Development of a Surveillance Definition for United States–Mexico Binational Cases of Tuberculosis

Abstract

Objectives:

Consistently collected binational surveillance data are important in advocating for resources to manage and treat binational cases of tuberculosis (TB). The objective of this study was to develop a surveillance definition for binational (United States–Mexico) cases of TB to assess the burden on US TB program resources.

Methods:

We collaborated with state and local TB program staff members in the United States to identify characteristics associated with binational cases of TB. We collected data on all cases of TB from 9 pilot sites in 5 states (Arizona, California, Colorado, New Mexico, and Texas) during January 1–June 30, 2014, that had at least 1 binational characteristic (eg, “crossed border while on TB treatment” and “received treatment in another country, coordinated by an established, US-funded, binational TB program”). A workgroup of US state, local, and federal partners reviewed results and used them to develop a practical surveillance definition.

Results:

The pilot sites reported 87 cases of TB with at least 1 binational characteristic during the project period. The workgroup drafted a proposed surveillance definition to include 2 binational characteristics: “crossed border while on TB treatment” (34 of 87 cases, 39%) and “received treatment in another country, coordinated by an established, US-funded, binational TB program” (26 of 87 cases, 30%). Applying the new proposed definition, 39 of 87 pilot cases of TB (45%) met the definition of binational.

Conclusion:

Input from partners who were responsible for the care and treatment of patients who cross the United States–Mexico border was crucial in defining a binational case of TB.

Surveillance is a cornerstone of public health practice¹ and an essential pillar of health systems.² Definitions used to accurately capture data on disease surveillance cases should be clear, simple, and practical, and list explicit criteria.¹ Standardization of case reporting through the use of good case definitions enhances data collection and facilitates data analysis and reporting.²

The Centers for Disease Control and Prevention (CDC) Division of Tuberculosis Elimination has a goal to reduce tuberculosis (TB) morbidity in the United States, particularly among non–US-born people and those in racial/ethnic minority groups.³ Among non–US-born people in the United States, those born in Mexico accounted for 6824 of 31 991 (21%) TB cases from 2010-2014.⁴ Non–US-born people from Mexico have lower rates of completing TB treatment⁵ and higher rates of moving or being lost to follow-up during TB treatment than other non–US-born people.^5-7 Furthermore, TB among people born in Mexico may substantially influence the incidence of TB cases among non–US-born people in the United States.⁸

The relationship of United States–Mexico border crossings on TB epidemiology and control is documented.^5,6,9-12 Tuberculosis patients who travel or move between the United States and Mexico require more time and resources from TB control programs than what is typically provided for TB patients who receive all of their care and treatment from 1 country.⁹ Moreover, US TB programs in border areas provide care for TB patients who were diagnosed and started treatment in Mexico but subsequently crossed to the United States,⁶ and some US programs treat patients in Mexico who may never cross the border. These patients do not meet US TB case counting criteria and should not be counted as US TB cases.⁴

During the past 18 years, at least 3 groups of experts have developed definitions for binational cases of TB. In 1999, the Tuberculosis Along the United States–Mexico Border Work Group, with members from CDC and approximately 17 state and local programs in Arizona, California, New Mexico, and Texas, proposed that binational cases of TB be defined as cases that required communication or collaboration across the United States–Mexico border or patients who were source cases or contacts to TB patients on the other side of the border.¹¹ More recently, the United States–Mexico Binational Commission developed a set of guidelines in which they proposed a binational case definition (a patient who may have acquired or may transmit the disease in the other country, or who may require binational collaboration for investigation, management, or control) that can be applied to all notifiable diseases.¹³ Neither definition is being used in a consistent, standardized way to report cases of TB along the United States–Mexico border region.

