The annual incidence of TB cases among Singapore residents fell steadily from 306 per 100,000 population in 1960 to 56/100,000 in 1987 but has since remained at between 50 and 55/100,000. One of the possible reasons for this non-decline may be persistence of transmission of TB in the community due to delayed diagnosis, treatment and ineffective case holding.

Compared to non-defaulting patients as controls, defaulters were mostly non-Chinese, and those live on their own or with friends. There was no significant association of defaulting with age, sex, marital or employment status, disease characteristics, or treatment-related factors. Seventy per cent defaulted during the continuation phase of treatment.

The study was a retrospective patient record based case control study conducted in the TB Control Unit (TBCU), Singapore. This being the main treatment centre, which treats about 50% of the cases was the venue of the study. The objectives were to: (i) identify any demographic, social, disease or treatment-related characteristics which may be predictive of patients defaulting from treatment; (ii) assess the effectiveness of home visits as a means of defaulter recall; and (iii) ascertain outcome in these patients. TB treatment defaulters were defined as the patients who missed their scheduled appointments and required a home visit to recall for treatment. Equal number of controls were randomly selected from non-defaulting patients who started treatment on the same dates as the defaulters. Majority of the patients were supplied drugs for self-administration at home and there were about 10% of the patients who were on DOTS during the study period.

Of the 44 treatment defaulters, 6 (13.6%) were contacted directly, 20 (45.5%) through a person at home during the visit and for 18 (40.9%) a recall letter was slipped through the door due to no contact with patient or any other person at home. Following home visits, 20 (45.5%) returned within 7 days. The treatment outcome was not very encouraging as only 19 (43.2%) completed treatment, 21 (47.7%) were not traceable, 1 was dead and 3 were hospitalized. However, of the 21 patients who were lost to follow-up, all except one had culture negative results. The study identifies the future prediction of default as those who were non-Chinese, living alone, male and had a previous history of treatment.