|
|
| |||||||||||||||||||||||||||||||||||||||
Background After outbreaks of multidrug-resistant tuberculosis, the Centers for Disease Control and Prevention proposed the use of respirators with high-efficiency particulate air filters (HEPA respirators) as part of isolation precautions against tuberculosis, along with a respiratory-protection program for health care workers that includes medical evaluation, training, and tests of the fit of the respirators. Each HEPA respirator costs between $7.51 and $9.08, about 10 times the cost of respirators currently used.
Methods We conducted a cost-effectiveness analysis using data from the University of Virginia Hospital on exposure to patients with tuberculosis and rates at which the purified-protein-derivative (PPD) skin test became positive in hospital workers. The costs of a respiratory-protection program were based on those of an existing program for workers dealing with hazardous substances.
Results During 1992, 11 patients with documented tuberculosis were admitted to our hospital. Eight of 3852 workers (0.2 percent) had PPD tests that became positive. Five of these conversions were believed to be due to the booster phenomenon; one followed unprotected exposure to a patient not yet in isolation; the other two occurred in workers who had never entered a tuberculosis isolation room. These data suggest that it will take more than one year for the use of HEPA respirators to prevent a single conversion of the PPD test. Assuming that one conversion is prevented per year, however, it would take 41 years at our hospital to prevent one case of occupationally acquired tuberculosis, at a cost of $1.3 million to $18.5 million.
Conclusions Given the effectiveness of currently recommended measures to prevent nosocomial transmission of tuberculosis, the addition of HEPA respirators would offer negligible protective efficacy at great cost.
The CDC recently published a draft guideline proposing new measures to prevent nosocomial tuberculosis, including the use of respirators with high-efficiency particulate air filters (HEPA respirators) in isolation rooms for patients with possible tuberculosis20. The Occupational Safety and Health Administration (OSHA) announced in October 1993 that it would require the use of HEPA respirators and a respiratory-protection program. We used data from the University of Virginia Health Sciences Center to estimate the feasibility and cost effectiveness of these additional requirements in hospitals such as ours, which are complying with the control measures recommended by the CDC.
Methods
The University of Virginia Hospital is a tertiary care center that moved in 1989 to a new facility with 700 beds and 47 negative-pressure-ventilation rooms with anterooms. The airflow at the doors of these isolation rooms is checked semiannually and whenever patients known to have multidrug-resistant tuberculosis are admitted. Hospital policy requires the immediate isolation of any patient with possible tuberculosis, including HIV-infected patients with cough and a new respiratory illness.
The frequency with which patients with tuberculosis were admitted was obtained from lists reported to the health department. The number of patients in isolation rooms was obtained from computer records of isolation orders from June 1992 through May 1993. Weekly logs of patients in isolation were used to validate these records. To provide a record of the number of health care workers entering isolation rooms and the number of visits per day, such personnel were asked to sign a sheet each time they entered the room.
Data on the annual screening of health care workers with purified protein derivative (PPD), required by hospital policy, were obtained from previously published studies21,22 and from the employee health department at the hospital. A conversion was defined as a newly positive PPD test with induration of 10 mm or more at 48 to 72 hours23.
The costs of masks were obtained from the manufacturers and from the purchasing department of the hospital. Annual costs were derived by multiplying the estimated number of masks used in one year by the price of the mask. These estimates ranged from a minimal amount, which assumed adherence to recommended patterns of mask use, a minimal number of health care workers caring for each patient kept in isolation, or both, to a maximal amount, which assumed premature disposal of the mask, a maximal number of health care workers caring for each patient, or both. Estimates of lost time and costs associated with the respiratory-protection program were obtained from data on an existing program for maintenance workers with occupational exposure to hazardous materials. The average salary for full-time hospital employees, excluding physicians on the faculty, was obtained from the hospital administration. Costs were expressed in 1993 dollars.
Results
Tuberculosis was diagnosed in 11 of 28,000 patients admitted during 1992. During the previous four years, 51 patients with tuberculosis were admitted (mean, 12.8 per year). From June 1992 through May 1993, 76 patients were kept in isolation rooms during 82 admissions, for a total of 611 days (7.5 days per admission). An average of 25 health care workers visited each isolation room each day, making an average of 50 visits per room per day. The health care workers who entered a particular room differed from day to day.