The Council of State and Territorial Epidemiologists published a position statement in 2013 that included a binational definition for a case of infectious disease and a recommendation that states add a binational variable to their reporting systems.¹⁴ The purpose of the Council of State and Territorial Epidemiologists binational infectious disease case definition and variable was 2-fold: (1) to facilitate sharing of information about disease cases and outbreaks across the US borders with Mexico and Canada and (2) to help understand the burden of binational disease, which may aid public health programs in advocating for additional resources. The goals aligned with the needs of TB surveillance. However, the Council of State and Territorial Epidemiologists case definition was broad in scope (ie, it considered a binational case of TB to be [1] one in which potential exposure occurred while in Canada or Mexico or by a resident of one of those countries, [2] one who is a resident of the other country or has contacts in the other country, or [3] another situation that requires coordination or communication with another country) and applied to all infectious diseases. Therefore, it did not have the specificity desired for TB surveillance purposes.

State and local TB control programs in the United States receive federal funding for TB prevention and control activities through federal cooperative agreement grants.¹⁵ Funding is weighted to account for patients whose care and management may be more challenging than usual and require increased staff member time and resources. Binational status has never been added to the TB funding formula because of a lack of data to describe the increased time and effort associated with care and treatment of TB cases in the United States and Mexico in the border area. A national surveillance definition for binational cases of TB would provide quantitative data for such purposes.

Our objective was to develop a surveillance definition for binational cases of TB that would be easy to apply consistently across reporting jurisdictions. The steps that we took to develop the definition (ie, discussions during meetings, focus-group interviews with key TB program staff members, and a workgroup of subject matter experts from local, state, and federal TB control programs) helped ensure that the definition was relevant to program needs and would be used consistently across reporting jurisdictions. Our methodology is relevant to other diseases and could be useful to programs that want to include all stakeholders, from the local, state, and federal levels, in the development of surveillance definitions.

Methods

Pilot Project

For this project, we defined binational as a characteristic that was related to both the United States and Mexico. For each confirmed case of TB disease reported from January 1 through June 30, 2014 (the project period), the local TB program in each of 9 pilot jurisdictions (Yuma County, Arizona; Imperial County, California; Orange County, California; San Diego County, California; the state of Colorado; the state of New Mexico; Cameron County, Texas; El Paso County, Texas; and Hidalgo County, Texas) determined whether the patient had 1 or more binational characteristics. We selected a set of binational characteristics for the pilot study based on the expert opinion of TB control program staff members and the frequency with which a characteristic was found in local binational data collection forms, contact investigation forms, or notes in TB patients’ charts (based on information gathered during previous CDC site visits). Quantitative data to support the selection of binational characteristics were not available at the time of this study.

Pilot sites filled out a questionnaire that included items on basic demographic information and a checklist of 14 selected binational characteristics (in a “select all that apply” checklist format; Figure 1). The checklist included items on characteristics related to travel and treatment of TB, including separate checkboxes for “received treatment in another country, coordinated by local US TB program”; “received treatment in another country, coordinated by an established, US-funded, binational TB program”; “contact investigation performed in another country, coordinated by local US TB program”; and “contact investigation performed in another country, coordinated by an established, US-funded, binational TB program.” The sites sent the completed questionnaires to CDC, where they were entered into an EpiInfo database.

Figure 1.

Text of pilot questionnaire for surveillance of binational cases of tuberculosis (TB).

We also used demographic data on TB cases from the National Tuberculosis Surveillance System, which collects TB case information from all 50 states and the District of Columbia.⁴ We matched cases from the pilot sites to National Tuberculosis Surveillance System cases using the state case number, a case-specific identifier assigned by the reporting jurisdiction. Before the pilot sites filled out the questionnaire, they determined “count status” (ie, countable or noncountable as a US TB case) according to the US national case definition and recommendations for counting reported TB cases.¹⁶ We organized the 14 binational characteristics into 2 separate columns, depending on whether the case was countable or noncountable (ie, 8 characteristics in column A were for countable cases and 6 characteristics in column B were for noncountable cases). Binational characteristics were selected from the appropriate column in the questionnaire for either countable or noncountable cases of TB (Figure 1).

