Background: Successful follow-up of community-recruited participants in HIV prevention research could be influenced by the household visit schedule of field teams. The aim of this study is to investigate the impact of different field work shifts on successful follow-up of participants.
Methods: HPTN 071 (PopART) is a community-randomised trial of a household-based HIV-prevention intervention package in South Africa and Zambia. The primary outcome, HIV incidence, is measured in a randomly selected Population Cohort (PC). We used data from the 12 month follow-up survey (July 2015 - June 2016) of the PC in 9 South African communities. Follow-up household visits among individuals consented to participate at baseline were scheduled in three shifts: Midweek (Monday-Friday) shifts (8 hour duration) categorized as early shift if it ended before 4pm; as late shift if it ended after 4pm; and a Saturday shift (5 hours). During each shift, 4 teams (two individuals per team) were active per community. A participant was considered successfully followed-up if the survey was completed during the household visit. The number of successful follow-up visits were calculated for each shift type and standardized to account for variation in shift duration. Poisson regression analysis was used to calculate incidence rate ratio''s (IRRs) and 95% confidence intervals (CIs) of number of successful visits. Month of visit was included in the model to account for confounding with calendar time.
Results: 11,720 participants, including 8,251 females (70%), were successfully followed-up during 223 days. The rate of successful visits per hour was 5.1 for early shifts (CI=4.1-6.0), 6.8 for late shifts (CI=5.9-7.6) and 15.7 for Saturday shifts (CI=12.1-19.3). Follow-up visits during Saturday shifts were more successful as compared to visits during early or late midweek shifts (IRR=3.1, CI=2.5-3.9; IRR=2.4, CI=1.9-2.9, respectively). Among males, Saturday shifts resulted in 3.6 times as many successful visits compared to early midweek shifts (CI=2.8-4.5) and 2.9 times as many compared to late shifts (CI=2.4-3.5). Among females, these IRRs were 2.9 (CI=2.3-3.7) and 2.2 (CI=1.7-2.7).
Conclusions: To efficiently follow-up participants, in particular males, more household visits should be scheduled during Saturday shifts instead of midweek shifts. These findings can be used for implementation of community-based research.

Download the e-Poster