Changes in Tuberculosis Disparities at a Time of Decreasing Tuberculosis Incidence in the United States, 1994–2016

Abstract

Objectives. To assess national progress in reducing disparities in rates of tuberculosis (TB) disease, which disproportionately affects minorities.

Methods. We used Centers for Disease Control and Prevention (CDC) surveillance data and US Census data to calculate TB rates for 1994 through 2016 by race/ethnicity, national origin, and other TB risk factors. We assessed progress in reducing disparities with rate ratios (RRs) and indexes of disparity, defined as the average of the differences between subpopulation and all-population TB rates divided by the all-population rate.

Results. Although TB rates decreased for all subpopulations, RRs increased or stayed the same for all minorities compared with Whites. For racial/ethnic groups, indexes of disparity decreased from 1998 to 2008 (P < .001) but increased thereafter (P = .33). The index of disparity by national origin increased an average of 1.5% per year.

Conclusions. Although TB rates have decreased, disparities have persisted and even increased for some populations. To address the problem, the CDC’s Division of TB Elimination has focused on screening and treating latent TB infection, which is concentrated among minorities and is the precursor for more than 85% of TB cases in the United States.

The incidence of tuberculosis (TB), once the leading cause of death in the United States, has dropped so dramatically that in 1989 the Centers for Disease Control and Prevention (CDC) announced a national strategic plan for TB elimination, defined as less than 1 case per million population.¹ The 9272 TB cases reported in 2016 to the CDC’s National TB Surveillance System (NTSS) is the lowest annual number ever recorded, and the incidence rate of 29 per million is the lowest in surveillance history; this is considerable progress, although it is still well above the elimination threshold.²

Although under control overall, TB is differentially distributed in the US population; minorities bear the greatest burden.^3–7 In 2016, racial/ethnic minorities accounted for 85.9% of all reported TB cases; 68.5% of cases were among non–US-born persons.² Non–US-born persons had a TB rate of 14.7 per 100 000, compared with 1.1 per 100 000 for US-born persons.² Asians accounted for 34.5% of TB cases, although they represent only 5.7% of the US population⁸; 95.4% of Asians with TB were non–US-born.⁹ Hispanics accounted for 28.0% of TB cases (75.7% were non–US-born)⁷ and represent 17.8% of the US population.⁸ Blacks/African Americans (hereafter Blacks) accounted for 21.3% of cases (45.8% were non–US-born)⁷ but make up only 13.3% of the US population.⁸ Persons who are homeless, incarcerated, or suffering from substance use disorders are also overrepresented among TB cases.^9–12

As TB incidence drops, the question is whether gaps in rates narrow, as has happened with HIV among women, or whether the rates drop in parallel, so that the disparity remains the same even while morbidity and mortality drop, as has occurred with breast cancer rates.^13,14 In addition to their negative impact on the well-being of the affected populations, health disparities place burdens on the broader society in the form of increased medical expenditures, lost productivity, and premature death.¹⁵

We assessed the extent to which decreases in TB incidence in recent decades have been accompanied by reductions in TB disparities.

METHODS

The US Department of Health and Human Services defines a health disparity as

a type of difference in health that is closely linked with social or economic disadvantage. Health disparities negatively affect groups of people who have systematically experienced greater social or economic obstacles to health. These obstacles stem from characteristics historically linked to discrimination or exclusion such as race or ethnicity, religion, socioeconomic status, gender, mental health, sexual orientation, or geographic location. Other characteristics include cognitive, sensory, or physical disability.¹⁶

Source of Information About Tuberculosis Cases

Since 1953, through the cooperation of state and local health departments, the CDC has collected information on newly reported cases of TB disease in the United States. Sixty jurisdictions currently submit reports (the 50 states, the District of Columbia, New York City, Puerto Rico, and 7 other US-affiliated jurisdictions in the Pacific Ocean and the Caribbean Sea). We excluded the island jurisdictions for this analysis. Public health authorities report cases electronically to the CDC in a standard format, called the Report of Verified Case of TB form. In 1986, all reporting areas began submitting information on TB patients’ countries of birth. Since 1993, the CDC has collected additional sociodemographic and medical risk factor, clinical, and laboratory information on each reported case. These reports constitute the NTSS.

