Latent Tuberculosis Infection Screening in Foreign-Born Populations: A Successful Mobile Clinic Outreach Model

Abstract

Objectives. We evaluated the efficacy of a mobile medical clinic (MMC) screening program for detecting latent tuberculosis infection (LTBI) and active tuberculosis.

Methods. A LTBI screening program in a MMC in New Haven, Connecticut, used medical surveys to examine risk factors and tuberculin skin test (TST) screening eligibility. We assessed clinically relevant correlates of total (prevalent; n = 4650) and newly diagnosed (incident; n = 4159) LTBI from 2003 to 2011.

Results. Among 8322 individuals, 4159 (55.6%) met TST screening eligibility criteria, of which 1325 (31.9%) had TST assessed. Similar to LTBI prevalence (16.8%; 779 of 4650), newly diagnosed LTBI (25.6%; 339 of 1325) was independently correlated with being foreign-born (adjusted odds ratio [AOR] = 8.49; 95% confidence interval [CI] = 5.54, 13.02), Hispanic (AOR = 3.12; 95% CI = 1.88, 5.20), Black (AOR = 2.16; 95% CI = 1.31, 3.55), employed (AOR = 1.61; 95% CI = 1.14, 2.28), and of increased age (AOR = 1.04; 95% CI = 1.02, 1.05). Unstable housing (AOR = 4.95; 95% CI = 3.43, 7.14) and marijuana use (AOR = 1.57; 95% CI = 1.05, 2.37) were significantly correlated with incident LTBI, and being male, heroin use, interpersonal violence, employment, not having health insurance, and not completing high school were significantly correlated with prevalent LTBI.

Conclusions. Screening for TST in MMCs successfully identifies high-risk foreign-born, Hispanic, working, and uninsured populations and innovatively identifies LTBI in urban settings.

Foreign-born populations are at greatest risk for having both latent tuberculosis infection (LTBI) and developing tuberculosis (TB) disease within high-income countries and, in 2012, accounted for 63.0% of the 9951 TB cases in the United States.¹ Newly diagnosed and reactivated TB infection among foreign-born individuals in the United States is currently 12 times greater (15.8 vs 1.4 cases per 100 000 population) than among US-born persons.¹ Among foreign-born individuals, LTBI often reactivates within 5 to 10 years after arrival to the United States.^2,3 Undocumented migrants and visitors from high-TB-prevalence countries, however, do not undergo routine LTBI screening and thus remain outside traditional health care screening and treatment programs in primary or specialty care settings except when they are acutely ill.^3,4 Thus, identifying and treating LTBI cases among these high-risk populations before transforming to TB disease and resultant transmission to others is crucial to ending the cycle of ongoing TB infection within the United States.

Workplace screening,^4,5 mandatory criminal justice system screening,^6–8 screening for entry into medication-assisted therapy and drug treatment programs,⁹ and refugee and naturalization programs^10,11 have been successful for reaching legal and domestic populations, but innovative options are needed to target foreign-born populations that are not yet integrated into mainstream care.

Culturally and geographically isolated foreign-born groups may be overlooked especially if there is low self-perception of tuberculosis risk.¹² Tuberculin skin testing (TST), though imperfect, is internationally recognized and has been shown to be a reasonably accurate assessment of LTBI status in immunocompetent adults, despite receiving previous Bacillus Calmette-Guérin vaccine.¹³ Whereas other studies have focused on traditional clinics or statewide programs,¹⁴ we present an innovative mobile medical clinic (MMC) as a model to target “hidden” foreign-born populations for LTBI screening.

New Haven, Connecticut, the country’s fourth poorest city for its size, with a census of 130 000, is a medium-sized urban setting in New England that has experienced extraordinary social and medical disparities including a high prevalence of poverty, drug addiction, HIV/AIDS, and unemployment and is disproportionately comprised of people of color, including 35.4% and 27.4% being Black or Hispanic, respectively.¹⁵ As New Haven is an industrial city with low-paying jobs, there has been an influx of foreign-born people, now officially comprising 11.6% of the population, with many having an undocumented residency status. Health care access for this group is absent unless individuals pay directly for fee-for-service, and concern for deportation and arrest further hinders willingness to seek care.¹⁶

