RESUMO
The study was conducted in 4 hospital laboratories to assess the value of smear re-examination, duplicate smear and culture in quality control in the diagnosis of pulmonary tuberculosis. In each hospital 1 to 3 sputum specimens were collected from each suspect and examined by the Ziehl-Neelsen method. Smears along with corresponding specimens were sent to the reference laboratory for re-examination and examination of fresh duplicate smears and culture. A total of 994 specimens plus 23 known negative and 23 known positive controls were analysed. For each assessment method, both over-diagnosis and underdiagnosis were indicated with under-diagnosis being a more common problem in the four hospital laboratories. The limitations of these assessment methods are presented.