SETTING: All 44 non-private hospitals (four central, 22 district and 18 mission) in Malawi that registered and treated tuberculosis (TB) cases, October-December 2001. OBJECTIVES: To determine, in new smear-positive pulmonary tuberculosis (PTB) patients, for the 2-, 5- and 7-month smear examinations, 1) the proportion with smears examined and 2) the actual timing of smear examination. STUDY DESIGN: Retrospective data collection using TB registers, TB treatment cards and laboratory sputum registers. Timing of smear examinations was judged acceptable if 2-month smears were examined at 2 or 3 months, 5-month smears at 4, 5 or 6 months and 7-month smears at 6, 7, 8 or 9 months. RESULTS: Of 1994 patients, for those alive and on treatment, 78% had smears definitely examined at 2 months, 75% at 5 months and 74% at 7 months. Of these, 82% had smears examined at an acceptable time for the 2-month smear, 71% for the 5-month smear and 78% for the 7-month smear. Smears were examined after the 8-month treatment regimen for the 2- and 5-month smear in respectively 2% and 9% of patients. Smears were done more frequently in female than male patients, and in district/mission hospitals than central hospitals. Smears were done at acceptable times more frequently in younger than older patients and in mission/central hospitals than district hospitals. CONCLUSION: During supervision, the actual time of follow-up sputum smear examinations needs to be monitored more closely.