Spatial distribution of tuberculosis incidence in Los Angeles County

Abstract

Background: In Los Angeles County, the tuberculosis (TB) disease incidence rate is seven times higher among non-U.S.-born persons than U.S.-born persons and varies by country of birth. But translating these findings into public health action requires more granular information, especially considering that Los Angeles County is more than 4000 mile2. Local public health authorities may benefit from data on which areas of the county are most affected, yet these data remain largely unreported in part because of limitations of sparse data. We aimed to describe the spatial distribution of TB disease incidence in Los Angeles County while addressing challenges arising from sparse data and accounting for known cofactors. Methods: Data on 5447 TB cases from Los Angeles County were combined with stratified population estimates available from the 2005-2011 Public Use Microdata Survey. TB disease incidence rates stratified by country of birth and Public Use Microdata Area were calculated and spatial smoothing was applied using a conditional autoregressive model. We used Bayesian Poisson models to investigate spatial patterns adjusting for age, sex, country of birth and years since initial arrival in the U.S. Results: There were notable differences in the crude and spatially-smoothed maps of TB disease rates for high-risk subgroups, namely persons born in Mexico, Vietnam or the Philippines. Spatially-smoothed maps showed areas of high incidence in downtown Los Angeles and surrounding areas for persons born in the Philippines or Vietnam. Areas of high incidence were more dispersed for persons born in Mexico. Adjusted models suggested that the spatial distribution of TB disease could not be fully explained using age, sex, country of birth and years since initial arrival. Conclusions: This study highlights areas of high TB incidence within Los Angeles County both for U.S.-born cases and for cases born in Mexico, Vietnam or the Philippines. It also highlights areas that had high incidence rates even when accounting for non-spatial error and country of birth, age, sex, and years since initial arrival in the U.S. Information on spatial distribution provided here complements other descriptions of local disease burden and may help focus ongoing efforts to scale up testing for TB infection and treatment among high-risk subgroups.