Tuberculosis Screening at 2 San Diego High Schools With High-Risk Populations FREE

SCREENING HIGH-RISK populations for tuberculosis (TB) and treating those with infection has been shown to be a cost-beneficial component of a comprehensive TB elimination strategy.¹ For children living in areas with a high prevalence of TB, screening is recommended² at ages 4 to 6 years and 11 to 16 years. Children with risk factors, such as recent immigration, institutional residence, incarceration, or exposure to adults at high risk for TB, should be tested with greater frequency.² Unfortunately, many children, especially adolescents, do not receive routine medical care, which makes such screening possible. Even if patients are seen for routine health maintenance, skin tests can be problematic, because a return visit is needed to read the test results. The current recommendation of the American Academy of Pediatrics³ is to treat children and adolescents with positive tuberculin skin test (TST) results and no evidence of clinical disease with isoniazid. Unfortunately, even when appropriately prescribed, treatment for TB infection is often not completed since it involves patient compliance for a prolonged course (6-9 months). Past results of school screening have varied. Mass school screening in New York, NY, showed a less than 3% positive TST result rate and was clearly not effective in finding and preventing active disease.⁴ However, a study⁵ done in Boston showed positive TST result rates of 5.1% to 8.9% in 12- to 15-year-old students. Studies⁶ in Australia (subjects aged 12-14 years) and Canada⁷ (subjects aged 3-47 years; mean, 12.6 years) reported positive TST result rates of 10% and 22.5%, respectively. In Southeast Asian adolescents (age, 16-23 years) in Los Angeles, Calif, a 55% positive TST result rate was seen.⁸ Although mass school screening is ineffective in detecting active TB, testing high-risk students in the school setting may be an effective way to screen these individuals. However, future cases of active TB will not be prevented unless the infected children have access to and complete a course of preventive therapy. The incidence of TB in the United States increased dramatically since the mid-1980s. In California from 1985 to 1992, there was a 30% increase in reported cases, from 3492 (13.2 cases/100000 population) to 5382 (17.2 cases/100000 population). In San Diego County, the number of reported TB cases increased from 150 (7.1 cases/100000 population) to 433 (16.5 cases/100000 population) in the same period.⁹ In California, the increased incidence in children has been dramatic. Between 1985 and 1992, cases in children younger than 15 years increased from 314 to 566, an increase of 80%.⁹ Non–US-born persons represented 70% of reported cases of TB in San Diego in 1995; Latin American and Southeast Asian immigrants comprised 60% of the total.^9- 10 In the United States, TB reported in non–US-born persons increased from 21.6% of total cases in 1986 to 29.6% in 1993.¹⁰ Between those born in the United States and those born outside of the United States, differences in disease rates are highest in those younger than 15 years.¹⁰ A 1990 screening project in 5 elementary schools in San Diego showed a positive TST result prevalence of 0.2% to 10.4%, with no new cases of active disease detected. Higher positive TST result rates were seen in Southeast Asian (12.2%), Latino (7.7%), and Filipino (3.2%) ethnic groups than in the white (0.7%) and African American (0.2%) populations (American Lung Association of San Diego and Imperial counties, written communication, October 30, 1990).

The purpose of our study was to determine the prevalence of TB infection and disease in a high-risk population of non–US-born adolescents. It was part of a larger project to assess the administration of preventive therapy in schools.

The project was a cross-sectional design study, involving high school students during the 1994-1995 school year. Two high schools in San Diego with 35% to 45% non–US-born students were selected. These schools were selected purposely because of known ethnic diversity and expected higher rates of TB infection than in other schools. The protocol was approved by the institutional review board at the University of California San Diego Medical Center. All students enrolled in the high schools were eligible for the study. Participation in the project was voluntary, and informed consent in writing was obtained from parents and students. Tuberculin skin tests were performed by the Mantoux method (0.1 cm³ of 5-TU strength purified protein derivative). Placement of skin tests was performed by school nurses and a nurse from the University of California San Diego Medical Center's skin testing office. Tests were read by the nurses after 48 hours. A positive test result was defined as induration of more than 10 mm.³ Skin testing materials were supplied by the San Diego Department of Health Services.

Demographic information was collected on all participating students through school records and questionnaires. Students with positive TST results were given the option of being examined by 1 of 3 study physicians (B.J.A., R.E.B., or A.L.P.) and having chest radiographs done either at a local hospital or by the health department or seeing their own physicians for examinations and chest radiographs. These students who saw their own physicians were asked to return questionnaires completed by their physicians.

Statistical analysis was performed with Epi Info, Version 6.¹¹ Frequency distributions and 2×2 contingency tables were constructed and used to evaluate data. P values were considered significant at less than .05. Relative risk and 95% confidence intervals were calculated to evaluate potential risk factors for TB infection.