The Community Health Care Van (CHCV) is an MMC that provides free health care 5 days per week in 4 impoverished neighborhoods in New Haven. Though at inception the program was linked to the needle and syringe exchange program,¹⁷ it has since expanded over 20 years to become a vital bridge to a diverse array of health and addiction treatment services that includes services for medically underserved populations, including directly administered antiretroviral therapy to treat HIV,^18–21 buprenorphine maintenance therapy,^22–25 community transitional programs from the criminal justice system,^26–33 hepatitis B vaccination,³⁴ rapid hepatitis C screening,³⁵ and other ongoing primary health care programs such as screening and monitoring of sexually transmitted infections,³⁶ diabetes, and hypertension. In addition, the CHCV provides outreach and intensive case management services.³⁷ Screening for LTBI and TB disease began in 2003 to target high-risk undocumented and foreign-born clients, as well as clients entering drug treatment programs or homeless shelters, who were concerned about TB infection yet were reluctant to seek care in traditional health care settings for fear of deportation, prohibitive cost, or language barriers. The LTBI screening program shortly thereafter became successfully incorporated into the country’s first mobile buprenorphine maintenance therapy program.⁹

METHODS

We selected 2 primary outcomes: (1) we defined total LTBI prevalence as all positive TSTs in the study population and included any past or present diagnosis of either TB disease or LTBI regardless of treatment status, and (2) we defined LTBI incidence as a subset of newly diagnosed LTBI as part of CHCV’s new TST screening program. The TST screening algorithms were consistent with the Centers for Disease Control and Prevention (CDC) recommendation of screening for high-risk and recent immigrant populations but also included health care providers and those entering drug treatment programs who had not been TST screened in the previous 12 months.¹¹ Total prevalent LTBI reflected the total number of persons identified with a past or current positive TST result on the CHCV as part of routine clinical care, and newly diagnosed LTBI (LTBI incidence) reflected the effectiveness of the CHCV’s public health–oriented LTBI screening program. For each of these outcomes, we examined the independent correlates by using standard measures used in routine clinical practice.

Study Population

All CHCV clients complete a standardized medical intake survey that assesses demographic characteristics, country of origin, past history and treatment of TB disease or LTBI, time periods of previous TST placement and results, screening chest x-rays (CXRs), and medical history including HIV, pulmonary disease, mental illness, and substance use. Undocumented status was assessed with a unique, confidential coding mechanism to protect clients. Patients with positive TST (defined as ≥ 10 mm for HIV-negative individuals and ≥ 5 mm for HIV-infected individuals) underwent World Health Organization (WHO) screening for symptoms³⁸ and were immediately referred for a CXR. As previously reported, patients with symptoms or abnormal CXR results underwent sputum culture for TB disease, and those identified with LTBI were offered 9 months of treatment of LTBI free of charge either on the CHCV or at a local TB clinic.³⁹

Patients eligible for TST screening included those who either self-reported no previous TST or who had a CDC-defined need to be screened more than 1 year previously. Ineligibility for TST screening included those screened within the past year, who reported and confirmed previous treatment of TB disease, or who had a documented previous positive TST (with or without isoniazid preventative therapy). Previous documented positive TST was confirmed either by MMC medical record or by patient history with chart review confirmation.

Analysis

We first extracted medical information without unique identifiers and available within the electronic medical record from January 2003 to July 2011 into an SPSS version 18 (SPSS Inc, Chicago, IL) database and then analyzed it with Stata IC version 12.1 (StataCorp LP, College Station, TX). We selected independent variables selected from the comprehensive medical intake for clinical relevance to LTBI. As previously defined,²⁷ patients with stable housing reported living in one’s own apartment, in one’s own house, or with one’s family.⁴⁰ We defined unstable housing as living in a shelter, in a halfway house, with friends, in a hotel, or in a public place. We classified being married or living together as a committed relationship. We defined no relationship status as being divorced, separated, widowed, or single. We defined foreign-born by self-report of country of birth being outside the United States or Puerto Rico. We assessed standardized measures of type, route, and frequency of drug use, including the 30-day time period before TST screening, by using the drug component of the Addiction Severity Index.⁴¹ We defined hazardous drinking as 4 or more drinks daily for men and 3 or more drinks daily for women; a “drink” was defined as equivalent to 12 ounces of beer or 1.5 ounces of hard liquor. We defined sex work as exchange of sexual intercourse for money, drugs, or protection, and we defined sex solicitation by self-report of paying money for sexual intercourse. We defined recent incarceration and recent emergency department visits by self-report if such an encounter occurred within the previous 6 months. Interpersonal violence included anyone reporting a history of either domestic violence or sexual assault.

