METHODS Forty hospitals in Malawi (3 central, 22 district and 15 mission) performing smear microscopy and registering tuberculosis patients. OBJECTIVE To determine, in patients aged 15 years or above, 1) the proportion with smear-negative pulmonary tuberculosis (PTB) who had sputum smears examined, 2) the number of sputum smears examined per patient, and 3) the proportion of patients registered with smear-positive and smear-negative PTB. METHODS Data collection during three 6-month periods, from January 1997 to June 1998, using tuberculosis registers, laboratory sputum registers and quarterly reports. RESULTS Of 6301 smear-negative PTB patients, 84% had sputum smears examined, the rate increasing from 76% in January-June 1997, to 85% in July-December 1997, to 89% in January-June 1998. Of patients who submitted sputum (where the number of smears was recorded), 99% had two or more smears examined and 93% had three smears examined. In district and mission hospitals performance improved over time, while in central hospitals results were more variable. During the same 18-month period 21 422 patients aged 15 years or more were registered with PTB: 59% with smear-positive PTB and 41% with smear-negative PTB; this pattern was similar in each 6-month period. CONCLUSIONS The study suggests that it is reasonable to aim for a target of 90% or more of smear-negative PTB patients having sputum smears examined.