Source of Population (Denominator) Data

For US-born populations, TB rates per 100 000 persons used population estimates from the US Census Bureau.^8,17 Calculations of TB rates for non–US-born populations used data from the March Supplement of the Current Population Survey, which stratifies population estimates by origin of birth.^17,18 Denominator data for homeless populations used annual estimates of homelessness published by the US Department of Housing and Urban Development.¹⁹

Measures of Tuberculosis Disparities

Although disparities in health status and health care in the United States have been widely examined, no consensus exists on the most appropriate methods for measuring and monitoring them.^20,21 In the United States, reports that track health disparities rely on both relative and absolute measures.^20–22

CDC-recommended measures to assess health disparities include rate ratios (RRs) and indexes of disparity.^20,22,23 A RR is the relative difference between 2 measures and is calculated by dividing the rate in 1 population by the rate in the referent population. It is the most frequently used measure of health disparity.

An index of disparity is a summary measure that averages the sum of absolute differences between individual group rates and a reference rate and expresses the average as a proportion of the reference rate. The result is multiplied by 100. The formula is

where r_j indicates the rates of the jth subpopulation, r_ref indicates the rate of the reference group, and n indicates the number of subpopulation.

For example:

where

• 2.9 is the overall TB incidence per 100 000 persons in the year 2016,

• 0.6 is the TB incidence in White Americans,

• 4.9 is the TB incidence in Black Americans,

• 4.5 is the TB incidence in Hispanic Americans,

• 18.0 is the TB incidence in Asian or Native Hawaiian or other Pacific Islanders, and

• 4.7 is the TB incidence in American Indians or Alaska Natives (AI/ANs).

The higher the index of disparity value for a specific indicator, the greater its corresponding population disparity.

Using NTSS and population data, we calculated proportions, rates, RRs, and indexes of disparity to describe TB disparities in the United States and assess changes over time from 1994 through 2016. We excluded data from Puerto Rico and the US-affiliated jurisdictions in the Pacific Ocean and the Caribbean from all calculations.

When examining RRs for race/ethnicity, we chose Whites as the referent population because this group has historically had the lowest reported rates of TB. For indexes of disparity, we used the total population rate as the reference population.²⁰ If a trend was not linear, we fit a local estimation regression curve to estimate the index of disparity trend over time.²⁴

RESULTS

From 1994 through 2016, whereas TB rates decreased dramatically for both US-born persons (down 83.8%) and non–US-born persons (down 59.3%), the disparity between the 2 groups increased. The RR comparing non–US-born to US-born persons more than doubled, from 5.3 in 1994 to 13.4 in 2016 (Figure 1). The index of disparity for birth origin increased an average of 1.5% annually, from 159.2 in 1994 to 234.5 in 2016 (Table A, available as a supplement to this article at http://www.ajph.org).

FIGURE 1—

Comparison of Tuberculosis Rates in US-Born and Non–US-Born Persons: United States, 1994–2016

Source. Adapted with inclusion of rate ratio from Centers for Disease Control and Prevention.²

Disparities by Race/Ethnicity

During 1994 to 2016, Whites experienced an 82.3% decrease in TB rates, from 3.4 to 0.6 per 100 000, and Blacks an 81.8% decrease, from 26.9 to 4.9 per 100 000 (Figure A, available as a supplement to this article at http://www.ajph.org).² TB rates decreased 76.6% among Hispanics (from 19.2 to 4.5 per 100 000), 72.5% among AI/ANs (from 17.1 to 4.7 per 100 000), and 57.9% among Asians and Pacific Islanders (from 42.8 to 18.0 per 100 000).²

Over the same period, RRs for all racial/ethnic groups except Blacks significantly increased compared with Whites (Figure 2). The average annual percentage increase was 3.7% for Asians, 1.6% for Hispanics, 2.3% for AI/ANs, and a nonsignificant 0.1% for Blacks. The RR for Asians showed the most dramatic increase, from 12.6 in 1994 to 30.0 in 2016 (Figure 2).

