Getting the best possible evidence from observational studies of treatment of multidrug-resistant tuberculosis patients

Background: The development of clinical recommendations by the World Health Organization (WHO) requires the systematic review of evidence and the grading of its quality. There is no high-quality evidence on treatment for multidrug-resistant tuberculosis (MDR-TB) patients using only standard second-line anti-TB drugs, as results of randomized controlled trials are not available. Well executed observational studies can meanwhile provide useful information that is applicable to clinical practice. For observational data to provide the best possible guidance, measures need to be taken to limit bias, imprecision, and heterogeneity. Intervention: WHO has established an expert Working Group to develop principles that national TB control programmes and other implementers could use to ensure the quality of observational study data. Results: Based on past experience, the group identified a number of critical points that need to be improved upon. Patient and programme-related data should be captured and reported in a standardized manner. These include bacteriological endpoints, proxies of disease severity, information on adverse drug reactions, the duration of current and past use of individual drugs, reason for change of regimen, use of accompanying medications, costs, hospitalization, patient support and type and extent of surgery. Conclusions: The Working Group recommends that programmes undertaking observational studies record most of these parameters, that they register data electronically, that they strive to publish their work, and that they share anonymous, individual patient data so as to update the knowledge base on MDR-TB patient outcomes.