We used the Wilcoxon rank sum test to test differences between persons with positive and negative TST results (Table 1), and we used the Kruskall–Wallis test to discern differences between groups representing newly diagnosed LTBI (newly identified clients with positive TSTs on the CHCV considered incident LTBI) and total LTBI prevalence (all clients with previous positive TST, including TB disease; data not shown). To not overfit the final multivariate regression model, we entered covariates on bivariate analyses for each of the 2 dependent variables significant at P < .05 into stepwise backward and forward multivariate logistic regressions with the final model selected on the basis of goodness of fit by using the Akaike information criterion. In the instances where the data were not missing at random (i.e., substance use), the nonresponse predominantly reflected a negative outcome (i.e., nonuse); in other instances where the missing mechanism was hard to discern, we initially tried multiple imputations but these ultimately did not influence the results.

TABLE 1—

Demographic, Clinical, and Risk Factor Characteristics of All Clients Receiving Tuberculin Skin Testing, Stratified by Positive or Negative Result: Community Health Care Van Mobile Medical Clinic, New Haven, Connecticut, 2003–2011

Variable	Total Screened (n = 4650), No. (%) or Mean	Total Prevalent LTBI (All TST+) (n = 779), No. (%)	No LTBI (All TST-) (n = 3871), No. (%)	P
Age at TST screening, y	37.0	40.7	36.3	< .001
Gender
Male	2585 (55.6)	481 (61.7)	2104 (54.4)	< .001
Female	2065 (44.4)	298 (38.3)	1767 (45.6)
Race/ethnicity
White	1268 (27.3)	97 (12.5)	1171 (30.3)	< .001
Black	1858 (40.0)	288 (37.0)	1570 (40.6)	.062
Hispanic	1483 (31.9)	374 (48.0)	1109 (28.6)	< .001
Other	30 (2.4)	9 (1.2)	21 (0.5)	.051
Foreign-born	808 (17.4)	372 (47.8)	436 (11.3)	< .001
Undocumented immigrant	421 (9.1)	198 (25.4)	223 (5.8)	< .001
High-school completion	3221 (69.3)	498 (63.9)	2723 (70.3)	.008
Stable housing	2643 (56.8)	444 (57.0)	2199 (56.8)	.583
Committed relationship	791 (17.0)	219 (28.1)	572 (14.8)	< .001
Employed	1454 (31.3)	341 (43.8)	1113 (28.8)	< .001
Health insurance	2853 (61.4)	328 (42.1)	2525 (65.2)	< .001
Known HIV-positive	463 (10.0)	85 (10.9)	378 (9.8)	.271
Known hepatitis C–positive	658 (14.2)	101 (13.0)	557 (14.4)	.369
Injection drug use, ever	1087 (23.4)	135 (17.3)	952 (24.6)	< .001
Injection drug use, past 30 d	347 (7.5)	37 (4.7)	310 (8.0)	<.002
Drug use, ever
Heroin	1543 (33.2)	193 (24.8)	1350 (34.9)	< .001
Cocaine	2116 (45.5)	272 (34.9)	1844 (47.6)	< .001
Crack cocaine	1845 (39.7)	226 (29.0)	1619 (41.8)	< .001
Marijuana	3228 (69.4)	399 (51.2)	2829 (73.1)	< .001
Methamphetamine	234 (5.0)	25 (3.2)	209 (5.4)	.011
Hazardous drinking	2512 (54.0)	390 (50.1)	2122 (54.8)	.015
Opioid substitution therapy, past 30 d	594 (12.8)	67 (8.6)	527 (13.6)	< .001
Emergency department visit, recent	1647 (35.4)	202 (25.9)	1445 (37.3)	< .001
Incarceration, recent	1114 (24.0)	120 (15.4)	994 (25.7)	< .001
Sex solicitation	521 (11.2)	123 (15.8)	398 (10.3)	< .001
Sex work	503 (10.8)	57 (7.3)	446 (11.5)	< .001
Interpersonal violence	927 (19.9)	128 (16.4)	799 (20.6)	.007
Sexually transmitted infection	1203 (25.9)	168 (21.6)	1035 (26.7)	.005
Mental health diagnosis	1322 (28.4)	159 (20.4)	1163 (30.0)	< .001

