The increased incidence of TB in correctional facilities has been attributed to several factors (Ex. 7-25). One, correctional facilities have a higher incidence of individuals who are at greater risk for developing active TB. For example, the population in prisons and jails may be dominated by persons from poor and minority groups, many of whom may be intravenous drug users. These particular groups may also suffer from poor nutritional status and poor health care, factors that place them at increased risk of developing active disease. Two, special types of correctional facilities, such as holding facilities associated with the Immigration and Naturalization Services, may have inmates/detainees from countries with a high incidence of TB. For foreign-born persons arriving in the U.S., the case rate of TB in 1989 was estimated to be 124 per 100,000, compared to an overall TB case rate of 9.5 per 100,000 for the U.S. (Ex. 6-26). In 1995, TB cases reported among the foreign born accounted for 35.7 % of the total reported cases, marking a 63.3 % increase since 1986 (Ex. 6-34). Three, many correctional facilities have a high proportion of individuals who are infected with HIV. The CDC reported that in addition to the growing increase in AIDS among prisoners, the incidence of AIDS in prisons is markedly higher than that for the U.S. general population. In 1988, the incidence of AIDS cases in the U.S. population was 13.7 per 100,000 compared to an estimated aggregate incidence for state/federal correctional systems of 75 cases per 100,000 (Ex. 3-33). Individuals who are infected with HIV or who have AIDS are at an increased risk of developing active TB due to their decreased immune capacity. The likelihood of pulmonary TB in individuals with HIV infection is reflected in the CDC's Revised Classification System for HIV infection (Ex. 6-30). In this revised classification system, the AIDS surveillance case definition was expanded to include pulmonary TB. Moreover, X-rays of individuals infected with HIV who have TB often exhibit radiographic irregularities that make the diagnosis of active TB difficult (Exs. 7-76, 7-77, 7-78, and 7-79). HIV-infected individuals may have concurrent pulmonary infections that confound the radiographic diagnosis of pulmonary TB. In addition, it may be difficult to distinguish symptoms of TB from Pneumocystis carinii pneumonia or other opportunistic infections. This difficulty in TB diagnosis can result in true cases of active TB going undiagnosed in this population. Undiagnosed TB has been shown to be an important cause of death in some patients with HIV infection (Ex. 7-76). Fourth, environmental conditions in correctional facilities can aid in the transmission of TB. For example, many prisons are old, have inadequate ventilation systems, and are overcrowded. In addition, inmates are frequently transferred both within and between facilities, thus increasing the potential for the spread of TB infection among inmates and staff. This increased potential for mobility among inmates also enhances the likelihood that inmates undergoing therapy for active disease will either discontinue their treatment or inadequately follow their prescribed regime of treatment. The inadequacy of their treatment may give rise not only to relapses to an infectious state of active disease, but also potentially give rise to strains of MDR-TB. These strains of TB have a higher incidence of fatal outcome and are generally characterized by prolonged periods of infectiousness during which the risk of infection to others is increased.

The high incidence of TB among the inmate population presents an occupational hazard to the staff in these types of facilities. Recent outbreak investigations by the CDC have documented the transmission of TB to exposed workers. In an investigation of a state correctional facility in New York for 1991 (Exs. 6-3 and 7-136), eleven persons with TB were identified (10 inmates and one correctional facility guard). Nine persons (8 inmates and the guard) had MDR-TB. All eight inmates were HIV positive. The guard was HIV negative; however, he was also immunocompromised as a result of treatment for laryngeal cancer. Seven of the inmates and the guard died from MDR-TB. The eighth inmate was still alive and receiving treatment for MDR-TB 2 years after being diagnosed as having the disease. DNA analysis identified the strains of tuberculosis bacteria from these individuals to be identical.

The investigation revealed that the source case was an inmate who had been transferred from another prison where he had been previously exposed to MDR-TB. He arrived at the prison with infectious TB but refused evaluation by the infirmary staff. This inmate was placed in the general prison population where he stayed for 6 months until he was admitted to the hospital where he later died. However, before his hospitalization, he exposed two inmates living in his cell block who later developed MDR-TB. These two inmates continued to work and live in the prison until shortly before their final hospitalization. The other inmates who subsequently developed MDR-TB had several potential routes of exposure: social contact in the prison yard, contact at work sites in the prison, and contact at the prison infirmary where they shared rooms with other inmates before diagnosis with TB.

The guard who developed MDR-TB had exposure to inmates while transporting them to and from the hospital. The primary exposure for this guard apparently occurred when he was detailed outside the inmates' room during their hospitalization for MDR-TB. The inmates were hospitalized in an isolation room with negative pressure. However, upon investigation it was discovered that the ventilation system for the room had not been working correctly and had allowed air to be exhausted to the hospital corridors and other patient rooms.

