Martins 2009.

Methods	Randomized controlled trial
Participants	The study location was 3 community clinics in Deli, Timor Leste 137 participants were randomized to the intervention group and 133 participants were randomized to the control group The inclusion criteria were patients with newly diagnosed pulmonary tuberculosis, 18 years of age or older, and who agreed to receive tuberculosis treatment at the clinic for 8 months The exclusion criteria were pregnancy and previous treatment for tuberculosis for more than 1 month
Interventions	Intervention: FOOD INCENTIVE The intervention was a food incentive program in which the intervention group patients were provided with midday meals every time they attended the clinic. Patients were to report to the clinic 5 or 6 mornings a week to receive directly observed treatment in the intensive phase and fortnightly in the continuation phase Control: NUTRITIONAL ADVICE Control participants received usual care. They were also given nutritional advice (verbal and written) about the types of locally available food that would constitute a balanced diet and would be likely to assist cure of tuberculosis
Outcomes	The measures of adherence were clinical attendance, directly observed treatment, interview, and pill counts. Clinic attendance and direct observation were performed daily in the intensive phase and fortnightly in the continuation phase while interview and pill counts were performed in the continuation phase The patient outcomes were treatment completion (in terms of either completion of 8 months or treatment or clearance of acid fast bacilli from the sputum) and clinical improvement (in terms of improvement in weight gain, cough clearance, and adverse events). Weight gain was measured at 8 and 32 weeks while cough clearance was measured at 4, 8, and 32 weeks
Notes	The authors note that "...the civil conflict dramatically increased the default rate of those patients (68,25%) still receiving treatment after 28 April 2006 and led to a significant decrease in treatment completion (168/199 (84%) before conflict vs 35/66 (53%) after conflict; 1.58, 1.26 to 2.01; P<0.001)."
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computerized randomization was used. "At each study site a research assistant randomly allocated study participants to the intervention group (supplementary food) or control group (nutritional advice). An independent statistician computer generated a random allocation sequence with randomly varying block sizes in Stata (version 8)." (pg 2)
Allocation concealment (selection bias)	Low risk	"The sequence was concealed from all investigators with sequentially numbered opaque sealed envelopes prepared distant from the study site. Allocation was stratified by community health clinic and by diagnosis of tuberculosis (smear positive and smear negative). Both participants and treatment providers were aware of an individual's allocation status after randomisation." (pg 2)
Selective reporting (reporting bias)	Low risk	No protocol available; although it appears that everything was reported it is difficult to determine this without a protocol
Other bias	Low risk	The study seems to be free of other types of bias
Blinding of outcome assessment (detection bias) Adherence measure	Unclear risk	(PRIMARY) COMPOSITE ADHERENCE MEASURES ‐ No explicit information was given regarding who collected the data and how. There is insufficient information to permit judgment of 'Low risk' or 'High risk'
Blinding of outcome assessment (detection bias) Patient outcome	Low risk	(PRIMARY) TREATMENT COMPLETION ‐ "An independent observer (PM, based in Darwin), who was blinded to the intervention received by the patients, however, determined the primary outcome (treatment completion)." (pg 2)
Blinding of participants (performance bias) Adherence measure	Unclear risk	(PRIMARY) COMPOSITE ADHERENCE MEASURES ‐ Patients were aware of allocation after randomization; not enough details are given about the methods for this measure to judge
Blinding of participants (performance bias) Patient outcome	Low risk	(PRIMARY) TREATMENT COMPLETION ‐ Outcome is unlikely to be affected by patient's lack of blinding
Blinding of personnel (performance bias) Adherence measure	Unclear risk	(PRIMARY) COMPOSITE ADHERENCE MEASURES ‐ Providers were unblinded. No information on how the data were converted into a % adherent. (pg 2) "Both participants and treatment providers were aware of an individual's allocation status after randomisation."
Blinding of personnel (performance bias) Patient outcome	Unclear risk	(PRIMARY) TREATMENT COMPLETION ‐ Providers and patients were aware of allocation after randomization but it is unclear whether this would have an effect on this measure
Incomplete outcome data (attrition bias) Adherence measure	Unclear risk	(PRIMARY) COMPOSITE ADHERENCE MEASURES ‐ There is a high rate of data loss for the measure. ITT analysis was done; it is not clear if any imputation method was employed
Incomplete outcome data (attrition bias) Patient outcome	Low risk	(PRIMARY) TREATMENT COMPLETION ‐ Few data points missing for this outcome