Rationale: Mathematical modeling is used to understand disease dynamics, forecast trends, and inform public health prioritization. We conducted a comparative analysis of tuberculosis (TB) epidemiology and potential intervention effects in California, using three previously developed epidemiologic models of TB.Objectives: To compare the influence of various modeling methods and assumptions on epidemiologic projections of domestic latent TB infection (LTBI) control interventions in California.Methods: We compared model results between 2005 and 2050 under a base-case scenario representing current TB services and alternative scenarios including: 1) sustained interruption of Mycobacterium tuberculosis (Mtb) transmission, 2) sustained resolution of LTBI and TB prior to entry of new residents, and 3) one-time targeted testing and treatment of LTBI among 25% of non-U.S.-born individuals residing in California.Measurements and Main Results: Model estimates of TB cases and deaths in California were in close agreement over the historical period but diverged for LTBI prevalence and new Mtb infections-outcomes for which definitive data are unavailable. Between 2018 and 2050, models projected average annual declines of 0.58-1.42% in TB cases, without additional interventions. A one-time LTBI testing and treatment intervention among non-U.S.-born residents was projected to produce sustained reductions in TB incidence. Models found prevalent Mtb infection and migration to be more significant drivers of future TB incidence than local transmission.Conclusions: All models projected a stagnation in the decline of TB incidence, highlighting the need for additional interventions including greater access to LTBI diagnosis and treatment for non-U.S.-born individuals. Differences in model results reflect gaps in historical data and uncertainty in the trends of key parameters, demonstrating the need for high-quality, up-to-date data on TB determinants and outcomes.

Introduction

Preventing tuberculosis (TB) disease requires treatment of latent TB infection (LTBI) as well as prevention of person-to-person transmission. We estimated the LTBI prevalence for the entire United States and for each state by medical risk factors, age, and race/ethnicity, both in the total population and stratified by nativity.

Methods

We created a mathematical model using all incident TB disease cases during 2013-2017 reported to the National Tuberculosis Surveillance System that were classified using genotype-based methods or imputation as not attributed to recent TB transmission. Using the annual average number of TB cases among US-born and non-US-born persons by medical risk factor, age group, and race/ethnicity, we applied population-specific reactivation rates (and corresponding 95% confidence intervals [CI]) to back-calculate the estimated prevalence of untreated LTBI in each population for the United States and for each of the 50 states and the District of Columbia in 2015.

Results

We estimated that 2.7% (CI: 2.6%-2.8%) of the U.S. population, or 8.6 (CI: 8.3-8.8) million people, were living with LTBI in 2015. Estimated LTBI prevalence among US-born persons was 1.0% (CI: 1.0%-1.1%) and among non-US-born persons was 13.9% (CI: 13.5%-14.3%). Among US-born persons, the highest LTBI prevalence was in persons aged ≥65 years (2.1%) and in persons of non-Hispanic Black race/ethnicity (3.1%). Among non-US-born persons, the highest LTBI prevalence was estimated in persons aged 45-64 years (16.3%) and persons of Asian and other racial/ethnic groups (19.1%).

Conclusions

Our estimations of the prevalence of LTBI by medical risk factors and demographic characteristics for each state could facilitate planning for testing and treatment interventions to eliminate TB in the United States. Our back-calculation method feasibly estimates untreated LTBI prevalence and can be updated using future TB disease case counts at the state or national level.