A contact investigation in the prison was conducted to identify other inmates who might have been exposed during this outbreak of MDR-TB. Of those inmates with previous negative tuberculin skin tests and without active disease (306), ninety-two (30 %) had documented skin test conversions. There was no tuberculin skin test program for prison staff; therefore, conversions among prison employees could not be evaluated.

The primary factors identified as contributing to this outbreak were deficiencies in identifying TB among transferred inmates, laboratory delays, and lapses in isolating inmates with active TB within the facility. Inmates with symptoms of active disease were not sent for evaluation in some cases until they became so ill they could not care for themselves. Some of these inmates were placed in the infirmary with other inmates until their diagnosis with TB. On other occasions, drug susceptibility testing was not reported until after an inmate's death, which means that appropriate patient management was not initiated.

As a result of this outbreak, a retrospective epidemiological investigation was conducted to examine the potential extent and spread of MDR-TB throughout the New York State prison system during the years 1990-1991 (Ex. 7-137). This investigation revealed that 69 cases of TB were diagnosed in 1990 and another 102 were diagnosed in 1991, resulting in a combined incidence of 156.2 cases/100,000 inmate years for 1990 and 1991 combined. Of the cases, 39 were identified as being MDR-TB, 31 of which were shown to be epidemiologically linked. Thirty-three of the individuals with MDR-TB never received any treatment for MDR-TB, 3 were diagnosed at death, and 23 died before drug susceptibility results were known. These inmates were also discovered to be highly mobile. The 39 inmates lived in 23 different prisons while they were potentially infectious. Twenty transfers were documented for 12 inmates with potentially infectious MDR-TB (9 shortly before diagnosis, one after diagnosis with TB but before diagnosis with MDR-TB, and 2 after a diagnosis of MDR-TB).

Several factors were identified as contributing to the spread of MDR-TB throughout the New York prison system: delays in identifying and isolating inmates, frequent transfers without appropriate medical evaluation, lapses in treatment, and delays in diagnosis and susceptibility testing.

A similar investigation in a California state correctional institution identified three active cases of TB (two inmates and one employee) during September and October 1991 (Ex. 6-5). As a result, an investigation was commenced to determine whether transmission of TB was ongoing in the institution. Eighteen inmates with active TB were identified. TB in 10 of these inmates was recognized for the first time while they were in the institution during 1991, resulting in an annual incidence of TB of 184 per 100,000, a rate greater than 10 times that for the state (17.4 per 100,000). Two of the 10 inmates had negative tuberculin skin tests prior to their entry into the institution. Three of the cases were determined to have been infectious during 1991.

A review of skin test data revealed that for the 2944 inmates for whom skin test results were available, 324 tested positive for the first time while in the prison system. Of these, 106 were tuberculin negative before their entry into the prison system, 96 of which occurred in the previous two years.

The employee identified as having active TB had worked as a counselor on the prison's HIV ward, where he recalled exposure to one of the 3 infectious inmates. This employee could recall no known exposures outside the prison. Similarly, two other prison employees had documented skin test conversions while working at the prison. Neither recalled exposures outside the prison; one reported exposure to an inmate with possible TB.

No information was provided in this report as to whether any isolation precautions were implemented at this facility. However, the investigators concluded that their findings suggested the likelihood that transmission of TB had occurred in the prison. Their conclusion was based on the fact that a substantial number of skin test conversions were documented among the inmates and that at least two inmates with active TB became infected while at the prison.

The transmission of TB was also reported in another California prison among prison infirmary physicians and nurses and correctional officers (Ex. 6-6). In this investigation, an inmate with active MDR-TB spent 6 months during 1990-1991 in the infirmary. The infirmary had no isolation rooms and inmates' cells were found to be under positive pressure. Employees occasionally recalled wearing surgical masks when entering the rooms of TB patients.

An analysis of available skin testing data revealed that of the 21 infirmary health care providers, only 10 had been tested twice during the period from 1987 to 1990. Of these 10, two were newly positive, one of whom had recently converted in 1991 and had spent 5 months in the preceding year providing health care to the source case in this investigation. Another health care provider and a correctional officer who worked in the infirmary also were identified as having newly converted while at the prison. There was no yearly skin test screening, and thus their conversions could have occurred at any time between 1987 and 1991. However, 13 other inmates were diagnosed with pulmonary TB during that same period. An additional correctional officer who did not work in the infirmary also was found to have newly converted. His reported exposure occurred at a community hospital where he was assigned to an inmate with infectious TB. The officer was not provided with any respiratory protection. The lack of isolation precautions and the lack of appropriate respiratory protection suggest transmission of TB from infectious inmates in the infirmary to the prison staff, either as a result of exposure to the source case or other inmates with pulmonary TB who were also treated in the prison infirmary. Because of the lack of contact tracing or routine annual screening of inmates or staff, the full extent of transmission from the source case or other TB cases could not be determined.

