Risk of Tuberculin Skin Test Conversion among Health Care Workers: Occupational versus Community Exposure and Infection
- Nina M. Larsen1,
- Cara L. Biddle1,
- Mark J. Sotir1,
- Nancy White2,
- Patricia Parrott2, and
- Henry M. Blumberg1,2
- 1Division of Infectious Diseases, Department of Medicine, Emory University School of Medicine, Atlanta, Georgia
- 2Epidemiology Department, Grady Memorial Hospital, Atlanta, Georgia
- Reprints or correspondence: Dr. Henry M. Blumberg, Div. of Infectious Diseases, Emory University School of Medicine, 69 Butler St., Atlanta, GA 30303 (hblumbe{at}emory.edu).
Abstract
A prospective observational cohort study to assess rates of and risk factors for tuberculin skin test (TST) conversion among health care workers (HCWs) was conducted at an urban hospital located in a high tuberculosis—incidence area in 1994–1998. All hospital employees undergoing required testing every 6 months were included. A total of 69 (1.2%) of 5773 susceptible employees had a documented TST conversion (overall rate, 0.38 per 100 person-years worked). No significant difference existed in conversion rates among employees with frequent, limited, or no patient contact. HCWs with a TST conversion lived in zip codes with higher tuberculosis case rates (P ⩽ .05). In multivariate analysis, TST conversion was associated with history of bacille Calmette-Guérin vaccination (relative risk [RR], 11.63), annual salary <$20,000 (RR, 3.67), and increasing age. In the setting of an effective tuberculosis infection—control program, TST conversion rates were low, and risk of conversion among HCWs was associated most strongly with nonoccupational factors.
Tuberculosis (TB) is a potential occupational hazard for health care workers (HCWs) [1]. Nosocomial outbreaks reported over the past decade have heightened awareness of exposure to and risk of infection with Mycobacterium tuberculosis [2–6]. A number of reports have noted that implementation of a hierarchy of TB infection—control measures, including administrative controls (which consist of careful screening of patients for active disease, isolation of those with suspected disease, rapid diagnosis, and prompt institution of therapy), have been effective in preventing nosocomial transmission of TB [7–9]. Although it is clear that inadequate TB infection—control measures can be associated with high rates of TB infection among HCWs, the rate of and risk factors for tuberculin skin test (TST) conversion among HCWs in a nonoutbreak setting has been incompletely defined [10, 11], especially in highly endemic areas of the United States.
We performed our study to evaluate the rate of and risk factors for TST conversions among employees of a 1000-bed university-affiliated urban public hospital over a >4-year period in a nonoutbreak setting after implementation of an effective TB infection control program [7, 12]. Grady Memorial Hospital is located in inner-city Atlanta, and the surrounding neighborhoods that the hospital serves have extremely high rates of TB, including some zip code areas with >120 cases per 100,000 persons [13]. The high rate of TB in the city of Atlanta has resulted in the hospital caring for 200 new patients with TB annually over the past decade [13–15].
We prospectively collected TST data on all hospital employees who are required to have biannual TSTs and then merged this database with the hospital's Human Resources database, which contains demographic information, residence zip code, job category, and salary classification, to assess risk factors for TST conversion among employees working in a highly endemic area.
patients and Methods
The study took place between July 1994 and October 1998 at Grady Memorial Hospital. This study was approved by the Emory University Investigational Review Board and the Grady Health System Research Oversight Committee. There were 11,708 persons employed for variable periods of time during the study period. Zip code residence was not available for 913 HCWs (7.8%), including 276 HCWs who were at risk for TST conversion (i.e., 2 previous negative TST results). The 913 HCWs (whose demographic information did not differ significantly from the overall group) therefore were excluded, leaving a total of 10,795 persons employed during the study period. Demographic variables, the prevalence of ever having had a positive TST result, and TST conversion rates did not differ between those who were excluded and the entire group.
