Mobile phones enable engagement with adolescents through a familiar medium. Survey data are presented from 2,023 South African adolescents who were asked about phone ownership, usage, and their willingness to divulge sensitive information in short message service surveys. Barriers to participation are addressed as are recommendations for follow-up research.

This Viewpoint describes ways in which artificial intelligence–powered large language models may be used to improve the delivery of mental health services worldwide.

To examine the feasibility of providing young adults with mobile phones for the purpose of mobile phone-assisted self-interviewing to improve retention in a long-term birth cohort study, mobile phones with survey software were distributed to 1,000 twenty-year-olds in the Birth to Twenty birth cohort study. Eleven months later, a targeted sampling frame was used to randomly select 435 participants from the subset of 734 phones that were still functional as survey tools. Text message notifications were dispatched at two time points, 2 weeks apart, requesting the completion of a 60-item survey. From the 435 young adults invited to participate in the survey, 105 (36.5%) submitted data in response to the first request and 84 (30.9%) submitted data in response to the second. The overall survey response rate was 33.7%, and item response rate varied from 88.5% to 100%. Contributing to the low response rate were challenges faced by both participant, including device loss and overly complicated survey procedures, and research team such as the deletion of the survey app by participants and the swapping out of study phone subscriber identity module cards making device management difficult. Reducing the effort required by participants to complete a survey, improving participant engagement in the data collection process, and using participants own handset are all suggestions for improving mobile survey data quality and responses rates.