Background: Antiretroviral therapy (ART) reduces HIV-related morbidity and mortality and is essential for epidemic control. To prevent interruptions in treatment (IIT), the Masaka Region HIV Program in south-central Uganda, under PEPFAR funding, implemented client-centered care including differentiated service delivery models (DSD), appointment tracking systems, and psychosocial support. However, IIT remain a challenge. We studied correlates of return and time to return to care among clients with IIT. Methods: We analyzed clients with IIT during January 1, 2020-March 31, 2021 receiving ART at 95 facilities in 12 districts. IIT was defined as no clinical contact for >28 days from missed appointment. Follow-up by home visits and/or phone calls commenced the day of missed appointment; outcomes (i.e. returned to care, self-transferred, migrated, died, unknown) up to May 1, 2021 were included. Descriptive statistics, logistic regression, and non-parametric tests of medians were conducted for correlates of return and time to return. Results: Of 2,356 clients with IIT, 1,381 (59%) were women. Median age, ART duration and time to return were 35 years (interquartile range [IQR]: 27.2,41.7); 3.6 yrs (IQR: 1.4,6.0), and 3.3 months (IQR 2.1,5.9) respectively. Most clients 2,220 (94%) received ART in facility-vs. 6% in community-based settings; 2,128 (90%) were virologically suppressed (<1,000 copies/ml) prior to IIT. Outcomes were 1,266 (54%) returned to care, 116 (4.9%) self-transferred, 35 (1.5%) migrated, 20 (0.8%) died, 919 (39%) unknown. Return was more among: clients in facility-based models vs. community models (Odds Ratio [OR]: 4.34, Confidence Interval [CI]: 2.36-7.96, p<0.001); clients on ART for 4-6 yrs vs. those on ART < 2 yrs (OR: 1.70, CI: 1.13-2.56, p=0.011); and non-suppressed vs. suppressed clients prior to IIT (OR: 4.05, CI: 1.64-10.0, p=0.002). Time to return was longer in facility-based vs. community models (median 3.35 vs 2.97 months, p=0.032), and non-suppressed vs. suppressed clients prior to IIT (5.43 vs 3.23 months, p<0.001). Conclusion: Almost 40% could not be traced. There were higher proportions but slower return in facility-based models; 75% returned clients returned within 6 months. It is important to consider more effective procedures to mitigate IIT especially in community-based treatment models and to act quickly once IIT occurs among clients with VL-non-suppression pre-IIT and in facility-based clients.