Tuberculin skin testing. Tuberculin skin testing was mandatory for all employees every 6 months, unless there was documentation of a previously positive test. Two-step tuberculin skin testing [6] was required of all newly hired employees. Extremely high compliance (essentially 100%) by HCWs with the skin-testing requirements has been noted elsewhere [7]. Tuberculin skin testing of HCWs was performed at the hospital's Employee Health Service (EHS) via the Mantoux method [6]: a 0.1-mL (5 tuberculin units) solution of purified protein derivative (Tubersol; Pasteur Merieux Connaught USA) was placed intradermally on the volar surface of the forearm and was read 48–72 h later by EHS staff. Self-reporting of results by HCWs was not permitted. A positive TST result was defined as induration ⩾10 mm. A TST conversion was defined as a documented positive TST result after a documented negative TST result performed by EHS staff. Those HCWs with a TST conversion had at least 2 previous negative TST test results before testing positive, because 2-step testing was required of all newly hired HCWs, and susceptible HCWs hired before July 1994 had 2 documented previous negative test results. Data on HCW TST results were collected prospectively and were entered into a computerized database at the EHS. All HCWs were questioned by EHS staff as to whether they previously had received bacille Calmette-Guérin (BCG) vaccination.
Employee information. Employee demographic information (age, sex, race, and residence zip code), as well as job-related data (department, job description, duration of employment, and pay grade), was obtained from the Grady Human Resources Department database. This database was merged with the EHS database containing TST data, BCG vaccination history, and other medical information. Each individual employee's salary was determined by assigning the median salary for the specific pay grades of HCWs in 1998. Employee residence zip code—related data were obtained from the 1990 Census Bureau information for the Atlanta metropolitan statistical area. Annual TB case rates by zip code for the years 1993–1997 were obtained from the Georgia Department of Human Resources [13].
Job classification. By use of job descriptions from the Human Resources database, employees were classified as having frequent, limited, or no direct patient contact. The group with frequent contact included personnel in the Department of Nursing (nurses and nursing assistants), respiratory and occupational therapists, phlebotomists, social workers, chaplains, radiology technicians, and emergency medical transport personnel. The limited contact group included dietitians, personnel in housekeeping, biomedical equipment technicians, laboratory staff, outpatient pharmacists, and financial counselors. The group with no direct patient contact included those in administration, public relations, medical records, human resources, information technology services, and engineering; inpatient pharmacists; and laundry workers. Employees also were classified into the following divisions: “Nursing,” “Clinical,” “Support,” “Administration,” and “Other.” Clinical staff included respiratory and occupational therapists, emergency medical transport personnel, phlebotomists, chaplains, social workers, and radiology technicians. Support staff included dietitians, housekeeping staff, laundry workers, engineers, patient financial counselors, medical records personnel, laboratory staff, and biomedical equipment technicians. Administration staff included personnel in executive hospital administration, accounting, human resources, information technology services, public relations, and marketing. Employees whose job description did not fit the categories listed were placed into the “Other” category.
TB infection—control measures. TB infection—control measures in place at the time of the study and those that had been implemented at Grady Memorial Hospital after an outbreak of nosocomial transmission of TB have been reported elsewhere [5, 7, 12]. They included the following: (1) administrative controls —primarily including a respiratory isolation policy mandating the types of patients required to be placed in respiration isolation at the time of admission (e.g., patients with known TB, patients admitted with TB in the differential diagnosis or for whom an acid-fast bacillus respiratory specimen was ordered, and patients with HIV infection who were admitted with an abnormal chest x-ray), a dedicated nurse epidemiologist to coordinate TB infection—control activities, and mandatory tuberculin skin testing of all HCWs every 6 months; (2) engineering controls —negative-pressure respiratory isolation rooms that had ⩾12 air changes/hour; and (3) personal respiratory protection equipment —a 3M 1812 submicron mask was used until March 1997, when it was replaced by the use of a 3M 1860 N-95 respirator. HCW qualitative fit testing [6] was performed at the time of implementation of the N-95 respirator or at the time of HCW employment, if that was after 1 May 1997.
