Background <p>Achieving the UNAIDS 95-95-95 targets is challenged by suboptimal retention in antiretroviral therapy (ART) programmes, particularly in resource-limited settings. In Ghana, where ART coverage remains low, HIV treatment interruption (TI) poses a significant barrier to sustained viral suppression and HIV control. This study aimed to quantify the magnitude and identify predictors of HIV interruption in treatment across six administrative regions in Ghana.</p> Methods <p>We conducted a retrospective cohort analysis of ART program data in Ghana across 6 regions from January 2019 to December 2023. A total of 33,613 participants contributed 227,319 follow-up visits. The primary outcome was treatment interruption (TI), defined as a missed clinic visit more than 28 days after the scheduled appointment. We used the Andersen–Gill model for recurrent event analysis to assess the predictors of TI.</p> Results <p>We found a high rate of TI, with about 68.3% of participants experiencing at least one treatment interruption, especially early in their care journey within the first three to six months. The risk of TI was found to cluster in specific regions and clinics, with newer or less experienced sites showing significantly higher rates of interruption. Patients on NNRTI-based regimens were identified to have high hazard for TI. Conversely, factors like providing a phone number, receiving care at primary care facilities or at faith-based facilities were found to be protective against TI.</p> Conclusions <p>To improve retention, Ghana’s HIV programmes must intensify efforts, with a focus on targeting high-risk sites and sub-groups, optimising ART regimens by transitioning to dolutegravir-based alternatives and leveraging multi-month dispensing and differentiated service delivery (DSD) models. Addressing these determinants is crucial for sustaining lifelong ART and advancing Ghana toward its national and global HIV control goals.</p>