There were eight newly positive PPD tests in 1992 among 3852 health care workers (0.2 percent) (Table 1), but six of these conversions occurred among employees hired the previous year who had had only one previous negative PPD test (i.e., at the time of their hire). These persons (mean age, 42 years) were considered unlikely to have entered an isolation room. Five conversions were considered most likely to be due to the booster phenomenon; one worker (who was 39 years old) had a conversion on his second PPD test, five months after he was hired, but this occurred after exposure to a patient not yet in isolation and may have represented a true conversion. Two conversions occurred in employees with at least two negative PPD tests previously; neither person had worked with a patient isolated for possible tuberculosis. One worked on a surgical unit, and the other worked on an outpatient dialysis unit. No patient with known tuberculosis was cared for on those units during the year before the PPD conversions of these employees.
|
0.10 x [0.19 + (0.07 x 0.81)] = 0.02467.
The number of years needed to prevent a single case of occupational tuberculosis would then be the inverse of 0.02467, or 41 years. The only such case recognized at this hospital during the past two decades would not have been prevented by the use of HEPA respirators, because the exposure occurred before tuberculosis was suspected.
A simple isolation mask costs $0.06. A dust-mist respirator costs $0.92. Costs for HEPA respirators in 1993 ranged from $7.51 for disposable models to $9.08 for respirators with replaceable filters (3M Health Care, St. Paul, Minn.). Minimal and maximal estimates of the annual cost of each type of mask (Figure 1) were based on these prices. For simple isolation masks, the figure of $1,833 (i.e., 50 room visits per day times 611 days of isolation per year times $0.06 per mask) represents both the minimal and the maximal estimate, because these masks are not reusable.
|
At least $15,396 would be required to purchase disposable HEPA respirators if the same 25 workers cared for each patient throughout the entire hospitalization and discarded their respirators only when necessary (25 x 82 x $7.51). A maximum of $114,715 would be required if different workers cared for each patient every day or if the workers discarded their masks each day (25 x 611 x $7.51).
For respirators with replaceable HEPA filters, the minimum cost would be $18,614 if the same workers cared for each patient throughout the patient's hospitalization (25 x 82 x $9.08). A maximum of $138,697 would be required if different workers cared for each patient every day (25 x 611 x $9.08).
As part of a personal respiratory-protection program, OSHA requires training and testing of the proper mask fit for each worker. We estimate that testing of fit will require 20 minutes to perform, plus 20 minutes of transit time for each of the 3852 health care workers at our institution with potential exposure to a patient with tuberculosis, or 2568 hours of lost time for the workers plus 1284 hours for the tester of fit. During the first year of the respiratory-protection program, this would be the equivalent of the time worked by 1.89 full-time employees at a cost of $67,462 (given that 1 full-time hospital employee works 2040 hours per year for a mean annual pay of $35,694). The cost of testing the fit of the mask for 350 new employees each year would be $6,124. Thus, testing of fit would cost $312,422 in all over a 41-year period.
Training would require 193 half-hour sessions (assuming 20 employees per session) and an additional 20 minutes of transit time per employee. This would result in 3210 hours of lost time for health care workers, at a cost of $56,166 for the first year and 96 hours, or $1,680, for the trainer. The cost of training newly hired employees would be $5,256 per year. The estimated costs of training for 41 years would be $268,086.
OSHA also suggests an annual medical evaluation25. Given that in the current program for maintenance workers dealing with hazardous substances a medical evaluation costs $60 and takes one hour of the employee's time, the total cost would be $231,120 and the equivalent of the time lost by 1.89 full-time employees ($67,462) for the 3852 health care workers each year ($12,241,862 for a 41-year period).
The CDC draft guideline states that screening every five years with a questionnaire should suffice to identify workers who need further evaluation. If we estimate that 10 minutes are needed for each employee to fill out a questionnaire and 5 minutes for the questionnaire to be screened in the employee health department, the cost would be $134,797 for a 41-year period.
We thus estimate that preventing a single case of occupational tuberculosis during the next 41 years by implementing the proposed requirements for HEPA respirators and a respiratory-protection program would cost this hospital between $1,333,090 and $18,508,947 (Table 2). If the number of respirators could be reduced by 50 percent because workers were caring for more than one patient, the minimal and maximal estimates would be reduced to $1,030,923 and $15,665,659, respectively.
|
The CDC draft guideline specifies the use of HEPA respirators to prevent nosocomial tuberculosis but gives no epidemiologic data about their efficacy20. Recent tuberculosis outbreaks occurred in hospitals with inadequate administrative and engineering measures for control10,11,12,13,14. Each hospital had isolation rooms with positive pressure relative to the hallway11,12,13,14. In one outbreak, exhaust air from a sputum-induction room was recirculated into the HIV clinic11. In another outbreak, patients were permitted to go to common areas or group activities without keeping their masks on. When they were readmitted, patients with known tuberculosis were not always placed in isolation again14. In these outbreaks transmission was controlled when the 1990 CDC guidelines were implemented. HEPA respirators were not used10,11,12,13,14.