Note. LTBI = latent tuberculosis infection; TB = tuberculosis; TST = tuberculin skin test. Percentages represent total clients by column category and may not add to 100% because of client nonresponse. Total prevalent LTBI represents the sum of incident LTBI (newly TST-positive) and prevalent LTBI (previously TST-positive and previously active TB).

We calculated the marginal effects for covariates in the multivariate logistical regression to measure if the incremental change in probability of the outcome was associated with the change in each of the covariates.⁴² Testing for multicollinearity revealed that the estimation of parameters on foreign-born and race variables in our final model was robust and did not affect the final outcomes. Finally, we performed a more specific analysis for the top-5 countries of origin with the greatest percentage of LTBI for both newly diagnosed and previous LTBI compared with the United States by using the χ² and Fisher exact test as appropriate.

RESULTS

From January 2003 to June 2011, the CHCV’s screening program for LTBI had a total of 8322 unique clients, of whom 89.9% (n = 7484) completed full demographic intake questionnaires (Figure 1). Of these, 55.6% (n = 4159) of individuals were considered TST-eligible with 40.9% of these (n = 1699) having completed previous LTBI screening. The TST positivity rate for ineligible clients who had completed screening in the past 12 months was 13.2% (n = 440). An additional 1459 individuals (17.5%) who had undergone TST screening in the previous year underwent repeat TST screening on the MMC for regulatory requirements (employment, entry into drug treatment, etc.) and were not included in the TST screening evaluation because they did not meet clinical eligibility according to CDC recommendations.^10,13 Nearly 78.0% (n = 1325) of eligible patients who underwent TST returned 48 to 72 hours later for TST reading. Of those TSTs read, 25.6% (n = 339) had a positive test result; 1 of these was further determined to have TB disease, leaving 338 cases of LTBI diagnosed by the CHCV.

FIGURE 1—

Tuberculin skin test screening algorithm for latent tuberculosis infection on the Community Health Care Van mobile medical clinic (n = 7484), New Haven, Connecticut, 2003–2011.

Note. TB = tuberculosis; TST = tuberculin skin test.

^a459 additional patients who did not meet TST screening eligibility underwent TST screening and are not reported here.

Prevalent Latent Tuberculosis Infection

In total, 779 individuals with total prevalent LTBI (previously and newly diagnosed) compared with those with negative TST (previously and newly placed) were more likely to be male (61.7%), Hispanic (48.0%), foreign-born (47.8%), undocumented (25.4%), and without health insurance (57.9%). It was interesting that clients with a negative TST reported significantly more injection drug use behavior (24.6% vs 17.3%) and crack cocaine use (41.8% vs 29.0%) than those without LTBI (Table 1). Multivariate modeling demonstrated highly significant correlations with total LTBI prevalence (Table 2), including being foreign-born (adjusted odds ratio [AOR] = 6.27; 95% confidence interval [CI] = 4.92, 8.01), Hispanic (AOR = 2.16; 95% CI = 1.64, 2.84), Black (AOR = 2.15; 95% CI = 1.64, 2.84), a heroin user (AOR = 1.34; 95% CI = 1.06, 1.69), male (AOR = 1.31; 95% CI = 1.08, 1.59), and employed (AOR = 1.26; 95% CI = 1.03, 1.56); experiencing interpersonal violence (AOR = 1.33; 95% CI = 1.05, 1.70); not having health insurance (AOR = 1.30; 95% CI = 1.05, 1.63); not completing high school (AOR = 1.26; 95% CI = 1.04, 1.54); and increasing age (AOR = 1.04; 95% CI = 1.03, 1.05). Similar to the model for incident LTBI, being foreign-born contributed the greatest to prevalent LTBI with marginal effect modeling; being foreign-born was associated with an increase in the probability of having LTBI by 30% relative to being not being foreign-born.

