36 is 0.825 for physical health component and 0.881 for mental health component.

Social Network Index measures ties in four domains: marriage or partnership, friends and relatives, religious activity, and voluntary associations. Scores for the four dichotomized variables are summed with a range of 0 to 4, with a score of 4 indicating social ties across all four domains. Each type of social die significantly associated with lower mortality rate, independent of the other three.

Loneliness will be measured with the UCLA 3-item Loneliness Scale, with response categories being 1 (hardly ever), 2 (some of the time), and 3 (often). Responses to the 3 questions are summed, with higher scores representing greater loneliness. The Cronbach's alpha is 0.72. The 3-item loneliness scale significantly correlated with the 20-item Revised UCLA Loneliness Scale (r=0.82), depressive symptoms measured by a short form of the Center for Epidemiologic Studies-Depression Scale (CES-D) (r=0.48-0.49), and the 4-item Perceived Stress Scale (r=0.40-0.44).

All participants will receive a study-provided Fitbit activity tracker with heart rate and sleep monitor, assistance with account and device setup, and as-needed tech support. A study coordinator who is blinded to randomization will proactively contact participants by phone, email, and/or text to remind them to wear their activity tracker for at least 10 hours on 3 or more days during the assessment week at baseline (0), 6, 12 months. We will obtain participants' self-monitored activity and heart rate data using a developed program for automated daily synchronization via the Fitbit open API.

Sleep data will be obtained from Fitbit Alta HR activity trackers following the same procedures as described above for physical activity assessment at 0, 6, and 12 months.

Dietary intake will be measured using 17-item fruit, vegetable, fat, and fiber screener by self-report. Correlations between screener estimates and estimated true intake were 0.5-0.8.

Social Problem-Solving Inventory-Revised: Short Form (SPSI-R:S) assesses participants' problem-solving abilities. The scale is reliable and valid, containing 25 items in the following 5 subscales: positive problem orientation (PPO), negative problem orientation (NPO), rational problem solving (RPS), impulsive/careless style (ICS), and avoidance style (AS). Each item is rated on a 5-point scale ranging from "not at all true of me" (0) to "extremely true of me" (4). SPSI-R:S is a short version of the long SPSI-R:L form which has Cronbach's alpha coefficient of 0.95 for the total score and of 0.67-0.92 for subscales.

Food insecurity will be measured with the 2-item Seattle Short form at baseline and 12 months. The 2 questions are about the food eaten in your household in the LAST 12 MONTHS, and whether you were able to buy the food you needed. Compared with the 6-item screen, the 2-item screen was found to have 95.4% sensitivity and 83.5% specificity.