SETTING: Laboratorio de Investigación de Enfermedades Infecciosas, Universidad Peruana Cayetano Heredia (UPCH) and government health centres, Lima, Peru. OBJECTIVE: To evaluate the contribution of unselected (courtesy) microscopic observation drug susceptibility (MODS) testing to the diagnosis and/or drug susceptibility testing (DST) of tuberculosis and their subsequent impact upon patient management. DESIGN: Retrospective database analysis and case note review of MODS culture-positive cases. RESULTS: Mycobacterium tuberculosis was isolated in 28.9% of 225 samples (209 patients); 22.2% of 63 positive cases were multidrug-resistant. In 58 MODS culture-positive cases with follow-up data available, MODS provided culture confirmation of diagnosis, DST or both in 82.8%, before any standard method. In 41.4%, this result should have prompted a modification in patient management. Delays between laboratory result and initiation or change of treatment, where applicable, took on average 42 and 64 days, respectively, of which a delay of respectively 17 and 48 days occurred after the receipt of results by the health facility. CONCLUSION: MODS provides important data for clinical management within a meaningful timeframe and should contribute positively to patient outcomes due to earlier initiation of appropriate therapy. Although clinicians may successfully select patients likely to benefit from MODS, ongoing work is required to identify optimal implementation of the assay and to reduce logistical and health system derived delays.