TABLE 2—

Independent Correlates of Total Prevalent Latent Tuberculosis Infection (n = 4650) and Incident Latent Tuberculosis Infection (n = 1325) With Multivariate Logistic Regression and Marginal Effects Modeling

Variables	AOR (95% CI)	P	Marginal Effects (95% CI)	P
Total LTBI variablesa
Foreign-born	6.27 (4.92, 8.01)	< .001	0.301 (0.255, 0.3556)	< .001
Hispanic	2.16 (1.64, 2.84)	< .001	0.096 (0.059, 0.132)	< .001
Black	2.15 (1.64, 2.84)	< .001	0.092 (0.058, 0.127)	< .001
Heroin use	1.34 (1.06, 1.69)	.016	0.034 (0.005, 0.062)	.02
Interpersonal violence	1.33 (1.05, 1.70)	.02	0.034 (0.004, 0.065)	.028
Male gender	1.31 (1.08, 1.59)	.006	0.030 (0.009, 0.051)	.005
Uninsured	1.30 (1.05, 1.63)	.018	0.031 (0.005, 0.057)	.021
Not completing high school	1.26 (1.04, 1.54)	.02	0.027 (–0.004, 0.051)	.024
Employed	1.26 (1.03, 1.56)	.023	0.027 (0.003, 0.051)	.028
Age	1.04 (1.03, 1.05)	< .001	0.004 (0.003, 0.005)	< .001
Incident LTBI variablesb
Foreign-born	8.49 (5.54, 13.02)	< .001	0.412 (0.326, 0.497)	< .001
Unstable housing	4.95 (3.43, 7.14)	< .001	0.279 (0.213, 0.346	< .001
Hispanic	3.12 (1.88, 5.20)	< .001	0.191 (0.102, 0.280)	< .001
Black	2.16 (1.31, 3.55)	.002	0.124 (0.042, 0.210)	.003
Employed	1.61 (1.14, 2.28)	.007	0.078 (0.136, 0.370)	.009
Marijuana use	1.57 (1.05, 2.37)	.029	0.069 (0.020, 0.014)	.026
Age	1.04 (1.02, 1.05)	< .001	0.006 (0.004, 0.008)	< .001
Hazardous drinking	0.64 (0.47, 0.87)	.004	–0.070 (–0.022, –0.118)	.004
Crack cocaine use	0.59 (0.39, 0.90)	.015	–0.076 (–0.019, –0.133)	.009

Note. AOR = adjusted odds ratio; CI = confidence interval; LTBI = latent tuberculosis infection.

^aAkaike information criterion = 3228.71; Pearson’s χ² = 0.0754.

^bAkaike information criterion = 1069.73; Pearson’s χ² = 0.051.

Of the 372 total foreign-born individuals with a positive TST, 34.9% (n = 130) reported having a primary care physician, and 23.7% (n = 88) reported having had a CXR before their CHCV visit in the context of general health assessment for immigration or other medical reason. Of the 198 total undocumented immigrants with a positive TST, 29.8% (n = 59) self-reported having a primary care physician, and 13.6% (n = 27) self-reported having a previous CXR screening with a minority of 40.4% (n = 80) reporting previous TST screening before first CHCV visit. Having a primary care provider was not found to be a significant determinant of either previous or newly identified positive TST status.

Newly Diagnosed Latent Tuberculosis Infection

Individuals with newly diagnosed LTBI (incident LTBI as defined by newly TST-positive among those who were eligible for testing) were statistically more likely than those without newly diagnosed LTBI to be male (66.4%), Hispanic (57.2%), foreign-born (60.5%), undocumented (37.5%), and employed (50.4%), and to have never injected drugs (10.0%), and to not have had a recent emergency department visit (73.5%). Only 16.8% (n = 57) of the 339 individuals newly found to have a positive TST self-reported having a primary care doctor, and only 7.4% (n = 25) had received previous CXR screening (data not shown).