Thus, similar to the evidence for the hospital setting, the evidence on correctional facilities shows that the failure to promptly identify individuals with infectious TB and provide appropriate infection control measures can result in the exposure and subsequent infection of employees with TB. These employees include the correctional facility infirmary staff, guards on duty at the facility, and guards assigned to escort inmates during transport to other facilities (e.g., outside health care facilities and other correctional facilities).

Homeless Shelters

Tuberculosis has also been recognized as a health hazard among homeless persons. The growth of the homeless population in the United States since the 1980s and the subsequent increase in the number of shelters for the homeless, furthers heightens the concern about the potential for the increased incidence and transmission of TB among the homeless, especially in crowded living conditions such as homeless shelters.

A number of factors are present in homeless shelters which increase the potential for the transmission of TB among the shelter residents and among the shelter staff. A high prevalence of TB infection and disease is common among many homeless shelters. This is not surprising, since the residents of these facilities usually come from lower socio-economic groups and often have characteristics that place them at high risk. Screening of selected clinics and shelters for the homeless has shown that the prevalence of TB infection ranges from 18 to 51 % and the prevalence of clinically active disease ranges from 1.6 to 6.8 % (Ex. 6-15). The CDC estimates this to be 150 to 300 times the nationwide prevalence rate (Ex. 6-17).

In addition to having a high prevalence of individuals with TB infection in the shelters, many of the shelter residents possess characteristics that impair their immunity and thus place them at a greater risk of developing active disease. For example, homeless persons generally suffer from poor nutrition, poor overall health status and poor access to health care. Many also suffer from alcoholism, drug abuse and psychological stress. Moreover, a significant portion of homeless shelter residents are infected with the HIV. In 1988, the Partnership of the Homeless Inc. conducted a survey of 45 of the nation's largest cities and estimated that there were between 5,000 and 8,000 homeless persons with AIDS in New York City and approximately 20,000 nationwide (Ex. 7-55). Due to these factors, homeless shelter residents are at increased risk of developing active disease. Thus, there is the increased likelihood that these individuals will be infectious as a result of active disease and thereby present a source of exposure for other homeless persons and for shelter employees.

In addition to having factors which increase their risk of developing active TB disease, homeless persons also are a very transient population. Because they are transient, homeless persons are more likely to discontinue or to erratically adhere to the prescribed TB therapy. Inadequately adhering to TB therapy can result in relapses to an infectious state of the disease or the development of MDR-TB. Both outcomes result in periods of infectiousness, during which they present a source of exposure to other residents and staff. In addition, environmental factors at homeless shelters, such as crowded living conditions and poor ventilation, facilitate the transmission of TB.

Outbreaks of TB among homeless shelter residents have been reported. For example, during 1990, 17 individuals with active pulmonary TB were identified among residents of homeless shelters in three Ohio cities: Cincinnati, Columbus, and Toledo (Ex. 7-51). In Cincinnati, 11 individuals with active TB were identified in a shelter for homeless adults. The index case was a man who had resided at the shelter and later died from respiratory failure. He was not diagnosed with TB until his autopsy. Of these 11 individuals, of which the index case was one, 7 were determined to be infectious. There was no indication as to whether any infection control measures were in place in the shelter. DNA analysis of 10 individual M. tuberculosis isolates showed identical patterns. The similarity among these DNA patterns suggested that transmission of the TB occurred in the shelter.

While the primary focus of this investigation was on the active cases reported among the residents in this Cincinnati shelter, the risk of transmission identified in this shelter also would apply to the shelter staff. Possible transmission of TB infection from the infectious individuals to the shelter staff might have been identified through tuberculin skin test conversions. However, no tuberculin skin test information for the staff was reported in this investigation.

