SETTING: All 44 non-private hospitals in Malawi treating tuberculosis (TB) cases in which oral regimens were used allowing patients during the initial phase to receive directly observed treatment (DOT) from health centres or guardians at home. OBJECTIVES: A country-wide audit of the oral regimens to determine: 1) TB ward bed occupancy rates, 2) patient DOT options, 3) patients' knowledge of treatment and 4) treatment outcomes compared to those obtained with previous treatment regimens. DESIGN: Retrospective data collection using registers and treatment cards. Prospective interviews with patients. Inspections of TB wards. RESULTS: There were 1513 TB beds occupied by 807 (53%) TB patients. Over 50% of 4793 patients registered with different types of TB chose guardian-based DOT. For 266 patients with pulmonary TB the correct knowledge about total duration of treatment (45%), all three DOT options (62%) and the months for giving follow-up sputum (16%), was poor. There were differences in treatment outcomes between TB patients on oral compared with previous regimens. With oral regimens, rates of unknown outcome were high. CONCLUSION: Oral treatment regimens are associated with reduced bed occupancy rates on TB wards. However, rates of unknown outcome are increased, and TB control is therefore weakened.