Statistical analysis. For univariate and multivariate analyses, TST conversion rates among HCWs (no. of conversions per 100 person-years worked) were analyzed by proportional hazards regression by SAS software (SAS Institute). The Wilcoxon rank sum test was used to compare distributions for duration of employment, household income, per capita income, household size, unemployment rate, percentage of residents below the poverty level, and annual TB case rate per 100,000 population from 1993 to 1997. All risk factors assessed in the univariate analysis (tables 1 and 2) were included in the multivariate analysis. The final multivariate model was adjusted for sex, race, degree of patient contact, and employee residence zip code—related variables (annual TB case rate per 100,000 population from 1993 to 1997, household income, and household size). Employee division and the employee residence zip code per capita income, percentage below poverty level, and unemployment rate were not included in the final model because of colinearity. P ⩽ .05 was considered to be statistically significant.
Table 1
Tuberculin skin test (TST) conversion rates and risk factors for employees of Grady Memorial Hospital, Atlanta, 1994–1998.
Table 2
Continuous risk factor analysis for tuberculin skin test (TST) conversion among employees of Grady Memorial Hospital, Atlanta, 1994–1998.
Results
The prevalence of ever having had a documented positive TST result for HCWs employed by the Grady Health System for any duration between July 1994 and October 1998 was 20.3% (2194 of 10,795 employees had a baseline positive TST result, either before or at the initiation of the study in July 1994 or at the time of employment if hired after 1 July 1994). An additional 2828 employees were excluded from the analysis of risk factors for TST conversion because they worked <6 months or only had baseline testing performed. During the study period, 5773 employees had negative baseline TST results (either at the beginning of the study or later, if hired after July 1994), were employed ⩾6 months after baseline skin testing, and were considered to be at risk for TST conversion. The 5773 “susceptible” employees worked a total of 18,069 person-years during the >4-year study period. A total of 676 different patients with confirmed active TB disease were cared for at Grady Memorial Hospital during the study period.
Demographic characteristics of employees at risk for TST conversion are shown in table 1. During the study period, 69 (1.2%) of the 5773 susceptible employees had a TST conversion, resulting in a TST conversion rate of 0.38 per 100 person-years worked. Conversion rates were significantly higher among black employees (0.42 per 100 person-years worked) and other nonwhite employees (0.81 per 100 person-years worked), compared with those among white employees (0.19 per 100 person-years worked; table 1). Other variables associated with increased risk of TST conversion in the univariate analysis included an annual salary <$20,000, history of BCG vaccination (table 1), and shorter employment time (median of 58.3 months for those who did not convert vs. 39.0 months for those who had a TST conversion; P = .05; table 2). In the univariate analysis, sex, age, employee division, and degree of patient contact (frequent, limited, or none) were not associated significantly with risk of TST conversion (table 1). HCWs with TST conversions were more likely to live in zip code areas with higher annual rates of TB (annual median case rate of 17.8 vs. 11.9 cases per 100,000 population; P = .05; table 2).
In multivariate analysis, factors independently associated with an increased risk of employee TST conversion included a history of BCG vaccination (relative risk [RR], 11.63; 95% CI, 2.64–51.28; P = .001), an annual salary <$20,000 (RR, 3.67; 95% CI, 1.58–8.53; P = .003), and increasing age (table 3). The degree of patient contact by HCWs (frequent, limited, or none), as well as race, sex, and employee residence zip code—related variables (annual TB case rate per 100,000 population [1993–1997], household income, and household size) were not associated with increased risk of TST conversion in multivariate analysis. Longer duration of employment was associated with decreased risk (per additional year employed; RR, 0.91; 95% CI, 0.86–0.95; P < .001; table 3).
