Each week a policy outcome is chosen at random with ⅓ mood, ⅓ activity, ⅓ sleep - this determines which category of message a participant will receive.

Each day in the study, a reinforcement learning algorithm will determine 1) if a participant will receive a notification that day or no notification that day, 2) the therapeutic strategy employed by the notification (Behavioral Strategy, Cognitive Strategy, Mindfulness, Motivational Interviewing, Distanced Self-Talk), and 3) if personalized data feedback will be included.

If assigned to receive a notification, 1 core message set that meets the above criteria will be randomly selected from a pool of 358 core message sets. Each core message set will be comprised of 4 messages containing comparable content, however they will be tailored based on the participant's wearable (steps, sleep) or mood data for the specified time interval (7 days, 30 days, since the start of internship) as follows: 1) no data, 2) low, 3) moderate, or 4) high.

The study's mobile app will be used to deliver push notifications. The notifications appear on the participant's phone lock screen. The notifications include 3 categories: mood notifications, activity notifications, sleep notifications. Mood notifications aim to increase the participant's mood. Activity notifications aim to increase the participant's physical activity. Sleep notifications aim to increase the participant's sleep duration. All notifications are categorized as one of five therapeutic approaches: 1) CBT-Behavioral, 2) CBT-Cognitive, 3) Distanced Self-Talk, 4) Mindfulness, 5) Motivational Interviewing.

Through the mobile app, participants enter a mood score (scale 1 - 10) every day of the study. 1 corresponds to lowest mood and 10 corresponds to highest mood.

Participant's daily step counts are recorded through a fitness tracker. High step counts are considered a positive outcome as it indicates more physical activity.

Participant's nightly sleep duration (in minutes) is recorded through a fitness tracker. High sleep duration is considered a positive outcome.

Prior to the start of the intervention and at quarterly intervals throughout internship year, all participants complete the Patient Health Questionnaire 9. High scores on the PHQ-9 correspond to a larger number of depressive symptoms.