Tuberculin skin testing results were reported in the investigation of a Columbus, Ohio shelter. In this investigation, a resident of a Columbus homeless shelter was identified with infectious pulmonary TB at the local hospital in March of 1990. The patient also had resided in a shelter in Toledo. As a result, a city-wide TB screening was initiated from April to May 1990 among the residents and staff of the city's men's shelters. Tuberculin skin tests were conducted on 363 shelter residents and 123 shelter employees. Among 81 skin-tested residents of the shelter in which the index case had resided, 32 (40 %) were positive compared to 47 (22 %) of 210 skin-tested residents of other shelters in Columbus who had positive skin test reactions. Similarly, among 27 employees of the shelter where the index case resided, 7 (26 %) had positive skin test reactions compared to 9 (11 %) of 85 employees in other men's shelters. These skin test results suggest an increased risk of transmission of TB among residents and employees of the homeless shelter where the index case resided. However, due to the lack of baseline skin test information among these residents and employees it is not possible to determine when their conversion to positive status occurred and whether this index case was their source of exposure. These results, however, do indicate a high prevalence of TB infection among homeless residents (e.g., 40 % and 22 %). Many of these individuals are likely to have an increased risk of developing active TB and, as a result, they may present a source of exposure to residents and staff.

The transmission of TB has also been observed among residents and staff of several Boston homeless shelters (Exs. 7-75 and 6-25). From February 1984 through March 1985, 26 cases of TB were confirmed among homeless residents of three large shelters in Boston. Nineteen of the 26 cases occurred in 1984, thus giving an incidence of approximately 317 per 100,000, 6 times the homeless case rate of 50 per 100,000 reported for 1983 and nearly 16 times the 1984 case rate of 19 per 100,000 for the rest of Boston (Ex. 6-25).

Of the 26 cases of TB reported, 15 had MDR-TB. Phage typing of isolates from 13 of the individuals with drug-resistant TB showed identical phage types, thus suggesting a common source of exposure. As a result of this outbreak, a screening program was implemented in November 1984 over a four-night period. Of 362 people who received skin tests, 187 returned for reading, 42 (22 %) were found to be positive and 3 were recent converters. Screening also was reported for the shelter staff at the three homeless facilities. At the largest of the three shelters, 17 of 85 (20 %) staff members had skin test conversions. In the other two shelters, 3 of 15 (20 %) and 3 of 18 (16 %) staff members had skin test conversions.

Whereas MDR-TB was primarily involved in the outbreak in Boston, an outbreak of drug-susceptible TB was reported in a homeless shelter in Seattle, Washington (Ex. 7-73). From December 1986 to January 1987, seven cases of TB from homeless residents were reported to the Seattle Public Health Department. The report of 7 individuals with active TB in one month prompted an investigation, including: (1) A mass screening to detect undiagnosed cases, (2) phage typing of isolates from shelter clients to detect epidemiologically linked cases, and (3) a case-control study to investigate possible risk factors for the acquisition of TB.

A review of the case registries revealed that 9 individuals with active TB had been reported from the homeless shelter for the preceding year and four cases in the year previous to that. As a result of the mass screening in late January 1987, an additional 6 individuals with active TB were detected. Phage typing of 15 isolates from the shelter-associated cases revealed that 6 individuals with active TB diagnosed around the time of the outbreak were of the same phage type, suggesting that there was a predominant chain of infection, i.e., a single source of infection. However, there also were other phage types, suggesting several sources of infection. Therefore, the investigators suggested that there was probably a mixture of primary and reactivated cases.

In addition to the similarity of phage types among TB cases, tuberculin skin testing results suggested the ongoing transmission of TB in the shelter. For example, 10 shelter clients who were previously tuberculin negative in May 1985 were re-tested in January 1987 and 3 (30 %) had converted. In addition, 43 clients who were negative in January 1987 were re-tested in June 1987 or February 1988 and 10 (23 %) had converted. Factors identified as contributing to the outbreak were the increased number of men with undiagnosed infectious pulmonary TB, the close proximity of beds in the shelter, and a closed ventilation system that provided extensive recirculation of unfiltered air.

As a result of the outbreak, a control plan was implemented. This plan included repetitive mass screening, repetitive skin testing, directly observed therapy, preventive therapy and modification of the ventilation system to incorporate UV light disinfection in the ventilation duct work. After the control plan was in place, five additional individuals with active TB were observed over a 2-year follow-up period.

While the primary focus in this study was on clients of the shelter rather than the shelter staff, the risk factors present in the shelter before implementation of the control plan would have also increased the likelihood for transmission of TB to shelter employees from infectious clients.

Thus, similar to correctional facilities, homeless shelters have a number of risk factors that facilitate and promote the transmission of TB (e.g., high incidence of infected residents with an increased likelihood of developing active disease, crowded living conditions and poor ventilation). Also, similar to correctional facilities, the evidence in homeless shelters shows that the failure to promptly identify homeless residents with infectious TB and the lack of appropriate TB control measures (e.g., lack of isolation precautions or prompt transfer to facilities with adequate isolation precautions) resulted in the transmission of TB to shelter employees.

