BACKGROUND: Patients being treated for recurrent or multidrug-resistant tuberculosis (TB) require long courses of injectable anti-tuberculous agents. In order to maintain strong TB control programmes, it is vital that the experiences of people who receive long-term injectables for TB are well understood. To investigate the feasibility of a novel model of care delivery, a clinical trial (The TB-RROC Study) was conducted at two central hospitals in Malawi. Hospital-based care was compared to a community-based approach for patients on TB retreatment in which 'guardians' (patient-nominated lay people) were trained to deliver injections to patients at home. This study is the qualitative evaluation of the TB-RROC trial. It examines the experiences of people receiving injectables as part of TB treatment delivered in hospital and community-based settings. METHODS: A qualitative evaluation of the TB-RROC intervention was conducted using phenomenographic methods. Trial participants were purposively sampled, and in-depth interviews were conducted with patients and guardians in both arms of the trial. Key informant interviews and observations in the wards and community were performed. Thematic content analysis was used to derive analytical themes. RESULTS: Fourteen patients, 12 guardians and 9 key informants were interviewed. Three key themes relating to TB retreatment emerged: medical experiences (including symptoms, treatment, and HIV); the effects of the physical environment (conditions on the ward, disruption to daily routines and livelihoods); and trust (in other people, the community and in the health system). Experiences were affected by the nature of a person's prior role in their community and resulted in a range of emotional responses. Patients and guardians in the community benefited from better environment, social interactions and financial stability. Concerns were expressed about the potential for patients' health or relationships to be adversely affected in the community. These potential concerns were rarely realised. CONCLUSIONS: Guardian administered intramuscular injections were safe and well received. Community-based care offered many advantages over hospital-based care for patients receiving long-term injectable treatment for TB and their families.