FIGURE 2—

Trends in Rate Ratios for Tuberculosis Among Racial/Ethnic Minorities Compared With Whites: United States, 1994–2016

Note. AI/AN = American Indian/Alaska Native. Black: y = 7.957e^0.001x; R² = 0.0221. Hispanic: y = 5.5907e^0.016x; R² = 0.6911. AI/AN: y = 4.7031e^0.0229x; R² = 0.6339. Asian: y = 12.942e^0.0372x; R² = 0.9625.

The index of disparity across racial/ethnic groups showed a nonlinear trend, with an average annual decrease of 1.0% during 1998 to 2008 (P < .001) followed by an average annual increase of 0.4% (P = .33) during 2009 to 2016 (Figure 3).

FIGURE 3—

Trend in Tuberculosis Rate Index of Disparity by Race/Ethnicity: United States, 1994–2016

Disparities by Social or Medical Risks

Incarceration.

Since 2000, persons incarcerated at the time of TB diagnosis have accounted for approximately 4% to 5% of TB cases annually (Figure 4). Incarcerated populations have substantially higher TB incidence than does the general population. In a national study of men aged 18 to 64 years incarcerated at the time of TB diagnosis, estimated annual TB incidence from 2002 to 2013 was higher for local jail inmates (a median of 29 cases per 100 000 inmates) than for federal and for state prison inmates (median of 25 per 100 000 and 8 per 100 000, respectively).¹¹ Incarcerated persons often have additional risk factors for TB. During 2002 to 2013, 18.2% of US-born and 9.7% of non–US-born inmates with TB experienced homelessness in the year before TB diagnosis.²⁵

FIGURE 4—

Percentages of Tuberculosis Cases With Social or Medical Risks: United States, 2000–2016

Homelessness.

Although less than 1.0% of the US population has reported homelessness, approximately 5.0% to 6.0% of TB cases since 1994 have been in persons who reported they were homeless in the year before TB diagnosis (Figure 4).^2,10 The TB rate among homeless persons was an estimated 44 per 100 000 during 2007 to 2011.¹⁰ Both the absolute number and the proportion of TB cases among persons who experienced homelessness has decreased, from 1379 (5.7% of TB cases) in 1994 to 430 (4.9% of TB cases) in 2016.²

Persons experiencing homelessness have a higher prevalence of comorbidities such as HIV infection, mental illness, and substance use disorders that increase the risk of progression to TB.¹⁰ Despite efforts to expand housing options (e.g., the Housing First program), many persons experiencing homelessness continue to rely on large congregate shelters, which increases the potential for TB transmission and TB outbreaks.²⁵

Drug and excess alcohol use.

Substance use disorder in the year before TB diagnosis is 1 of the most commonly reported behavioral risk factors among TB patients in the United States.² The percentage of TB cases reporting excess alcohol use averaged 12.4% during 1994 to 2016 but has been decreasing (Figure 4). An average of 6.9% of TB cases reported noninjection drug use, and 2.2% reported injection drug use.

HIV comorbidity.

The percentage of those with TB who also have HIV infection decreased overall from 1994 to 2016 (Figure 4). In particular, Blacks, who have the highest HIV rates, experienced a 53.5% decrease (Figure B, available as a supplement to this article at http://www.ajph.org). Although HIV and TB rates per million population decreased substantially during 1994 to 2016, the disparity between Blacks and Whites remained. The RR comparing TB and HIV in Blacks and Whites increased from 21.7 in 1994 to 30.5 in 2016 (Figure C, available as a supplement to this article at http://www.ajph.org).

