Abstract
BACKGROUND: The first year after HIV treatment initiation or re-initiation is the period of highest risk of a treatment interruption or disengagement, yet little is known about the timing, patterns and effects of interruptions in the early treatment period. METHODS: Using routinely collected electronic medical record data from 543 Zambian facilities from 2018 to 2023, we described patterns of engagement during the first year of HIV treatment. We defined engagement patterns for months 0-6 and months 7-12 after initiation or reinitiation as (1) continuous (attended all scheduled clinic and medication pickup visits as planned; (2) cyclical (attended ≥1 visits late >28 days but returned to and remained in care) or (3) disengaged (missed a scheduled visit by >28 days and had no evidence of return). RESULTS: Our sample population comprised 159 429 adult participants (61% female, median age 33). Of the 513 322 interactions observed ≤12 months after initiation, 53% occurred as planned, 22% were late ≤28 days late, 9% were >28 days late, and 17% were scheduled but never attended. In 0-6 months after initiation, 51% clients were continuously engaged, 12% cyclically engaged and 33% disengaged. Two-thirds of disengagers (21% of cohort) did not return after the initiation visit. During months 7-12, most clients who had been continuously engaged in months 0-6 (54%) remained continuous, while 18% moved to cyclical engagement. Among cyclical engagers in months 0-6, nearly half (47%) moved to being continuously engaged by month 12. Only 34% of the study population remained engaged continuously by the end of the 12-month period. CONCLUSIONS: Fewer than 60% of clients initiating antiretroviral therapy care between 2018 and 2022 at Zambian facilities remained continuously engaged at month 6 and 34% at month 12. Cyclical engagement and frequent interruptions should be accepted as the norm and models of service delivery designed to accommodate them.