Table 3
Multivariate analysis of independent risk factors for risk of tuberculin skin test conversion among health care workers for employees of Grady Memorial Hospital, Atlanta, 1994–1998.
Discussion
Although TB has long been recognized as a potential occupational hazard, the risk of TB infection among HCWs, as measured by TST conversion, has been poorly defined, especially in nonoutbreak settings [11]. There was a low TST conversion rate among HCWs (0.38 per 100 person-years worked or a risk of 0.38% per year for each HCW) at our hospital, even though we cared for large numbers of patients with TB. This provides additional data to support the efficacy of a hierarchy of TB infection—control measures (administrative, engineering, and respiratory), as recommended by the Centers for Disease Control and Prevention [6]. This low TST conversion rate during the study period compares with a previous annual conversion rate of 13.2% among nurses at our institution during 1991–1992, before implementation of an effective TB infection—control program [5].
Previous studies from institutions in the United States that care for relatively few patients with TB annually (<15) have reported annual TST conversion rates ranging from 0.11% to 1.7%, whereas reports from institutions that care for large numbers of patients with TB (>100 per year) have reported rates ranging from 1% to 10% [3, 13, 16–21]. Many of these studies are limited by having participation rates that were not specified, were variable, or were low, which could result in substantial selection bias [3], and by lack of baseline 2-step tuberculin skin testing [20]. Our prospective study included 2-step testing and had the advantage of including all susceptible (i.e., a previous negative TST result) employees who worked >6 months. Without baseline 2-step testing for employees who have not had a TST in the previous year, subsequent “conversions” actually may represent boosting, which is thought to represent remote tuberculous infection. The booster effect also may result from previous exposure to nontuberculous mycobacteria [22], as well as BCG vaccination [23], in addition to previous tuberculous infection. A single TST may elicit little response yet stimulate an amnestic response so that a second test, when performed as soon as 1 week later, may elicit a much larger induration or booster reaction [22]. Without 2-step testing, the resultant misclassification of TST results may lead to upwardly biased estimates of the incidence of new tuberculous infection, as well as erroneous characterization of associated risk factors [20]. Two-step testing corrects this upward bias. In fact, at our institution, more HCWs who have a positive TST result because of boosting on the second part of 2-step testing than because of a positive test from a TST conversion [24].
We found that the degree of patient contact was not associated with the risk of TST conversion. There were no significant differences in TST conversion rates among employees who had frequent (e.g., nurses) or limited patient contact, compared with those with no patient contact. Factors that were associated independently with TST conversion among employees at our institution included history of BCG vaccination (RR, 11.6), annual salary <$20,000 (RR, 3.67), increasing age, and shorter duration of employment (RR, 0.91 for each additional year of employment). Thus, in the setting of an effective TB infection—control program, such as that at Grady Memorial Hospital, which has been documented to prevent nosocomial transmission [7, 12], community or nonoccupational factors appear to be most important in determining risk of TST conversion.
To our knowledge, our study is the first to assess employee salary as a risk factor for TST conversion among HCWs. We suspect that those employees who are in the lowest salary category (<$20,000 per year) were found to be at greatest risk for TST conversion as a result of community exposure to TB. This includes inner-city areas in Atlanta that have extremely high rates of active TB, primarily occurring among the black population [13]. The incidence of TB in the city of Atlanta during 1992–1997 was 54 per 100,000 population per year, which was >6 times the national average, and a few zip code areas that surround our institution have had annual rates of TB of >100 per 100,000 population per year [13].
In the univariate analysis, nonwhite race was associated with increased risk of employee TST conversion, but race was not independently associated with increased risk of TST conversion in the multivariate analysis. Our finding that HCW salary, a measure of socioeconomic status, is associated independently with risk of TST conversion is consistent with the report of Cantwell et al. [25], which evaluated national TB case data in the United States for the years 1987–1993. The authors concluded that socioeconomic status accounts for much of the increased risk of TB disease previously associated with race.