A total of 1645 skin tests were placed between February 21, 1995, and March 14, 1995. Thirty-nine (2.4%) students did not return for skin test readings. Of the remaining 1606 (97.6%) students whose tests were evaluated, 2 were not included in further analysis because of late (>72 hours) readings of induration. Seven hundred forty-four (36.4%) students had skin tests placed and read at high school 1 and 860 (56.7%) at high school 2. Positive TST results were found in 95 (12.8%) students tested at high school 1 and 207 (24.1%) students tested at high school 2. Skin test induration ranged from 10 to 76 mm (median, 17 mm). Of the students with positive results, 71.5% had induration of 15 mm or more and 38.7% had induration of 20 mm or more. There were 767 male and 837 female students tested. Equal proportions of male and female students had positive skin test results. Ages ranged from 11 to 20 years (median, 16 years). Increasing age was significantly associated with positive TST reactions (χ² test for trend: high school 1, P=.007; high school 2, P=.05). However, when controlled for ethnicity, this relationship remained only for Vietnamese and Filipino ethnic groups at high school 1.

Ethnicity data of the schools are shown in Table 1. The ethnic distributions of the students tested were representative of the overall ethnic distribution of each school. Skin test results by ethnicity are shown in Table 2 and Table 3. At high school 1, Latino, Filipino, and Vietnamese ethnic groups were more likely to have positive TST results than whites. At high school 2, Latino and Vietnamese ethnic groups were more likely to have positive TST results than whites.

Skin test results by ethnicity and place of birth are shown in Table 4 and Table 5. Except for the Latino group at high school 1, students born outside of the United States were more likely to have positive TST results than those of the same ethnic group born in the United States. Among non–US-born students, there were no significant differences in positive TST result rates between different ethnic groups. Among US-born students at high school 1, Latino and Vietnamese ethnic groups were more likely to have positive TST results than whites. At high school 2, only Latino students were at higher risk.

Of all students whose test results were positive, 67 (22.1%) gave histories of positive TST results. However, only 25 (37.3%) students with a positive TST history produced documentation of completed therapy with isoniazid. Chest radiographs were completed in 230 (76.1%) students with positive TST results. One hundred forty-four (62.6%) chest radiographs and physical examinations were performed as part of school-based directly observed preventive therapy, the other 86 (37.3%) radiographs and examinations were performed by private physicians. Only 1 student had findings compatible with active TB. Smears and cultures were negative for acid-fast bacteria. Forty-seven students (15.6% of those with positive TST results) failed to provide documentation of a chest radiograph or documentation of completed therapy. Excluding these 47 students, the rate of TB in those receiving the TST was 0.06%.

Results of skin testing in these 2 high schools suggest a high rate of TB infection among subpopulations within San Diego. The average positive TST result rate of 18.4% was higher than rates seen in elementary schools. The active disease rate was low at 0.06%, demonstrating, as in other studies, that TST screening is not highly effective in detecting active disease.

In a nationwide survey of school-based TSTs, Driver et al¹² reported 6-fold to 24-fold higher rates of positive TST results in non–US-born populations. Results of the present study confirm the increased prevalence of TB infection in non–US-born populations in all ethnic groups. The effect of birthplace, however, appeared to be less dramatic in the Latino population. Compared with other ethnic groups in high school 1, positive TST results were noted in fewer non–US-born (15.4%), and in a larger percentage of US-born (16.7%), Latino students. This may not be significant since there were relatively few Latino students tested in this school (37 students). In high school 2, however, which enrolled a greater number of Latino students (211 students), positive TST results were found in 44.4% of non–US-born and 12.5% of US-born students. The rate of positive TST results in the US-born Latino students was higher than in other US-born ethnic groups. A previous study¹³ of TSTs in Latino children reported similar rates of 53% of non–US-born and 9% of US-born subjects. The higher rate of positive TST results seen in US-born Latino students may reflect the close relationship between Latino populations on either side of the US-Mexico border. It is more difficult for other ethnic groups to travel regularly back and forth from their countries of origin, and, therefore, those born in the United States are less likely to be exposed to populations with a high prevalence of TB. An explanation for the higher incidence of positive TST results in US-born students who are of Vietnamese ethnicity in high school 1 is unclear, but may be related to immigrant contacts.

Past vaccination with the BCG vaccine is a potential source of false-positive TST results. However, the vaccination is usually given in infancy, and its effect is unlikely to persist 5 to 10 years after it is given.¹⁴ We were not able to obtain reliable data on BCG vaccination, since most of the students could not recall having had the vaccine and documentation of scars was incomplete. Correlation of BCG vaccination with size of skin test induration could not be observed.