CDC staff members conducted data cleaning, data management, and descriptive analyses using Excel, EpiInfo, and R software (http://www.R-project.org). We shared results with pilot sites on an ongoing basis. Because data used in this evaluation were collected as part of disease surveillance activities and did not involve human subjects, the project did not require institutional review board approval.

Interviews With Key TB Program Staff Members

After data collection, we conducted 1 focus-group interview of TB program staff members in each of 3 pilot sites (Yuma County, San Diego County, and El Paso County) to better understand the effect on staff workload of identifying and reporting binational cases of TB, determine whether there were additional data elements not included on the pilot questionnaire that might have been useful in identifying a binational case of TB, and assess whether TB staff members had trouble interpreting or using the pilot reporting form. We selected these 3 sites because they reported the highest number of patients with 1 or more binational characteristics during the project period. The focus-group interviews used 13 questions, and we grouped responses to individual questions into categories based on main ideas or themes. The questions were as follows:

1. Who filled out the binational TB pilot form (job category)?

2. What information sources were used to fill out the binational TB pilot form (eg, patient charts, interviews with patients, interviews with case managers)?

3. How much time, on average, did it take to fill out the binational TB pilot form?

4. Was the binational TB pilot form easy to understand?

5. Were any of the items on the form confusing to interpret or fill out?

6. Was the binational TB pilot form easy to complete?

7. Do you have any suggestions for how to improve the binational TB pilot reporting form?

8. How are binational TB cases defined by your program?

9. How are binational TB cases initially identified or detected by your program?

10. Did you have cases that were a lot of work to manage and you thought should qualify as binational but were not captured by this form?

11. Is there any information about binational TB cases (ie, data elements that may help identify a binational TB case) that were not included on the binational TB pilot form?

12. What would you propose as a national surveillance definition for binational TB cases?

13. Would it be useful to continue to report binational TB cases as part of routine reporting of TB surveillance data?

Workgroup

At the end of the project period, we formed a workgroup to develop a surveillance definition that best met the goals of accurately identifying binational cases of TB. The workgroup also determined the additional TB program staff member time and effort associated with reporting binational cases of TB. The workgroup was also tasked with deciding whether it was useful to include binational cases of TB in routine national TB surveillance reporting. Other objectives of the workgroup were to (1) determine which data elements should be included in the national surveillance definition for binational cases of TB, (2) recommend methods for reporting binational TB case surveillance data to CDC in the absence of a formal reporting mechanism (ie, no item on the national TB case reporting form exists for a binational case, and few state and local reporting jurisdictions have a variable in their electronic reporting systems to capture binational cases of TB), and (3) propose a structure for a data collection method within the National Tuberculosis Surveillance System to formalize the reporting of binational cases of TB. Workgroup participants from pilot sites, state TB programs, and CDC participated in 4 one-hour conference calls from January through March 2015.

Results

From January 1 through June 30, 2014, the 9 pilot sites reported 431 cases of TB. Of these, 344 were patients with no binational characteristics (reported through the National Tuberculosis Surveillance System only), and 87 were patients with at least 1 binational characteristic (reported through the National Tuberculosis Surveillance System and the pilot-project questionnaire). Of the 87 cases, the most common binational characteristic was “traveled across border for any length of time during infectious period” (n = 70, 80%). “Traveled across border for any length of time while on TB treatment” was reported for 34 (39%) patients. Five (6%) reported “source case or other epidemiologically linked case identified in another country” (Figure 2).

Figure 2.

Binational characteristics associated with 87 cases of tuberculosis (TB) reported from 9 pilot sites, January 1-June 30, 2014. The 9 pilot sites were Yuma County, Arizona; Imperial County, California; Orange County, California; San Diego County, California; the state of Colorado; the state of New Mexico; Cameron County, Texas; El Paso County, Texas; and Hidalgo County, Texas. Binational was defined as a characteristic that was related to both the United States and Mexico.

