Youth Ending the HIV Epidemic

Rate of Participant Retention will be calculated as the percent of those who were retained in the study and completed the final survey.

Mobile app data from the AiCure Application (paradata) will be used to measure the number of participant logins per week and thereby calculate mean logins per participant per week.

Mobile app data from the AiCure Application (Paradata) will be used to measure the number of seconds each participant spends in that app each day, thereby calculating the mean number of seconds in app per day per participant.

The AiCure mobile application platform uses an artificial intelligence or AI platform to recognize dosing patterns recorded by users that do not correspond to what has been defined as 'normal' dosing by the AI. The AI accordingly flags any video recording featuring abnormal dosing for review by the AiCure Video Review team. Upon confirmation of abnormal dosing, the AiCure Video Review team notifies the research coordinator in the form of a "Red Alert". Each of these "Red Alerts" message will be counted as an event of intentional nonadherence, and the mean number of times a participant may have "falsified" medication taking will be calculated and compared against the number of 'normal' doses.

The acceptability of the AiCure Mobile application will be through the System Usability Scale (SUS). The intervention was considered acceptable if ≥80% had a SUS score >68, which is considered above average acceptability.

Client satisfaction will be measured using a Client Satisfaction Questionnaire (CSQ-8). The CSQ-8 is an 8-item scale with higher values indicating higher satisfaction. Acceptability was set as a cutoff of 80% having a score of ≥17, which is the score cutoff considered above average and acceptable for this measure.

Recommend study to a friend. 7-point Likert Scale of how likely participants are to recommend the study to a friend. The intervention was considered acceptable if ≥80% reported likely or very likely to recommend.

Client satisfaction with the app+incentives. 7-point Likert Scale values from 1-7 with lower values corresponding to least satisfaction and higher values corresponding to greater satisfaction.

Adherence data will be abstracted from the automated directly observed therapy platform within the AiCure mobile health application. Percent adherence will be calculated as the percentage of days over the study period that participants recorded taking their medication within the app.

The preliminary effect on ART adherence will be measured through self-report using a 3-item scale that has been previously validated. Questions ask about frequency of missed medications in the last 30 days, adherence frequency in the last 30 days, and adherence rating in the last 30 days. Reponses to the 3 questions were transformed to a 0-100 scale, with higher scores indicating better adherence. A summary score was calculated as the mean of the 3 individual items.

Number of seconds in app collected using mobile app data (paradata) from AiCure platform.

How easy/difficult was it to use your personal phone; use adherence monitoring; receive incentives? Measured using a 7-point Likert Scale ranging from easy to difficult. Higher values correspond to greater ease of use, lower values correspond to greater difficulty.

Did you ever have trouble accessing app, using adherence monitoring, receiving reminders, receiving incentives, or finding a private place? 7-point Likert of frequency ranging from often to never with high value corresponding to app related issues occurring with greater frequency and lower value corresponding to app related issues issues occurring with lesser frequency.