Setting: Tuberculosis (TB) treatment center at Coast Provincial General Hospital in Mombasa, Kenya.
Objectives: To describe TB management practices in a facility in coastal Kenya and identify factors associated with poor treatment outcomes.
Design: Retrospective review of patient treatment records from January 2008 to June 2009. Treatment outcomes of patients were classified as treatment success (cure or treatment completion) or poor treatment outcome (treatment failure, death or default). Relative risk regression was used to determine the association between exposures of interest and poor treatment outcomes.
Results: Records were obtained from a total of 183 patients: 142 (78%) had pulmonary TB, 68 (37%) were human immunodeficiency virus (HIV) infected and 81 (44%) had acid-fast bacilli (AFB) positive smear micros- copy. Most treated individuals (86%) achieved a successful treatment outcome as defined by the World Health Organization. Of those with poor treatment outcomes, 64% defaulted, 32% died, and 4% failed treatment. Initial negative AFB smear and HIV co-infection were associated with poor treatment outcomes (RR 3.32, 95%CI 1.22-8.99 and RR 4.61, 95%CI 1.69- 12.59, respectively).
Conclusion: Strategies to accelerate accurate diagnosis of smear-negative TB and increase patient retention during treatment, especially in HIV co-infected individuals, are needed to reduce poor treatment outcomes in Kenya.