Multivariate modeling identified highly significant correlates for incident LTBI (Table 2), which included being foreign-born (AOR = 8.49; 95% CI = 5.54, 13.02), having unstable housing (AOR = 4.95; 95% CI = 3.43, 7.14), being Hispanic (AOR = 3.12; 95% CI = 1.88, 5.20) or Black (AOR = 2.16; 95% CI = 1.31, 3.55), marijuana use (AOR = 1.57; 95% CI = 1.05, 2.37), and being employed (AOR = 1.61; 95% CI = 1.14, 2.28). Hazardous drinking, however, was inversely correlated with clients having a newly identified positive TST (AOR = 0.64; 95% CI = 0.47, 0.87). Marginal effect calculation implied that being foreign-born was associated with an increase in probability of being newly identified as having a positive TST by 0.41 or, equivalently, 41.2%.

Latent Tuberculosis Infection by Country of Origin

Foreign-born status was confirmed to have a large marginal effect on both TST-positive prevalence and incidence. Countries of origin with the most numerous foreign-born patients with incident LTBI identified on this MMC were Mexico (19.2%; n = 65), Ecuador (14.7%; n = 50), Guatemala (3.5%; n = 12), and Jamaica (1.8%; n = 6). The United States and Puerto Rico as a single region contributed the largest absolute number of incident LTBI cases at 49.0% (n = 128) because of the large volume screened from the United States.

Nearly two thirds of newly diagnosed LTBI cases were among foreign-born individuals (62.2%; n = 211 of 339), and being foreign-born increased one’s odds of having a positive TST, from 4.9- to 10.1-fold (P < .001), depending on the country of origin, based on WHO tuberculosis prevalence category (table not shown). Among the 779 total individuals with a positive TST, 25.4% (n = 198) were undocumented immigrants, and 47.8% (n = 372) were foreign-born.

In terms of LTBI prevalence, participants from foreign-born countries had the greatest likelihood of having newly identified LTBI but were similar to the United States and Puerto Rico in terms of absolute numbers (50.6%; n = 394). A graphical representation of TST-positive incidence and prevalence by the top 5 regions of origin reveals statistically significant greater percentage yield for identification of clients with newly diagnosed positive TST (P < .01; Figure 2).

FIGURE 2—

Newly diagnosed and previously diagnosed latent tuberculosis infection by US-born and top-4 foreign-born countries of origin.

Note. LTBI = latent tuberculosis infection. Puerto Rico is shown separately because it has significantly higher incidence and prevalence than the US mainland. Newly diagnosed LTBI (incident LTBI = new TST-positive) plus prevalent LTBI (previous TST-positive) equals total prevalent LTBI (all TST-positive).

^aP < .01 for newly diagnosed LTBI for all foreign countries compared with United States.

^bP < .01 for previously diagnosed LTBI for Mexico, Jamaica, and Ecuador compared with the United States; P < .17 for Puerto Rico and Guatemala compared with the United States.

DISCUSSION

Although other MMCs have effectively screened for TB, one such program in Rotterdam, Netherlands, operational since 2002,^43,44 targeted exclusively the homeless and people who injected drugs by using chest radiographs to detect active TB disease, but did not screen for LTBI. By contrast, our program included a more heterogeneous sample and assessed clients longitudinally over a longer time period; it focused on LTBI identification using TST and was overrepresented by foreign-born Hispanics, which is the predominant immigrant population in New Haven. Although other TB screening programs in MMCs exist domestically and internationally, little peer-reviewed data are available, and such programs typically focus on active TB case finding through CXR screening.

Domestically, the greatest incidence of TB disease is among foreign-born persons from Asia and Africa.⁴⁵ In New Haven, however, the foreign-born populations screened by the CHCV were largely of Hispanic origin, which may reflect both the inner city’s demographics and the strong community referral factor within the local Hispanic community by the CHCV’s Spanish-speaking bicultural staff. A longstanding community presence by this program has been previously described for TB screening programs by creating community-based trust and referral networks for both TB screening and treatment where financial incentives are not provided.⁴⁶

Moreover, among a population with a large number of undocumented immigrants, substance misusers, and hazardous alcohol drinkers, more than three quarters (78.0%) returned for TST readings, higher than reports elsewhere.^47–49 The high acceptance of LTBI testing and return rates, however, may reflect a bias as many clients may have required TST for work, school, or drug treatment programs.

