Monthly culture is usually recommended to monitor treatment of multidrug-resistant tuberculosis (MDR-TB). As mycobacterial laboratory capacity is limited in many settings, TB programs need evidence to decide whether monthly cultures are necessary compared to other approaches. We simulated three alternative monitoring strategies (culture every 2 or 3 months, and monthly smears alone) in a cohort of MDR-TB patients in Estonia, Latvia, Philippines, Russia and Peru from 2000 to 2004. This retrospective analysis illustrated that less frequent testing delays confirmation of bacteriological conversion. This would prolong intensive treatment, hospitalization and respiratory isolation, increasing cost and toxicity. After conversion, less frequent testing could delay diagnosis of possible treatment failure.