• Michael C Tan, Carlo A Marra, Mohsen Sadatsafavi, Fawziah Marra, Onofre Morán-Mendoza, Susanne Moadebi, R Kevin Elwood, and J Mark FitzGerald.
• Faculty of Medicine, University of British Columbia, Vancouver, BC, Canada.
• Value Health. 2008 Sep 1; 11 (5): 842-52.

BackgroundContacts of patients with active tuberculosis ("TB contacts") with a tuberculin skin test (TST) size > or = 5 mm are currently recommended treatment for latent TB infection (LTBI). Knowing the cost-effectiveness of LTBI therapy for specific TB contact subpopulations may improve the use of limited resources by reducing the treatment of persons at low TB risk.ObjectiveTo evaluate the cost-effectiveness of LTBI therapy for different TB contact populations defined by important risk factors, and to propose an optimal policy based on different recommendation for each subgroup of contacts.MethodsA 6-year Markov decision analytic model simulating the quality-adjusted life years (QALYs), number of active TB cases prevented, and costs for hypothetical cohorts of Canadian TB contacts defined by TST size, age group (< 10 y/o or above), ethnicity, closeness of contact, and Bacillus Calmette-Guérin (BCG) vaccination status.ResultsFor the majority of subgroups, the current policy of preventive therapy in those with positive TST was the most cost-effective. Nevertheless, our analysis determined that LTBI treatment is not cost-effective in nonhousehold Canadian-born (nonaboriginal) or foreign-born contacts age > or = 10 y/o. On the other hand, empirical treatment without screening of all non-BCG-vaccinated household contacts age < 10 y/o appeared cost-effective. Such an optimal approach would result in an incremental net monetary benefit of $25 for each contact investigated for a willingness-to-pay of $50,000/QALY. Results were robust to several alternative assumptions considered in sensitivity analyses.ConclusionsThe current practice of LTBI treatment for TB contacts with a TST size > or = 5 mm is cost-effective. A customized approach based on excluding low risk groups from screening and providing treatment to high risk contacts without screening could improve the performance of the program.

31 keywords...

hide…