Optimal Targeting of Individual and Population Level TB Prevention
|Project Period||December 2019 to November 2024|
|Principal Investigator||Nick Menzies|
|Prime Institution||Harvard T.H. Chan School of Public Health|
|Location of Interest||United States|
Public Health Relevance: This research is intended to provide individually- and locally-tailored evidence on TB risks, to optimize TB prevention services in the United States and internationally. We will conduct analyses of empirical data, and develop and train mathematical models to simulate individual and population level TB outcomes. Based on this evidence, we will develop web-based interfaces predicting the health benefits, risks, and costs of LTBI screening and treatment, and the population-level effects of user-defined TB control strategies.
Description: Within the same community, TB risks can differ by several orders of magnitude due to differences in infectious exposure and immune competence, and TB control depends heavily on targeting services to those most at risk. Priority groups described by the CDC and other agencies capture major TB risk factors, but these broad categories include many individuals with low TB risk, and exclude others who would benefit from screening. Our long-term objective is to provide individually- and locally-tailored evidence on TB risks and intervention effects, to optimize TB prevention services.
In prior work we have demonstrated the feasibility of estimating TB risks for small population groups, and in Aim 1 we will create granular estimates of TB risk for the US population, via a Bayesian evidence synthesis combining time series data on TB cases and population size, prevalence of latent infection (LTBI), and the fraction of cases due to recent infection. This analysis will allow us to produce individually-tailored risk predictions to better target preventive services, and provide patients with quantitative information on the risks they face. The number of patients to whom this applies is substantial—approximately half of all US residents have been tested for LTBI, and of those testing positive only half initiate treatment. This represents a large number of people facing decisions about LTBI testing and treatment.
Aim 2 will directly address these questions, creating highly-disaggregated estimates of the costs, harms, and benefits of LTBI testing and treatment. To do so we will construct a Markov microsimulation model of LTBI screening and treatment. Using this model we will estimate long-term patient-level outcomes, including changes in TB risk, survival, costs, and adverse events. Based on these analyses we will develop a user-friendly web tool to provide patients and clinicians prompt, validated, and individually-tailored information on possible treatment outcomes. We will also conduct analyses and develop a companion tool that will report the impact and cost-effectiveness of LTBI screening for user-defined target groups for the purpose of guiding program decision-making. To increase the reach and impact of these tools we will adapt them for other countries with TB incidence below 20 per 100,000.
In Aim 3 we will develop a transmission-dynamic simulation model to predict long-term outcomes for a broad set of TB control options (including but not limited to LTBI treatment) and risk factor trends. The model will be calibrated for multiple jurisdictions, and a web-based interface will allow users to specify scenarios and visualize outcomes. By identifying how current and novel interventions can be most effectively deployed to improve health, this research addresses the NIH’s highest priority area of health economics research, and responds directly to the need for computational tools and models to better understand and respond to infectious disease risks.