Studies were included if they reported TB treatment outcomes for CB DOT compared to clinic DOT. There was no restriction on the type of TB (pulmonary, extra-pulmonary, drug-sensitive or drug-resistant) or method of diagnosis (microbiological or radiological with typical clinical presentation). We included RCTs and non-randomised studies (interventional and observational). Studies were included independent of data analysis method (intention-to-treat or other). CB DOT and clinic DOT had to be given for the same time period within a study and for a minimum of the intensive phase of treatment to be included.