The CHCV mobile screening program serves as an important, ongoing public health intervention with more than a quarter (25.6%) of all eligible clients resulting in newly diagnosed cases of LTBI, which confirms the success of the LTBI screening program that targets high-risk populations that may not access traditional health care settings, evidenced by the low proportion having a primary physician. Established linkages to the statewide TB control program enabled those with LTBI to receive further treatment, free of charge, through provision of free CXR and laboratory testing and provision of isoniazid preventive therapy for interested clients.³⁹ Such proactive public health programs that target foreign-born populations, especially those with undocumented status that are not linked to interdiction efforts, were critical to the success of the LTBI screening program, especially for previously unscreened foreign-born populations.

It is interesting that the decreased likelihood of newly diagnosed LTBI among those using crack cocaine and taking part in hazardous drinking may not reflect low LTBI prevalence, but more likely reflects that these high-risk persons may have previously interfaced with structural settings such as drug treatment and criminal justice settings where TST screenings are routine. Furthermore, we note the high prevalence of foreign-born populations with positive TST among those not using drugs, perhaps suggesting the need to target this population more proactively. The increased association of total positive TST prevalence among heroin users similarly suggests that these individuals were previously screened in structured settings and did not contribute to newly identified positive TSTs on the MMC.

In categorizing LTBI cases by country of origin, we were able to identify a remarkably high prevalence of newly diagnosed LTBI cases from high TB disease incidence countries among foreign-born clients. This finding highlights the importance of LTBI screening among foreign-born immigrants from high TB prevalence countries, recognizing that WHO only reports TB disease and not LTBI prevalence and incidence worldwide. Important in TB-control programs is the use of TB preventive therapy among those with LTBI, especially foreign-born persons. A 9-month course of isoniazid preventive therapy for those with LTBI is fraught with numerous adherence and completion challenges. Newer approaches to reduce LTBI treatment duration to 12 weekly doses of combined isoniazid with rifapentine for those without HIV is promising, though cost and implementation challenges remain, especially in the setting of current national drug shortages.⁵⁰

Limitations

Though these findings are compelling for scaling up LTBI screening domestically, this study is not without limitations. First, among 4159 individuals who were considered TST-eligible, only 40.9% of these (1699 of 4159) had a TST placed and 31.9% overall (1325 of 4159) followed through with TST reading. This low LTBI completion percentage reflects, in part, a calculation bias in estimating TST-eligible patients within subgroups in our population who may not necessarily need yearly screens (e.g., 1-time drug treatment program entry). Nonetheless, among those targeted for LTBI screening, follow through was 78%. Methods that could increase LTBI screening, such as enhanced community outreach, educational initiatives, and mobile health technologies should be considered.⁵¹

Second, previous positive TST screening relied on self-report, but previous studies suggest that self-reported TST-positive status has a high predictive value positive rate associated with the inflammatory skin reaction observed after testing. Last, we did not use LTBI screening strategies that deploy a 1-time serum interferon-γ release assay, which performs only slightly better than TST in predicting future TB disease.⁵² Such strategies remain cost-prohibitive for many MMCs and public health departments that have limited budgets.

Conclusions

Latent tuberculosis infection in high-income countries such as the United States has the greatest prevalence and incidence among foreign-born populations, many of whom are missed through current screening guidelines and within traditional health care settings. Screening initiatives such as on the CHCV’s mobile medical clinic provides an innovative strategy to improve TB detection and promote public health to reduce TB disease. The CHCV, a confidential, free health care venue for both documented and undocumented foreign-born individuals provides an alternative to current traditional health care settings. Such low threshold and grassroots programs are effective in detecting both newly and previously known LTBI, and findings here reflect high uptake utilization by not only underserved foreign-born populations but also by high-risk US-born populations. As the Affordable Care Act increases health insurance coverage for many Americans, it will fail to do so among the many who are likeliest to develop TB disease (e.g., newly arrived and undocumented immigrants) and facilitate transmission to others. Lack of supplemental community-based health care programs may undermine current public health efforts unless immigration reform is achieved. Public health authorities should consider MMCs as an effective method to promote screening of at-risk, vulnerable populations for LTBI and thus provide innovative outreach that can further reduce TB disease among foreign-born and vulnerable populations.