Most of the 87 TB patients with at least 1 binational characteristic were male (n = 63, 72%) and had an average age of 45 (range, 4-91) (Table). Of the 344 TB patients who did not have a binational characteristic, 225 (65%) were male and the average age was 47 (range, 0-93). Most (93%) of the 87 TB patients with at least 1 binational characteristic were born in the United States (n = 25, 29%) or Mexico (n = 56, 64%). Most (n = 184, 53%) of the 344 TB patients without a binational characteristic were born in countries other than the United States or Mexico.

Table.

Demographic characteristics of tuberculosis (TB) patients with no binational characteristics, with ≥1 binational characteristic, and who met the proposed binational TB surveillance definition as reported by 9 pilot sites, January 1-June 30, 2014^a

Characteristic	No. (%)
	Patients With No Known Binational Characteristics	Patients With ≥1 Binational Characteristic	Patients With ≥1 Binational Characteristic Who Met the Proposed Binational TB Surveillance Definitionb
All	344 (100)	87 (100)	39 (100)
Sex
Male	225 (65)	63 (72)	28 (72)
Female	119 (35)	24 (28)	11 (28)
Age, mean (range), y	47 (0-93)	45 (4-91)	44 (11-86)
Country of birthc
United States	65 (19)	25 (29)	12 (31)
Mexico	95 (28)	56 (64)	25 (64)
Other	184 (53)	6 (7)	2 (5)
Honduras	21 (11)	3 (50)	1 (50)
Philippines	35 (19)	1 (17)	0 (0)
China	7 (4)	1 (17)	0 (0)
Bolivia	1 (1)	1 (17)	1 (50)
Other	120 (65)	0 (0)	0 (0)

^aData sources: binational questionnaires submitted by pilot sites to the Centers for Disease Control and Prevention; National Tuberculosis Surveillance System (unpublished data). The 9 sites were Yuma County, Arizona; Imperial County, California; Orange County, California; San Diego County, California; the state of Colorado; the state of New Mexico; Cameron County, Texas; El Paso County, Texas; and Hidalgo County, Texas.

^bTB disease patients who crossed the border into the United States from another country during treatment, patients who were referred to a US-funded, binational TB program for treatment continuity (ie, a patient who was being treated in the United States but it was known that he or she would cross the border to Mexico), or patients who did not physically reside in the United States and did not receive treatment in the United States but received treatment in another country through a US-funded, binational program (ie, funding for patient’s treatment in Mexico came from the United States).

^cTwo cases reported in the National Tuberculosis Surveillance System were excluded because of missing data on country of birth.

Interviews With Key TB Program Staff Members

Tuberculosis program staff members who participated in focus group interviews agreed that standardized national surveillance for binational cases of TB was important and should be a routine part of TB reporting. Staff members from one site reported that it took too much time to fill out the pilot questionnaire (30-60 minutes); staff members from 2 other sites reported that it took 5-10 minutes to fill out and that it was an appropriate amount of time. Staff members from all 3 sites reported that the pilot questionnaire was difficult to interpret and complete. Specifically, they found the 2-column format for countable and noncountable cases to be confusing. Also confusing to them was the distinction between “local US TB program” and “US-funded, binational TB program” described in the binational characteristics related to coordinating treatment or contact investigation. Based on these comments, we combined the binational characteristics for treatment completion and the binational characteristics for contact investigations (Figure 2).

Workgroup

The Surveillance Definition for Binational Tuberculosis Cases Workgroup recommended the following definition for a binational case of TB:

• Tuberculosis disease in a patient who crossed the border into the United States from Mexico during treatment, or

• Tuberculosis disease in a patient who was referred to a US-funded, binational TB program for treatment continuity (ie, a patient who was being treated in the United States but it was known that he or she would cross the border to Mexico), or

• Tuberculosis disease in a patient who does not physically reside in the United States and did not receive treatment in the United States but received treatment in Mexico through a US-funded, binational program (ie, funding for patient’s treatment in Mexico came from the United States; unknown if patient will ever cross the border to the United States).

