The primary data source for this analysis was the National Tuberculosis Surveillance System (NTSS), which receives reports of verified cases of TB from health departments. NTSS collects deidentified demographic, TB risk factor, and clinical information on all TB cases in the United States.^2,10

We analyzed NTSS data for all verified TB cases reported during 1993–2014 by the 50 U.S. states and the District of Columbia among persons who were residents of a correctional facility at the time their diagnostic evaluation for TB was initiated. The type of facility is also recorded: local jail, state prison, federal prison, juvenile detention facility, or a category called other facility type. The other facility type is a broad category that includes not only ICE detention centers but also other incarceration settings such as Indian reservation or military facilities.¹⁰ Additionally, the NTSS records (beginning with cases reported in 2009) whether an incarcerated TB patient was in ICE custody at the time of diagnosis, regardless of the correctly facility type.¹⁰

For this retrospective cross-sectional analysis of existing NTSS data, we grouped patients into two broad categories based on self-reported ethnicity — Hispanic ethnicity or non-Hispanic ethnicity — without regard to race (e.g., white, black). Because the epidemiology of TB in the United States differs markedly between the U.S.- and foreign-born,^2,11–12 we stratified our analysis by these 2 groups. We compared the prevalence of demographic and clinical characteristics, as well as the social risk factors that are routinely reported in the NTSS (ie, homelessness or substance use in the year leading up to TB diagnosis), between Hispanic and non-Hispanic incarcerated TB patients.

We calculated odds ratios and their corresponding 95% confidence intervals for the individual associations of demographic, risk factor, and clinical characteristics with being a Hispanic person who receives a TB diagnosis while incarcerated. We present unadjusted odds ratios because multivariable logistic regression did not change individual odds ratio point estimates > 10%, suggesting adjustment for other characteristics was unnecessary. All analyses were conducted using SAS version 9.3 (SAS Institute, Cary, NC).

Data for this analysis were collected by local and state health departments as part of routine public health surveillance. This secondary analysis was not considered by the Centers for Disease Control and Prevention to constitute research involving human subjects.

During 1993–2014, a total of 12,847 TB cases among persons who were incarcerated at diagnosis were reported in the United States. The Figure shows how the national proportion of TB cases that occurred in somebody incarcerated at TB diagnosis declined during 1993–2002 for both Hispanic and non-Hispanic populations. However, between 2003 and 2014, the annual proportion of Hispanic TB patients who were incarcerated at diagnosis grew from 4.9% to 8.4%, while the corresponding proportion among non-Hispanic TB patients continued to decline. Because of this trend, and preliminary analysis demonstrating no notable differences during 1993–2002 between Hispanic and non-Hispanic incarcerated TB patients (data not shown), we focused our subsequent analyses on the TB cases reported among incarcerated persons during 2003–2014.

A total of 5,502 TB cases were reported among incarcerated persons during 2003–2014. After excluding the 21 patients with unknown birth origin, 11 with unknown race/ethnicity, and 7 with unknown age or who were ≤ 14 years old, 5,463 TB cases remained for the rest of the analysis.

During 2003–2014, 530 (19%) of the 2,813 incarcerated U.S.-born TB patients were Hispanic, and 2,289 (86%) of the 2,650 foreign-born were Hispanic. The Table shows the characteristics of these 5,463 patients, stratified by birth origin and Hispanic ethnicity. Nearly all the 2,289 Hispanic foreign-born patients originated from one of four countries: Mexico (1,495 patients, or 65.3%), Honduras (306, or 13.4%), Guatemala (196, or 8.6%), or El Salvador (142, or 6.2%). Nearly half (44.0%) of the foreign-born Hispanic patients with a reported U.S. arrival date received a TB diagnosis within a year of arrival, in contrast to most of the other foreign-born who received a TB diagnosis after arriving in the United States > 2 years prior.

Regardless of birth origin or ethnicity, the preponderance of incarcerated persons to receive TB diagnoses were men between the ages of 15 and 44 who were in local jails. About a third of all foreign-born were in the other facility type category that includes ICE detention facilities. Our internal analysis with the new ICE custody variable introduced to the NTSS in 2009 suggests that for the majority of cases in this review, being in other facility type does indeed correspond to an ICE facility (i.e., during 2009–2014, 91% of TB patients in other facility type were also in ICE custody). Although a smaller proportion were in federal (non-ICE) prisons, Hispanic patients, regardless of birth origin, had twice the odds of being in federal prisons, compared with non-Hispanic incarcerated persons (Table).

Incarcerated Hispanic patients were generally younger than non-Hispanics. A large proportion of U.S.-born Hispanic patients had excess alcohol use (35.8%) and noninjection drug use (40.2%). Both U.S.-born and foreign-born Hispanic patients had a higher prevalence of injection drug use within the year before TB diagnosis. Foreign-born Hispanic patients had less substance use than either U.S.-born group but more than was reported among other the foreign-born (Table).