Tuberculosis patients in primary care do not start treatment. What role do health system delays play?

SETTING: Primary health care facilities in five provinces of South Africa. OBJECTIVE: To investigate the association between the proportion of sputum results with a prolonged smear turnaround time and the proportion of smear-positive tuberculosis (TB) cases initially lost to follow-up. DESIGN: The unit of investigation was a primary health care facility and the outcome was the initial loss to follow-up rate per facility, which was calculated by comparing the sputum register with the TB treatment register. A prolonged turnaround time was defined as more than 48 h from when the sputum sample was documented in the sputum register to receipt of the result at the facility. RESULTS: The mean initial loss to follow-up rate was 25% (95%CI 22-28). Smear turnaround time overall was inversely associated with initial loss to follow-up (P = 0.008), when comparing Category 2 (33-66% turnaround time within 48 h) with Category 1 (0-32%) (OR 0.73, 95%CI 0.48-1.13, P = 0.163) and when comparing Category 3 (67-100%) with Category 1 (OR 0.62, 95%CI 0.39-0.99, P = 0.045). The population preventable fraction of initial loss to follow-up (when turnaround time was <48 h in >= 67% of smear results) was 21%. CONCLUSION: Initial loss to follow-up should be reported as part of the TB programme to ensure that patients are initiated on treatment to prevent transmission within communities.