The resistance of Mycobacterium tuberculosis to antibiotics, which reflects the quality of the chemotherapy applied in the community, is one of the elements of epidemiological surveillance used in national tuberculosis programmes. Measurement of drug resistance poses problems for biologists in standardization of laboratory methods and quality control. The definition of rates of acquired and primary drug resistance also necessitates standardization in the methods used to collect information transmitted by clinicians. Finally, the significance of the rates calculated depends on the choice of the patients sample on which sensitivity tests have been performed. National surveys of drug resistance therefore require multidisciplinary participation in order to select the only useful indicators: rates of primary resistance and of acquired resistance. These indicators, gathered in representative groups of patients over a long period, are a measurement of the impact of modern chemotherapy regimens on bacterial ecology.