Background

While targeted testing of high-risk groups for latent tuberculosis infection (LTBI) has long been a priority set forth by the Centers for Disease Control and Prevention [1], LTBI testing has not traditionally been part of community-based HIV outreach programs [2]. Several barriers to community-based integrated screenings incorporating tuberculosis (TB), human immunodeficiency virus (HIV), and sexually transmitted diseases (STDs) exist. Aside from administrative separation of TB, HIV and STD control programs in state and county health departments, perhaps the most significant challenge to incorporating TB testing has been the two-visit process of the tuberculin skin test (TST), requiring participants to return to the community site for interpretation [3]. Follow-up rates for TST have been reported as low as 33% in studies of TB community outreach initiatives [3–5] and have even been problematic when a "mandatory" approach is used: A study in Oregon restricted access to shelters, soup kitchens, detoxification programs, and transitional housing and programs based on participation in TB screening, but still suggested that homeless persons shifting their place of sleep posed a barrier to follow-up [6].

Another barrier is the cost of screening. Many of the persons at highest risk for TB, HIV, and STDs are socioeconomically disadvantaged and either do not have adequate resources to pay for screening or choose not to use their limited resources for such screening. In a survey of 248 symptomatic persons with active tuberculosis in Los Angeles, delay in seeking care was associated with concern about cost [7]. Concern about cost has similarly been demonstrated as a barrier to HIV testing [8].

Furthermore, persons in low-TB incidence settings often have limited knowledge regarding TB, particularly regarding risk reduction measures [9, 10]. However, populations targeted for HIV screening do express interest in tuberculosis screening. In a survey of 292 homeless shelter residents in San Francisco, subjects were generally knowledgeable about TB, 57.8% expressed concern about contracting TB infection, and 47.1% described willingness to take measures to avoid TB transmission [11]. We aimed to characterize community acceptance and willingness to pay for and participate in LTBI screening as part of a larger geographic information systems (GIS)-based community outreach integrated screening study.

Methods

In the context of a larger GIS-based study, residences of individuals diagnosed with TB disease, HIV, and syphilis 1/1/05-12/31/07 in Wake County, North Carolina were mapped using ArcMap 9.3 GIS software, with the areas of highest densities of all three diseases (defined as areas with greater than ten cases per square mile over the three-year period) designated as "hot spots." A map of Wake County streets was overlaid on the study map to determine which street intersections and local businesses were within the "hot spots." In collaboration with nurses and disease intervention specialists from the HIV, syphilis, and TB clinics at Wake County Human Services in North Carolina, we identified sites for local screenings. Sites were chosen based on location within a "hot spot," availability of an area within the site to administer confidential surveys, and acceptability by the site owner/manager. Community-based advertising and small incentives (snacks, beverage and a $5 grocery gift card) were used to attract participants. At each screening event, written informed consent was obtained from each participant. Participants were asked questions regarding medical and epidemiologic TB risk, behaviors, willingness to pay, and health care utilization. After the interview, participants received HIV, syphilis, and LTBI testing. Participants were asked to come to the health department evening clinic for HIV results, and in cases of positive HIV or syphilis results, participants were also contacted directly by a county disease intervention specialist.

In the initial group of patients enrolled in the study (N = 78), two tests for LTBI were performed. Blood was drawn for the Quantiferon Gold In-Tube (QFT-GIT) (Cellestis Limited, Carnegie, Victoria, Australia) assay, an interferon gamma release assay (IGRA) for latent TB. A TST was subsequently placed by the Mantoux method, and persons were asked to return to the same community site or the health department TB Clinic in 48-72 hours for TST interpretation. After an interim analysis (N = 24), the written questionnaire was modified for the remainder of participants (N = 89) due to concern about reliability of responses (see Results).

Patient responses were analyzed with descriptive statistics using SAS software (Cary, NC). The study was approved by the Duke University Medical Center Institutional Review Board and the Wake County Human Rights Consumer Affairs/Human Research Committee. QFT-GIT assays were performed according to manufacturer's instructions, and HIV and syphilis tests were performed according to routine health department protocols.

Results

A total of 113 participants were recruited through eleven community outreach screenings conducted at six different sites, including a church, a night club, several community centers, and an apartment complex. Of the participants, 54% were male, the majority were African American (90%) and U.S. born (92%), and mean age was 41.4 ± 12.6 (Table 1). The study group represented a population at high-risk for LTBI and STDs, with 71% reporting prior incarceration, 27% reporting current or prior cocaine use, and 35% reporting current or prior homelessness. Furthermore, 53% of persons admitted to unprotected sex within the past year and 49% reported a prior STD. Most subjects (62%) did not have a regular healthcare provider.

Table 1 Demographics and Behaviors in Participants of an HIV/syphilis/TB screening (N = 113)
Full size table

The initial 24 subjects enrolled stated they would be willing to pay a median of $20 to get tested for HIV and $10 to get tested for TB (Table 2). After a data quality review, the open-ended questions were removed from the questionnaire because some participants were responding with amounts of money beyond a reasonable range. For example, one participant stated he would pay "two million dollars" for TB and HIV testing. When the survey was modified from open-ended questions (Q1, Q2) to multiple choice (Q4) for the remaining 89 participants, 23% said they would not pay any amount for testing.

