Food incentives to improve completion of tuberculosis treatment: randomised controlled trial in Dili, Timor-Leste

Randomized controlled trial in adults with newly diagnosed pulmonary tuberculosis in Dili, Timor-Leste, recruited from three community clinics. The intervention arm included 137 randomized participants and 136 analyzed; the population was largely poor, malnourished, and untreated before diagnosis.

Adults receiving standard tuberculosis treatment were given culturally appropriate food support in addition to usual care. During the 8-week intensive phase, the intervention group received one bowl of feijuada at the clinic daily; during weeks 9-32, they received daily food parcels containing red kidney beans, rice, and oil, intended to provide one meal per day for one adult.

Food supplementation did not improve the primary outcome of tuberculosis treatment completion. Completion was 103/136 (76%) in the intervention group versus 100/129 (78%) in controls, RR 0.98 (0.86 to 1.11), P = 0.7. Intensive-phase adherence was actually lower with food support, 86.7% (18.6) versus 91.4% (13.3), P = 0.02. The intervention modestly improved weight gain at 8 weeks, 5.2% (6.2) versus 3.5% (6.3), P = 0.04, and at 32 weeks, 10.1% (9.6) versus 7.5% (7.6), P = 0.04. Itch was more common in the intervention group, 21% versus 9%, RR 2.27 (1.20 to 4.26), P = 0.008.

The trial was conducted in one city among predominantly poor, malnourished adults with pulmonary tuberculosis, so generalizability to other settings is limited. Some subgroup analyses were small, and baseline BMI data were missing for 18 participants. The intervention improved weight gain but did not improve the main clinical endpoint of treatment completion.