Background: Among U.S. residents, tuberculosis (TB) disease disproportionally affects non-U.S.-born persons and varies substantially by country of birth. Yet TB disease incidence rates by country of birth are not routinely reported despite these large, known health disparities. This is in part due to the technical challenges of using standard regression analysis with a communicable disease. Here, we estimate tuberculosis disease incidence rates by country of birth and demonstrate methods for overcoming these challenges using TB surveillance data from Los Angeles County which has more than 3.5 million non-U.S.-born residents.
Methods: Cross-sectional data on 5,447 cases of TB disease from Los Angeles County were combined with population estimates from the American Community Survey to calculate TB disease incidence rates for 2005 through 2011. Adjusted incidence rates were modelled using Poisson and negative binomial regressions. Bayesian models were used to account for the uncertainty in population estimates.
Results: The unadjusted incidence rate among non-U.S.-born persons was 15 per 100,000 person-years in contrast to the rate among U.S-born persons, 2 per 100,000. The unadjusted incidence rates were 44 and 12 per 100,000 person-years among persons born in the Philippines and Mexico, respectively. In adjusted analysis, persons born in the Philippines were 2.6 (95% CI: 2.3-3.1) times as likely to be reported as a TB case than persons born in Mexico. Bayesian models showed similar results.
Conclusion: This study confirms substantial disparities in TB disease by country of birth in Los Angeles County. Accounting for age, gender, years in residence and year of diagnosis, persons from the Philippines, Vietnam and several other countries had much higher rates of reported TB disease than other foreign countries. We demonstrated that incidence rates by country of birth can be estimated using available data despite technical challenges.