Evaluating the Impact of Strategies for Tuberculosis Prevention and Control in High-Burden, Low-Resource Settings: Data for Evidence-Based Decision-Making in Local Contexts
Embargo until
Date
2015-10-29
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Johns Hopkins University
Abstract
Background: While incidence rates of tuberculosis (TB) are on the decline globally, the TB burden in sub-Saharan Africa and Southeast Asia remains high. If the goal of reducing the global TB incidence rate to < 10/100,000 population per year is to be achieved by 2035, additional TB control interventions will need to be deployed in high burden settings. Research is needed to identify effective, efficient interventions that prevent additional TB cases, identify and properly diagnose incident cases, as well as to provide timely, appropriate treatment and ensure treatment completion. We sought to evaluate several TB control interventions as implemented in local contexts, including a household contact tracing in rural South Africa, a cost-effectiveness analysis of interferon-γ release assays (IGRAs) in India, and predictors of isoniazid preventive therapy (IPT) completion in rural Malawi. This research aims to provide data for policy-makers and government officials tasked with the deployment of scarce TB control resources in local contexts, with the goal of identifying strategies to integrate TB case finding and prevention activities into programs with limited resources.
Methods: We recruited 130 newly diagnosed TB patients (“index cases”) from public clinics in Vhembe District, Limpopo Province, South Africa, and visited their homes to test their household contacts for TB via sputum smear microscopy and culture. Clinical and demographic characteristics, including HIV status, were assessed via self-report. We calculated the yield of previously undiagnosed TB disease among household contacts (defined as the number of new TB cases identified for every 100 index cases traced) and evaluated risk factors for TB disease among household contacts using multilevel logistic regression.
Next we evaluated the incremental cost-effectiveness of IGRAs compared to a base-case scenario of empirical diagnosis (without microbiological testing), as well as sputum smear microscopy and Xpert MTB/RIF. We performed our analysis from the perspective of the Indian TB Control Program, and evaluated the cost, disability-adjusted life years, deaths, and secondary cases averted, and false positive diagnoses resulting from the use of these diagnostics in a hypothetical cohort of 1 million adult Indian TB suspects. We performed one-way sensitivity analyses, as well as a probabilistic sensitivity analysis to generate uncertainty ranges around our estimates.
Finally we evaluated factors associated with IPT completion in a cohort of 974 newly diagnosed adult HIV patients in Malawi who were started on IPT after active TB disease was excluded. Participants were recruited as part of a larger cluster randomized trial of TB screening being conducted in 12 clinics across rural Malawi. IPT completion was defined as receipt of ≥150 doses of isoniazid during routine clinical visits. We assessed factors associated with IPT completion using a multilevel logistic regression model adjusted for patient clinical and demographic characteristics.
Results: A total of 282 household contacts were enrolled in our household contact tracing study between December 1, 2013 and September 30, 2014. A total of 11 individuals tested positive for TB disease, for a household TB disease prevalence of 3.9% (95% CI: 2.0-6.9%) and a yield of 8.5 cases per 100 index cases traced (95% CI: 4.2-15.1). The majority of TB cases identified by the study (10/11, 90.9%) were smear-negative/culture-positive. The presence of TB symptoms was not significantly associated with increased odds of active TB disease in our population (aOR: 0.3, 95% CI: 0.1-1.4).
Our cost-effectiveness analysis found that IGRAs were less cost-effective than sputum smear microscopy or Xpert MTB/RIF when diagnosing active TB in India. This was largely due to the poor specificity of IGRAs for active TB in a setting with high background rates of latent TB infection. Relative to sputum smear microscopy, IGRAs resulted in 315,700 (95% uncertainty range [UR]: 118,300 – 388,400) false-positive TB diagnoses, at an incremental cost of US$49.3 million (95% UR: $34.9 - $58.0 million) per 1 million population tested. Relative to Xpert MTB/RIF (including the cost of treating drug resistant TB), IGRAs averted 70,400 (95% UR: [-7,900] – 247,200) fewer disability adjusted life years and cost US$14.6 million (95% UR: [-$7.2] - $28.7 million) more.
In Malawi, 732 of the 974 (75%) individuals who started IPT completed their course of therapy. Individuals completing IPT were significantly older than non-completers (34 vs. 31, p<0.01) and less likely to have experienced an interruption of >2 months (7.1% vs. 80%, p=0.01). After controlling for potential confounders, participants younger than 25 years (compared to those over 45 years, aOR: 0.33, 95% CI: 0.18-0.60) and males (compared to non-pregnant women, aOR: 0.57, 95% CI: 0.37-0.88) had significantly lower odds of IPT completion. Concomitant receipt of ART drugs, being a current or former smoker, and self-reported alcohol use were not significantly associated with IPT completion in our study.
Discussion: Identification of effective and cost-effective interventions operationalizing case finding and prevention of TB will be vital in controlling TB and meeting ambitious global targets by 2035, especially in high-burden settings. We evaluated potential prevention and case-finding interventions in local settings, providing data useful to TB control programs and governments in sub-Saharan Africa and Southeast Asia, where high TB burdens and scarce resources present substantial challenges to meeting global TB control targets.
Description
Keywords
Tuberculosis, case finding, operational research, cost effectiveness, IGRA, HIV, isoniazid, IPT