Applying the proposed surveillance definition, 39 of the 87 patients with at least 1 binational characteristic were deemed binational. Of these 39 patients, 28 (72%) were male and the average age was 44 (range, 11-86). Of the 39 patients, 25 (64%) were born in Mexico (Table). Hence, of 431 cases of TB reported by 9 pilot sites during the pilot period, 39 (9%) would be considered binational if the proposed definition were implemented. The 48 patients who had at least 1 binational characteristic but did not meet the definition for binational were similar in characteristics to those who did meet the definition: 36 (75%) were male; 35 (73%) were non–US-born, 31 (65%) of whom were born in Mexico; and the average age was 45.

Discussion

Partners from local, state, and federal US TB control programs along the United States–Mexico border worked together to develop a surveillance definition for binational cases of TB that can be interpreted and reported consistently across reporting jurisdictions. Focusing the definition on TB patients who cross the United States–Mexico border during treatment may not capture data on all TB patients in the United States who have a binational characteristic (ie, a characteristic that is related to both the United States and Mexico) or require additional resources for follow-up, care, and treatment, but it was deemed the most relevant indicator of impact on local and state TB program resources. A TB patient who crosses the United States–Mexico border often requires TB program staff members to conduct activities above and beyond what is normally done for a non–border-crossing TB patient, such as locating contacts who may be frequent border crossers and, therefore, are difficult to find and working with binational programs and TB control program staff members in Mexico to ensure follow-up and treatment completion of patients who reside in, move to, or frequently travel to Mexico.^6,17

Some TB patients in Mexico who do not cross the border into the United States will meet the surveillance definition for binational cases of TB if their treatment is coordinated or supplied by a US-funded, binational TB program. For example, Grupo Sin Fronteras, a federal- and state-funded, binational TB program in Harlingen, Texas, treats patients in Mexico who have not crossed the border to the United States.¹⁸ Currently, these cases are not countable, meaning that they do not meet the criteria to be considered US TB cases and are not included in state or national TB surveillance.¹⁶ Treating patients before they cross the border reduces the risk of transmission within the United States¹⁰ and is a cost-effective way to reduce TB morbidity in this population.¹⁹ Including binational cases that are not eligible to be counted as US TB cases (eg, patients who were diagnosed and began treatment in Mexico before arriving in the United States) in the surveillance definition will allow for a more accurate description of the epidemiology and increased effort associated with binational cases of TB and demonstrate the need for additional funding and resources.⁶

To make our case definition, we combined quantitative and qualitative results collected during the pilot period. For example, although “traveled across border for any length of time during infectious period” was the characteristic most frequently reported during the pilot period, it was determined during discussions with TB program staff members that, in practice, this does not often indicate increased case management complexity and, therefore, was not included in the definition. In comparison, coordinating treatment across the border was determined to represent substantial effort to TB control programs in terms of staff member time and financial resources. Based on pilot data and personal experiences of TB program staff members, the workgroup developed the definition that best met the goals of the project. The surveillance definition for binational cases of TB described here provides a feasible solution to the collection and monitoring of binational case data.

Limitations

This study had several limitations. First, it can take up to 2 years to obtain complete treatment information on a TB patient. The pilot project for assessing binational characteristics of TB patients occurred during a 6-month period, which was too short a time frame for most of the pilot sites to gather complete treatment information on all patients. As such, the percentage of TB patients during the pilot period who met the surveillance case definition was likely underestimated because of a lack of complete information during patients’ treatment.

Second, we do not know if the low percentage of patients reporting “source case or other epidemiologically linked case identified in another country” was the result of a lack of data, incomplete contact investigations, or few epidemiologically linked cases associated with TB among our study population. Although TB control program staff members might be able to identify epidemiologically linked cases through routine contact investigations,²⁰ most TB disease among this population might have been caused by reactivation of latent TB infection acquired many years earlier.²¹

Conclusion

The successful collection of binational TB case data using our surveillance definition will depend on state TB control programs’ commitment to using the proposed definition, as well as proper training and instructional materials to ensure consistency in reporting. Consistently collected binational surveillance data can be useful in educating policy makers and partners in TB control and advocating for appropriate resources to manage and treat binational cases of TB.