Treatment completion continues to be an unmet challenge in tuberculosis (TB) control worldwide. Since the introduction of DOTS by the World Health Organization in 1993, directly observed therapy (DOT) has been promoted as the centerpiece of the strategy. The results of three randomized controlled studies designed to evaluate the effects of DOT on treatment completion, and carried out in South Africa, Thailand and Pakistan, showed conflicting results. It is doubtful that any of the studies provided valid measures of the effectiveness of DOT. Each study operationalized DOT in ways that did not match the study definition and failed to account for additional study interventions in their analyses. Given the evidence, there is no one factor that is 'the' answer to the challenge of TB treatment completion.