Objective To describe electronic diary study adherence, completion and sexual behaviour patterns among a high-risk sample clinically recruited for a prospective study of incident STI.

Method Participants (N=243; 18–29 yrs; 61% female; 88% African American) completed 12-weeks of weekly STI testing and 3x/day electronic diaries assessing individual and partner-specific affect, daily activities, substance use and sexual and contraceptive behaviour. We analysed event- and subject-level completion rates, and subject-level STI and behaviour response effects among subjects who had completed the study and who had not been incarcerated during study.

Results In total, 82.3% (200/243) of participants individually submitted at least 250 diaries (252 expected) during the study; as a group, participants submitted 89.7% (54 900/61 236) of the total number diaries expected. The average number of diaries weekly submitted varied between 17.03 and 19.45 (21 expected). Total and weekly submissions were similar by gender, age and race. The overall median survey completion time was 1.77 min (1.68 min: if no partners; 3.00 min: if any partners). Among those with only one partner, median completion time when no sex occurred was 2.47 min and was 3.87 min when any sex occurred. About 35% (85/243) of participants had an STI at some point in the study. Subjects reported the following sexual behaviours: vaginal sex (86.4%: 210/243), anal sex (29.6%: 72/243), giving oral sex (70.8%: 172/243) and receiving oral sex (82.7%: 201/243). Of those reporting vaginal sex, 60.4% (127/210) had only one partner and 23.3% (49/210) reported two partners. The median number of partner changes was one. Event level data suggest no response effects: sensitive behaviours, such as substance use or vaginal sex, exhibited similar stability in reports over time as compared to non-sexual activities such as eating or taking a walk.

Conclusion Data suggest excellent study protocol adherence and stable behaviour reporting among a high-risk sample. These findings demonstrate the feasibility of prospective, incident STI data collection methods, particularly in clinically recruited samples, which are often assumed to be non-compliant or unreliable around such methods.