An Unanswered Health Disparity: Tuberculosis Among Correctional Inmates,1993 Through 2003

Jessica R. MacNeil; Mark N. Lobato, MD; Marisa Moore, MD, MPH


Am J Public Health. 2005;95(10):1800-1805. 

In This Article


From 1993 through 2003, US jurisdictions reported 210976 cases of TB to the national TB surveillance system. Information about residence at a correctional facility at the time of TB diagnosis was reported for 208468 (98.8%) patients, of whom 7820 (3.8%) were inmates. The percentage of TB cases in males aged 15 to 64 years reported as residing in a correctional facility was 9.2% for those born in the United States and 4.2% for those not born in the United States (odds ratio [OR]=2.2; 95% confidence interval [CI]=2.08, 2.33; P<.001).

From 1993 through 2003, the percentage of TB cases among local jail inmates increased from 42.8% of all inmates with TB to 53.5% (χ2 for trend= 57.8; P<.001), whereas cases among federal inmates increased from 2.9% to 11.8% ( Table 1 ). Case rates for the 11 years studied were 29.4 per 100000 for federal prisons and 24.2 for state prisons. In contrast, federal prisons in 1993 and in 2003 had nearly level TB case rates. In state prisons, case rates decreased from 52.3 in 1993 to 6.6 in 2003, a decline of 87.4%.

Two states, California and Texas, accounted for 42.7% of the 7820 reported TB cases among inmates from 1993 through 2003, and another 4 states (Florida, Georgia, Illinois, and New York) accounted for an additional 28.6% of reported TB cases. These same 6 states accounted for 56.9% of the 200648 reported TB cases among noninmates.

The characteristics of individuals with TB in correctional facilities differed from those with TB who did not reside in correctional facilities: inmates with TB were more likely to be male (89.4% vs 61.9%, respectively), US born (77.0% vs 58.3%, respectively), younger (median: 37 vs 45 years of age, respectively), and from racial and ethnic minority groups (81.7% vs 75.5%, respectively). Over time, an increasing proportion of TB cases were among foreign-born persons in both inmate and noninmate populations. The proportion of foreign-born inmates with TB had increased from 15.5% in 1993 to 40.3% by 2003; similarly, 30.6% of those with TB in the general population were foreign-born in 1993, and 53.6% were foreign-born in 2003.

Excess alcohol use, injection drug use, noninjection drug use, and homelessness within 1 year prior to TB diagnosis in adult males aged 15 to 64 years were all more frequent in inmates with TB than in noninmates with TB ( Table 2 ). Additionally, inmates were more likely than noninmates to report at least 1 TB risk factor including HIV-infection (60.1% vs 42.0%, respectively).

Inmates with TB were also more likely than noninmates with TB to be HIV infected. From 1993 through 2003, HIV infection was documented in 35.8% of inmates with TB in state prisons, in 20.7% of those in jail, and in 13.2% of those in federal prisons. Overall, of males with TB aged 15 to 64 years, 25.2% who were inmates were known to be HIV infected versus 18.0% of those who were noninmates. A positive finding, however, is that HIV prevalence is declining in this setting. Among those with TB in state prisons, the prevalence of HIV infection decreased from 43.1% in 1993 to 11.6% in 2003. In federal prisons, the prevalence of HIV infection decreased from 23.8% in 1993 to 9.3% in 2003. In local jails in 1993, 22.8% of inmates with TB also were infected with HIV, whereas in 2003, 12.4% of inmates with TB were HIV infected.

A higher proportion of inmates (90.3%) than noninmates (84.4%) had pulmonary TB. Results of sputum smears for acid-fast bacilli and sputum cultures were reported more often for inmates than for noninmates. Inmates compared with noninmates were less likely to have extrapulmonary TB (OR=0.60; 95% CI=0.55, 0.66; P<.001).

Table 3 presents the frequency of drug resistance of M tuberculosis isolates among inmate and noninmate patients. In general, drug-resistance levels were higher in those with a prior history of TB, those not born in the United States, those with HIV infection, and inmates. From 1993 through 2003, declines in drug resistance were greater for inmates than for noninmates: isoniazid resistance decreased (inmates, from 10.9% to 6.7%; noninmates, from 9.9% to 8.8%), as did levels of multidrug-resistant TB (inmates, from 3.3% to 0.6%; noninmates, from 3.4% to 1.1%), and resistance to any drug (inmates, from 15.2% to 11.5%; noninmates, from 14.9% to 13.6%).

In 1993, among patients for whom extended treatment was not warranted, treatment was completed within 12 months in 47.9% of inmates compared to 60.4% of noninmates. In 2001, rates of completion of therapy had improved to 63.6% and 80.1%, respectively. Rates of completion of therapy within 12 months were lower in persons with TB risk factors and lowest for those who had HIV infection at the time of TB diagnosis, in both inmates and noninmates, but lower among inmates ( Table 4 ).

Inmates were more likely to receive directly observed therapy for at least part of their therapy than were noninmates ( Table 4 ); however, inmates were less likely to complete therapy. Lower completion rates among inmates compared with noninmates reflect higher levels of "incomplete treatment" categories (moved, lost, other, or unknown). A higher percentage of federal inmates (27.9%) were classified as "treatment incomplete" owing to a reported treatment outcome of "moved out of jurisdiction" compared with other inmates (9.0%) and noninmates (4.4%). In addition, 11.0% of local jail inmates had a reported treatment outcome of ," compared with 7.1% of other inmates and 3.9% of noninmates.