Gustavo Diaz, PhD


GHES LMIC Fellow 2019-2020

FELLOWSHIP SITE: Centro Internacional de Entrenamiento e Investigaciones Medicas (CIDEIM), Cali, Colombia
U.S. INSTITUTION: Yale University

Project Title: Improving the yield of tuberculosis contact investigation in Cali, Colombia

The incidence of tuberculosis (TB) in Cali was 48 per 100,000 in 2017. TB is concentrated in low-income neighborhoods, where incidence rates of ~165 per 100,000 are comparable with high-burden countries. Household TB contact investigation is an evidence-based strategy for early detection and treatment of undiagnosed TB patients. We recently reviewed program data on all household contact investigations performed in Cali in 2017. We found an overall diagnostic yield of contact investigation of 0.4% (7/1,864), which is more than 80% below other low and middle-income countries. In a follow-up analysis, we found multiple steps in the evaluation process where contacts are lost. For instance, only 35% (191/541) of referred contacts completed evaluation for TB in clinics. To understand the causes and evaluate novel solutions for contact investigation in Cali, we propose two specific aims. First, to identify barriers to delivery of household contact investigation. We will conduct focus group discussions with stakeholders from the Cali Health Secretariat, including TB program coordinators and front-line health workers involved in TB contact investigation. We will also carry-out key informant interviews with contacts and index-patients. Second, we aim to evaluate the feasibility of molecular testing for TB diagnosis to improve TB detection among contacts. We will introduce sputum Xpert-Ultra testing for TB and conduct an exploratory study of its incremental yield over microscopy in 200 household contacts (samples will also be evaluated by standard culture). We will also collect data on time savings, incremental costs, and feasibility of sampling and testing by frontline health workers. We hypothesize that home sputum collection and adding this highly sensitive molecular test to the contact investigation workflow will increase the detection of contacts with low bacillary load in sputum. With the gathered information, we plan to prototype a “home-based strategy” to address barriers to household contact investigation.