DISCUSSION

The goal of the CDC Division of TB Elimination’s Strategic Plan for 2016 to 2020 is to “reduce TB morbidity in the United States with a particular focus on reducing disparities in TB morbidity among disproportionately affected groups, including non–US-born persons and racial and ethnic minorities.”²⁶

Disparities

We found that, whereas TB disparities did decline for racial/ethnic minorities from 1998 to 2008 as evidenced by a drop in the index of disparity, the later experience did not keep pace with the decline in TB incidence. Overall, TB incidence declined from 1998 through 2008. The other measure of disparity, RRs, showed increases over time for all groups except Blacks. For non–US-born persons, both RRs and indexes of disparity demonstrated increases in disparities over time.

A major reason for the persistence of TB disparities is the higher prevalence of latent TB infection (LTBI) among minority populations, particularly non–US-born persons.²⁷ More than 90% of TB cases among non–US-born persons is attributable to reactivation of LTBI, most likely acquired in their countries of origin.² An estimated 6.3 million of the approximately 40 million non–US-born persons in the United States have LTBI.²⁷ Mathematical models suggest that identification and treatment of persons with LTBI will have the greatest impact on TB elimination in the United States.²⁸

Therefore, reduction of TB disparities and progress toward TB elimination will require increased focus on LTBI testing and treatment. To meet this challenge, the Division of TB Elimination (DTBE) has instituted the following targeted efforts:

• The Tuberculosis Epidemiologic Studies Consortium (TBESC), a CDC-funded collaboration with TB programs and academic institutions in 11 states, has begun an outreach project to establish partnerships between TB programs and community medical facilities that test and treat populations at high risk for LTBI. These partnerships will include research to identify interventions that will increase LTBI testing, diagnosis, and treatment among high-risk populations.

• The CDC-funded Tuberculosis Trials Consortium (TBTC) demonstrated the effectiveness of a short-course regimen for LTBI that has been shown to improve patient adherence over the standard 9 months of treatment with isoniazid: a 3-month, 12-dose regimen of isoniazid and rifapentine.²⁹ TBTC and TBESC are currently collaborating on a clinical trial of the safety, tolerability, and effectiveness of an even shorter LTBI regimen, 6 weeks of daily rifapentine.

• A CDC-funded collaborative effort with 3 universities, begun in 2014, models future incidence of TB and LTBI by population; this will assist efforts to target geographic areas with high TB incidence and LTBI prevalence for LTBI testing and treatment initiatives.³⁰

Additional DTBE activities to address TB disparities and TB elimination are summarized in Table B (available as a supplement to this article at http://www.ajph.org). Global progress in reducing TB rates in countries that contribute large numbers of immigrants to the United States will also likely decrease TB rates in the United States.³¹

Other Factors

Although addressing LTBI in non–US-born persons is critical to TB elimination, other factors also play a role in maintaining disparities. The roots of disparity are complex and include differential exposure to TB, historical discrimination, a lack of access to health care or quality of care, lifestyles, and cultural practices.³² Large proportions of persons with TB have more than 1 social or medical condition that increases their risk of infection with Mycobacterium tuberculosis or of progression to TB disease once infected.^12,25 Substance use disorders are a particular concern, especially with the challenges of the opioid-dependence epidemic.^12,32

Progress toward TB elimination will require continued monitoring and reporting of TB disparities, engagement of the affected populations and communities, and development of broader partnerships with medical facilities and agencies that serve minority populations.

CONCLUSIONS

It is interesting to note that overall TB rates in the United States have leveled off at the same time that TB disparities, measured by RRs and indexes of disparity, have increased. It may be that progress toward TB elimination will improve only in proportion to reductions in disparities; the 2 likely will go hand in hand.