The two important components of good TB control programme are early diagnosis and prompt institution of effective treatment. Operational research in Malawi between 1995-96 revealed that the guideline on TB suspects (chest symptomatics) were not followed by many health care staff as a result, there was a median delay of 4 months between onset of cough and diagnosis. Before sputum examination was performed, either the patient had visited traditional healer or he was not offered the sputum examination when he visited the centre. Intensive efforts were made from 1997 to overcome this problem to train health care staff to collect sputum specimens from patients with cough for more than three weeks on one hand and educate the general population and traditional healers about TB through mass media and conduct meetings on the other. One year later, it was felt essential to assess the health care seeking behaviour and diagnostic processes in patients with smear positive TB to assess the impact of the intervention measures undertaken. Government hospitals in five districts in Malawi were chosen as study sites.

During the study period between January to September 1998, 1518 patients were registered with pulmonary TB, of whom 1099 (72%) were interviewed using structured questionnaires. The median delay between onset of cough and diagnosis was 8 weeks. There was a variable pattern of care seeking behaviour, with 70% of patients initially visiting a place of health facility and 30% visiting traditional healers, grocery shops, etc. Of these, 867 (79%) patients made one or more visits for relief more often to medical care. At all stages, antibiotics resulted in symptomatic improvement in up to 40% of cases. There was a median time of 7 weeks between cough and first submission of sputum specimens. Almost all patients received sputum smear results after a median length of 4 days; 474 (43%) patients were only aware of their diagnosis at the time of receiving smear results. This observation being significantly associated with lack of schooling and not knowing another person with TB. Although the median delay between onset of cough and diagnosis was reduced from 4 months to 8 weeks, there is a need for more interactions between TB programme and other important stakeholders like non-orthodox care providers to improve case detection rates and reduce the delay in diagnosis. More needs to be done to educate community and include un-orthodox care providers for participation in the TB control by either carrying out diagnosis and treatment of TB patients or referring them to the programme.