At our hospital, administrative and engineering controls have been used for decades to prevent nosocomial tuberculosis. Simple isolation masks were used until 1993, when the hospital switched to dust-mist respirators. Our PPD screening has shown declining rates of new positive tests for 25 years (Table 1). Many conversions occurred in employees without known exposure and were possibly due to exposure in the community after the employee was hired or to exposure before hiring (i.e., with conversion on the second PPD test after hiring because of the booster phenomenon). In 1992, only three health care workers were considered actually to have had a conversion, one of which was due to unprotected exposure. The other two employees had not worked with a patient in isolation for tuberculosis; thus, HEPA respirators would not have prevented them from having a conversion.
The National Jewish Center for Immunology and Respiratory Medicine in Denver documented only two PPD conversions in the decade after 1983. There, employees use simple isolation masks while caring for the many patients with tuberculosis, who stay in negative-pressure rooms with ultraviolet lights (Burton LJ: personal communication).
In 1992 our hospital spent $1,833 on simple isolation masks. In 1993 dust-mist respirators were used that cost almost 14 times as much as the simple isolation masks. In turn, HEPA respirators could cost six times more than the dust-mist respirators. Implementing a respiratory-protection program would cost still more. At hospitals that treat more patients who require isolation for tuberculosis, the costs would obviously exceed our estimates.
Such costs are difficult to justify, given the lack of epidemiologic data demonstrating the effectiveness of either HEPA respirators or a respiratory-protection program and the strong epidemiologic evidence for the effectiveness of the currently recommended administrative and engineering controls. Moreover, when tuberculosis is transmitted, it is often transmitted by patients who have been given an incorrect diagnosis and who have not been isolated. HEPA respirators would not alter the risk of exposure to such patients or the risk that is present early in an outpatient visit, before a history has been taken that suggests tuberculosis (as in the only case of occupational tuberculosis documented at our hospital in the past two decades). Because the CDC specifies that the new guidelines should be followed in ambulatory care clinics, dental clinics, home health care settings, emergency medical services, and other facilities, such as treatment centers for substance abuse and medical areas in correctional facilities, the potential cost to the nation could be very high.
HEPA respirators have inconvenient aspects that are important but difficult to quantitate precisely. They are bulky and less comfortable than isolation masks. They muffle the voice and interfere with communication with the patient. They may cause respiratory compromise in some workers. Overall, HEPA respirators are cumbersome, and a requirement to use them would interfere with practical aspects of the daily delivery of health care.
In an era of cost control by the federal government, this proposal would lead to tremendous, unnecessary increases in hospital expenses. We are concerned that many hospitals may respond by eliminating other, more important parts of their infection-control programs that actually do prevent infection. This would be most unfortunate and could lead to further increases in hospital costs and excess in-hospital mortality.
Conclusions
Our data show that current administrative and engineering controls are very effective in preventing the nosocomial transmission of tuberculosis. Hospitals that use such controls have no need for HEPA respirators or a respiratory-protection program. In our opinion, the draft guideline represents an overreaction to recent outbreaks of nosocomial tuberculosis in hospitals that were not complying with recommended control measures. Those outbreaks were controlled by the implementation of the current guidelines, without the use of HEPA respirators. Data from hospitals that comply with existing guidelines suggest that the proposed measures would add negligible protective efficacy at great price. Such costly, unproved measures should not be required unless epidemiologic data demonstrate their efficacy and cost effectiveness.
Supported in part by a training grant (T32AI07046) from the National Institutes of Health.
We are indebted to Vickie Pugh from the Employee Health Department of the University of Virginia Hospital, Michelle Whitlock from the Department of Environmental Health and Safety, and Barbara Strain from the Microbiology Laboratory for their assistance in obtaining data.
Source Information
From the University of Virginia Health Sciences Center, Box 473, Charlottesville, VA 22908, where reprint requests should be addressed to Dr. Farr.
References
| |||||||||||||||||||||||||||||||||||||||
Related Letters:
HEPA Respirators and Tuberculosis in Hospital Workers
Sobel E., Sherertz R. J., Streed S. A., Brown V., Bishop C., Rutala W. A., Weber D. J., Adal K. A., Anglim A. M., Farr B. M.
Extract |
Full Text
N Engl J Med 1994;
331:1658-1660, Dec 15, 1994.
Correspondence
This article has been cited by other articles:
HOME | SUBSCRIBE | SEARCH | CURRENT ISSUE | PAST ISSUES | COLLECTIONS | PRIVACY | HELP | beta.nejm.org Comments and questions? Please contact us. The New England Journal of Medicine is owned, published, and copyrighted © 2008 Massachusetts Medical Society. All rights reserved. |
Previous