The efficacy of the screening process would improve if testing were restricted to non–US-born students. The value of any screening test is dependent on the prevalence of disease in the population. Therefore, the positive predictive value of the Mantoux skin test decreases in populations with a low prevalence of TB infection. Mohle-Boetani et al¹ showed that screening all children is only cost beneficial if the reactor rate is 20% or higher. This situation exists in the non–US-born population tested for this study (positive TST rate of 40.2% for both schools combined), but not in the US-born population (positive TST rate of 4.1% for both schools combined). The Latino population represents an exception since positive TST result rates in US-born Latinos were significantly higher (mean, 13.2%), compared with the US-born white students. The Latino population probably represents a high-risk group independent of birthplace.

Access to health care is limited in adolescents, especially in minority ethnic groups.^15- 16 Many students tested did not have primary care physicians, and most had not received TSTs before this screening. Schools are one of the few places where adolescents are consistently present. It is notable that, of the 1645 TSTs placed on students for this study, 97.6% were read, showing the potential benefit of using schools as sites for providing accessible health care services. Participation in this study was voluntary, and only 36.4% and 56.7% of students participated at each school. This may have introduced a self-selection bias in the study population. If, for example, all those not tested were TB infected, the prevalence of positive TST results in high schools 1 and 2 would have increased to 68.2% and 56.9%, respectively. Alternatively, if all untested students had negative TST results, the prevalence of infection would decrease to 4.7% and 13.7% in high schools 1 and 2, respectively. Students attending high school 2 would still be considered a high-risk population, but those at high school 1 may not. It is unlikely, however, that nonparticipating students would be more or less likely to have positive TST results than those in the study.

Screening is only worthwhile if those identified as infected with TB can be adequately evaluated and treated. Sixty-seven students gave histories of positive TST results, yet few (25 students) had a history of treatment. In fact, only 37.3% of those with positive TST results in the past had been evaluated and completed therapy. In this study, follow-up evaluation for evidence of TB was obtained in 76.1% of those with positive skin test results, with 47 students (15.6%) unavailable for follow-up. Successful evaluations achieved by this project depended greatly on having a full-time coordinator, on-site physicians, accessible radiograph facilities, and extra responsibilities for school nurses. This level of financial and personnel support does not currently exist in many schools.

Furthermore, even when evaluations are obtained, compliance in the adolescent population with daily medications necessary for TB preventive therapy is a potential problem. School-based delivery of medication by school personnel could theoretically be an ideal way to provide directly observed preventive therapy and follow-up nursing care.

The American Academy of Pediatrics has recently shifted the focus for TB prevention away from universal screening toward more focused approaches aimed at high-risk groups.³ The results of this study support this recommendation and show that, in San Diego, non–US-born adolescents represent a particularly high-risk group for TB infection. In addition, the Latino population had a high rate of TB infection regardless of birthplace.

The value of school-based screening is to prevent future cases of active disease. Therefore, screening these adolescents for TB will only be worthwhile if treatment services are linked to screening. Feasibility and cost analysis of directly observed therapy in the schools will be important determinants of whether school-based screening is worthwhile. Additionally, further studies are needed to test other innovative approaches to delivering preventive therapy to infected children.

Clearly, the highest priorities for health departments are insuring completion of therapy for persons with active TB and conducting contact investigations. These measures will do more to prevent the current spread of TB than population-based screening for infection. However, after these 2 areas have been addressed, targeted screening linked to preventive therapy should be considered for populations with elevated rates of TB infection because future cases of active TB will come from these groups. School-based screening of high-risk individuals linked with evaluation and treatment is one way this can be accomplished.

Accepted for publication January 26, 1998.

This study was conducted with financial support from the American Lung Association, San Diego City Schools, and the County of San Diego Department of Health Services, San Diego, Calif. Dr Pong was supported in part by Institutional National Research Services Award Training Grant P32-PE10005 from the Health Resources and Services Administration, Washington, DC.

Data presented as a poster at the National Research Service Award Trainees Research Conference, Chicago, Ill, June 3, 1995, and as a poster at the meeting of the Ambulatory Pediatric Association, Washington, DC, May 2-6, 1997.

We thank Judy Beck, PhD, Susie Horn, RN, Elizabeth Spurgeon, RN, and Jan Stollenwerk, RN, from the San Diego City Schools for their enthusiasm and hard work in completing this project.

Editor's Note: This study shows that all national recommendations must be viewed in the context of local populations. It also provides an example of how effective school-based screening can be.—Catherine D. DeAngelis, MD

Reprints: Richard E. Besser, MD, University of California San Diego Medical Center, 200 W Arbor Dr, Mail Code 8454, San Diego, CA 92103-8454 (e-mail: rbesser@ucsd.edu).