Two previous studies, by Bailey et al. [26] and Louther et al. [19], evaluated employee zip code—residence data as a marker for community exposure but had conflicting findings regarding the role of community transmission of TB among HCWs. Bailey et al. [26], at Barnes Hospital in St. Louis, MO, reported that the proportion of persons below the poverty level within the employee's postal zone of residence was the only independent risk factor for TST conversion among HCWs. This led Bailey et al. to conclude that, for certain groups of employees, an exposure to TB in the community probably poses a greater risk than exposure in the hospital setting. Louther et al. [19] examined risk factors for HCWs in New York City at St. Claire's Hospital during 1991–1994. Louther et al. reported a high TST conversion rate among HCWs (5.2 conversions per 100 person-years worked) and that occupation was associated with increased risk of TST conversion, whereas postal zone of residence was not. However, the study by Louther et al. took place before full implementation of the infection-control measures recommended by the Centers for Disease Control and Prevention and during a period after an outbreak at that institution. Two-step tuberculin testing was not implemented before the initiation of either of the studies reported by Bailey et al. and Louther et al.
Limitations of our study include a small number of susceptible HCWs with a history of BCG vaccination. Additional studies are needed at institutions with larger numbers of susceptible HCWs who have received BCG vaccination. Other reports have noted that BCG vaccination was associated with an increased risk of having a positive TST result [27–29]. We previously reported that house staff physicians who were graduates of foreign medical schools (a marker for foreign birth and BCG vaccination) were 5 times more likely than house staff physicians who were US medical graduates to have a TST conversion [12]. A second limitation of our study is the lack of history of birth country for employees. This information is not collected by the Grady Human Resources Department. Finally, no physician HCWs were included in this study because physicians are not Grady employees (they are employees of Emory University or Morehouse Schools of Medicine) and therefore are not included in the hospital's Human Resources Database. A previous report by our group [12] found that the rate of TST conversion among Emory house staff (who rotate through Grady) who are US medical school graduates was 0.72 per 100 person-years worked in January 1993–June 1997, but the time period assessed was different than that in this study involving hospital employees. Over the 4 years 1997–2001, the conversion rate among Emory house staff was ∼0.2 per 100 person-years (authors' unpublished data), which is similar to that reported in this study for Grady employees.
In summary, TST conversion rates did not differ among hospital employees who had frequent or limited patient contact, compared with those who had no patient contact. The overall employee TST conversion rate was low (0.38 per 100 person-years) at an institution that cares for large numbers of patients with TB. Factors independently associated with TST conversion included BCG vaccination, lower salary (<$20,000 per year), and older age. In the setting of an effective TB infection—control program that prevents nosocomial transmission, risk of TB infection among HCWs, as measured by the TST, appears to be related to community rather than occupational exposure and risk.
Footnotes
-
Financial support: National Institutes of Health (K07 HL03078), American Lung Association of Georgia, and Georgia Department of Human Resources (contract 427-93-41861).
- Received November 9, 2001.
- Revision received April 23, 2002.
- © 2002 by the Infectious Diseases Society of America
References
This Article
- AbstractFree
- » Full Text (HTML)Free
- Full Text (PDF)Free
Classifications
Services
Citing Articles
Google Scholar
PubMed
Related Content
Share
- Email this article
-
CiteULike
Connotea
Delicious
Facebook
Google+
Mendeley
Twitter
Published on behalf of
IDSA Members: For your free access to this journal, log in via the IDSA members area.
The Journal
Impact Factor: 8.186
5-Yr impact factor: 7.898
For The Media
For Authors
- Services for authors
- Self-archiving Policy
- IDSA Conflict of Interest policy
- Instructions for Authors
- ICMJE Form
- Submit Now!
Open access options for authors visit Oxford Open
This journal enables compliance with the NIH Public Access Policy
Corporate Services
Alerting Services
Related IDSA Publications
