Developing a Flexible App Data Collection Tool for Substance Use Research

Substance use among adolescents and emerging adults remains an alarming public health issue which is associated with the leading causes of death. Although mobile applications are a promising data collection tool for understanding substance use among youth, existing mobile applications are limited in that they are not readily modifiable nor designed for long-term use. To address these challenges, our research team has developed a prototype mobile application named SARA (Substance Abuse Research Assistant) that contains multiple data collection tools (e.g., ecological momentary assessments, active tasks) and features various engagement strategies (reinforcers or notifications) designed to encourage data collection adherence over an extended period of time.

Baseline reinforcers which are administered to all participants include: (1) a virtual aquarium that participants see after opening the app; (2) virtual fish to populate the aquarium, which participants can earn points to unlock by completing daily data collection activities; (3) a new aquarium interface (i.e., leveling up) that participants unlock after using the app for a predetermined amount of time; (4) small financial rewards for each 3-day streak of completing daily data collection activities; and (5) reminder text messages/phone calls from research staff following predetermined reminder schedules. Key reinforcers and notifications are described in detail in the "intervention" section.

This study examines the impact of these key engagement strategies on participant compliance with daily data collection activities over a 30 day field testing period. The primary analyses will use a log linear generalization. The primary longitudinal binary outcome for the reinforcers will be whether on the following day the participant completes one or more self-report validated survey items. The secondary longitudinal binary outcome for the reinforcers will be whether on the following day the participant completes both active tasks. Survey questions answered prior to the randomization as well as the completion of one or more survey questions on the day prior to randomization, and the completion of both active tasks on the day prior to randomization, may be used as control variables in the above analyses of reinforcers. The primary longitudinal binary outcome for the notifications will be whether on the same day the participant completes one or more self-report survey questions. The secondary longitudinal binary outcome for the notifications will be whether on the same day the participant completes both active tasks. The completion of one or more survey questions on the day prior to, and the completion of both active tasks on the day prior to, the randomization will be included as control variables for the notifications.