Background: Few studies have examined predictors of latent tuberculosis infection (LTBI) treatment completion in inner city populations in the United States.
Objective: To assess LTBI treatment completion rates and predictors in an inner city cohort.
Methods: Data from control groups of two sequentially conducted randomized controlled trials of LTBI treatment were analyzed for treatment completion rates. Participants in Study A (n = 191), conducted in 1996-1999, self administered daily isoniazid (INH) for 6-12 months, while participants in Study B (n = 123), conducted in 2002-2005, self administered daily INH for 9 months.
Results: Overall, 44.6% of participants completed therapy, with significantly higher completion rates in Study B than Study A (37.0% vs. 56.1%, P = 0.001). Marriage and alcohol use were significant predictors of completion (aOR = 2.153, 95%CI 1.301-3.562) and non-completion (aOR = 0.530, 95%CI 0.320-0.877), respectively; multivariate analysis indicated increased completion among married persons of foreign birth and among alcohol users who were homeless. Knowledge of and attitudes to tuberculosis were not significant predictors.
Conclusions: The design provided an opportunity to assess predictors of LTBI treatment completion in this inner city population. Social circumstances were the strongest predictors of treatment completion, suggesting that tangible social services may be more effective than educational programs in encouraging treatment completion.