Background We investigated the impact of community tracing as recommended in national guidelines on outcomes within a prospective HIV cohort in Zambia.

Methods HIV-positive, antiretroviral therapy-naïve adults were enrolled at 2 Lusaka clinics. Per national guidelines we collected detailed baseline patient locator information including patient phone number, address, church, and a map from the clinic to their home. Study visits were aligned with routine ART program schedules and 3 telephonic attempts were made if visits were missed. Per guidelines, a lay health worker conducted a community home visit on lost to follow-up (LTFU) patients. Transfers to other clinics and deaths were ascertained when reported by the clinic staff, patients, or family members. At one year, we measured the percentage retained, transferred out, withdrawn (stopped ART), dead, and LTFU (>6 months absent). A lay health worker went into the community to make a home visit on LTFU patients. We recorded the change in mortality after tracing. We also estimated the time and costs per patient traced.

Results We enrolled 795 patients (median age 34 years; 53.7% were female; median CD4 228 cell/mm3). Prior to tracing, we recorded 45 deaths, 23 transfers, 1 withdrawal, and 83 LTFU who could not be reached by phone. At 63 attempted home visits, we learnt that 9 (14.3%) had died, 5 (7.9%) had transferred, and 2 had withdrawn. We could not locate 32 (50.8%) but neighbours/family reported that 12 of these had relocated (HIV care status unknown). After successful tracing, 15 (23.8%) returned to clinic and HIV care. Community tracing increased known mortality from 5.7% to 6.8% (95% CI: 5.1–8.8%) and increased retention at 1-year from 80.9% to 82.8%. Tracing required an average of 5 person-hours and K150.00 (∼15 USD) in bus/taxi fares per patient.

Conclusions Community tracing was limited by patient mobility and had a modest impact on cohort mortality and retention.