Table 2 Preferences and Willingness to Pay in Participants of an HIV/syphilis/TB screening (N = 113)
Full size table

At these high-risk community sites, there was also a slight preference for receiving a tuberculosis blood test over skin test (39% vs. 23%), with more persons trusting the blood test to give the "right answer" versus the skin test (47% vs. 9%, p < 0.001). Participants reported they would even pay a median of $5 (interquartile range (IQR): 0,20) to get their preferred TB test (skin test or blood test). When asked whether they would be (theoretically) willing to give up the free study-provided snack in exchange for testing, 38 (84%) participants agreed that they would give up a free snack for their preferred test.

Seventy-eight persons had a TST placed. For the remainder of participants enrolled in the study (N = 35), TST was omitted from the screening procedure due to poor follow-up in the initial group. Of the total 113 participants, there were four phlebotomy failures and five indeterminate QFT-GIT® results, so 109 persons had valid HIV tests obtained and 104/113 (92%) of persons had valid QFT-GIT® results. Of the 109 persons who had blood drawn for HIV testing, 14 (13%) presented to the health department to receive their HIV results. Similarly, while 89% (98/110 who responded to the question) of participants reported that it would be easy to return to the same site in 2-3 days for TST interpretation, only 14/78 (18%) of participants actually returned for an interpretation. Potential barriers to follow-up for the TST interpretation and HIV result notification visits were reported relatively infrequently, with only 38% reporting full-time employment, 19% reporting children at home, and 40% reporting difficulty with transportation. When asked about perceived risk to get sick with TB, 83/106 (78%) thought their risk was low, and 81/106 (76%) thought their risk to get sick with HIV or syphilis was low.

Results for participants' HIV, latent TB, and syphilis tests were not available at the time of this study, as this data was being collected as part of a larger geographic-based integrated screening study. All of the screenings, however, were performed in two identified "hot spots": the first in the northern part of the county with an average case density of all three diseases of 15.7 cases per square mile, and a second in the central part of the county with an average case density of 19.2 cases per square mile. The remainder of the county had an average case density of less than 10 cases per square mile.

Discussion

As tuberculosis control efforts move closer to elimination in low-incidence countries, LTBI treatment in relatively high-risk but low-healthcare access populations will become more important to achieve reductions in incident tuberculosis. Community-based screenings may be a reasonable strategy to target such populations, but significant challenges must be overcome to make such screenings feasible and cost-effective. Our study highlights two such challenges: 1) the value that participants place on receiving these screening tests and 2) low return rates for tuberculin skin test readings and HIV result notifications. Our subjects, who reported a number of risk factors for TB exposure as well as for progression to TB disease, perceive their risk of TB to be low. While the test results of specific subjects in this study were not available for analysis, the case density of all three diseases (based on public health surveillance) in our participants' neighborhoods was higher than other parts of the county. Participants in this study also reported relatively low healthcare utilization. The survey responses suggesting that healthcare is a relatively low priority in this population were supported by the low return rates for both TST interpretation and discussion of HIV test results.

While almost all participants expressed a commitment to return for TST interpretation on initial surveys, most did not return for interpretation of tests actually placed in the study, with no clear logistical barriers. Therefore, the only way to effectively reach this at-risk population may be through innovative community outreach programs. With participant preference for a blood-based test that would also eliminate a return visit for interpretation, IGRAs would likely be well-utilized in this kind of community-based outreach effort, and feasibility and acceptability of such tests has been demonstrated [12]. Incentivizing community health workers responsible for connecting LTBI patients with health infrastructure, and using non-incentive based strategies to engage patients in care, such as cellular phone text messaging, are other potential ways to improve patient adherence.

Incentives have enjoyed moderate success in increasing adherence to TB therapy in high-risk populations. In a randomized control trial of homeless persons with LTBI, usual care at a TB clinic was compared to community follow-up with a $5 incentive; completion of isoniazid therapy was significantly higher in the monetary incentive arm [13]. In a California study of LTBI treatment in drug users, therapy completion rate was 53% when active outreach was used with a financial incentive of $5 per visit, compared to 4% when active outreach was used alone (p < 0.0001) [14]. Likewise, in a Haitian study of 60 patients with active TB, one group received free medical care alone, while the second group also received financial aid for three months, travel expenses, nutritional supplements, and monthly clinic reminders [15]. By one year of follow-up, all of the patients in the financially-supported group were cured of their TB, compared to only 57% of patients in the control group. Consistent with these studies, our data support a higher likelihood of participation in TB and STD testing offered in the community with a monetary incentive rather than free medical care alone in a county clinic.

Conclusion

TB elimination in developed countries will require creative interventions in disadvantaged populations who do not access healthcare regularly and may not place a financial priority on preventive care. TB screening efforts in such populations should employ blood-based methodologies that can simultaneously screen for TB, HIV, and syphilis with a monetary incentive to enhance patient participation.