Long-Term Care Facilities for the Elderly

Long-term care facilities for the elderly also represent a high-risk population for the transmission of TB. TB disease in persons over the age of 65 constitutes a large proportion of TB in the United States. Many of these individuals were infected in the past, before the introduction of anti-TB drugs and TB control programs when the prevalence of TB disease was much greater among the general population, and have harbored latent infection over their lifetimes. However, with advancing age, these individuals' immune function starts to decline, placing them at increased risk of developing active TB disease. In addition, they may have underlying disease or overall poor health status. Moreover, residents are often clustered together and group activities are often encouraged. TB case rates are higher for this age group than for any other. For example, the CDC reports that in 1987, the 6,150 cases of TB disease reported for persons ?65 years of age accounted for 27 % of the U.S. TB morbidity although this group only represented 12 % of the U.S. population (Ex. 6-14).

Because of the higher prevalence of TB cases among this age group, employees of facilities that provide long-term care for the elderly are at increased risk for the transmission of TB. More elderly persons live in nursing homes than in any other type of residential institution. The CDC's National Center for Health Statistics reports that elderly persons represent 88 % of the nation's approximately 1.7 million nursing home residents. As noted by the CDC, the concentration of such high-risk individuals in long-term care facilities creates a high-risk situation for the transmission of TB (Ex. 6-14).

In addition to having a higher prevalence of active TB, the recognition of TB in elderly individuals may be difficult or delayed because of the atypical radiographic appearance that TB may have in elderly persons (Exs. 7-59, 7-81, 7-82, and 7-83). In this situation, individuals with active TB may go undiagnosed, providing a source of exposure to residents and staff.

While the increased incidence of TB cases among the elderly in long-term care facilities may be a result of the activation of latent TB infections, the transmission of TB infection to residents and staff from infectious cases in the facilities has been observed and reported in the scientific literature.

For example, Stead et al. (1985) examined the reactivity to the tuberculin skin test among nursing home residents in Arkansas (Ex. 7-59). This study involved a cross-sectional survey in which tuberculin skin tests were given to all current nursing home residents. In addition, all newly-admitted nursing home residents were skin tested. For the three year period evaluated, 25,637 residents of the 223 nursing homes in Arkansas were tested.

Of 12,196 residents who were tested within one month of entry, only 12 percent were tuberculin positive, including those for whom a booster effect was detected. However, among the 13,441 residents for whom the first test was delayed for more than a month, 20.8 % were positive. In addition, the results of retesting 9,937 persons who were tuberculin negative showed an annual conversion rate of approximately 5 % in nursing homes in which an infectious TB case had been recognized in the last three years. In nursing homes with no recognized cases, the authors reported an annual conversion rate of approximately 3.5 %. The authors concluded that their data supported the contention that tuberculosis may be a rather common nosocomial infection in nursing homes and that new infections with tuberculosis is an important risk for nursing home residents and staff.

Brennen et al. (Ex. 5-12) described an outbreak of TB that occurred in a chronic care Veteran's Administration Medical Center in Pittsburgh. This investigation was initiated as a result of two skin test conversions identified through the employee testing program. One converter was a nurse working on ward 1B (a locked ward for neuropsychiatric patients) and the other was a physician working in an adjacent ward, 1U, who also had significant exposure to ward 1B. The source of infection in this investigation was traced to two patients who had resided on ward 1B and who had either a delayed or undiagnosed case of TB. The contact investigation revealed 8 additional conversions among patients, 4 in ward 1B and 4 in wards 2B and 4B (units on the floor above 1B).

Because the source cases were initially unidentified, no isolation precautions were taken. Smoke tracer studies revealed that air discharged from the window air conditioning unit of one of the source patients discharged directly into the courtyard. Air from this courtyard was the air intake source for window air conditioning units in the converters' room on ward 2B and thus was one of the possible sources of exposure.

In addition to the contact investigation on ward 1B and the adjacent units, hospital-wide skin testing results were evaluated. Of 395 employees tested, 110 (28 %) were positive. The prevalence in the surrounding community was estimated to be 8.8 %. Of those employees initially negative, 38 (12 %) converted to positive status. Included among these were employees in nursing (18), medical (3), dental (1), maintenance/engineering (3), supply (1), dietary (9), and clerical (2) services.

Occupational transmission of TB was also reported in a nursing home in Oklahoma (Ex. 6-28). In August 1978, a 68 year old female residing in the east wing of the home was diagnosed with pulmonary TB. She was subsequently hospitalized. However, by that time she had already had frequent contact with other residents in the east wing. As a result, a contact investigation, in which all residents of the home were given skin tests, was initiated.

The investigation revealed that the reaction rate for residents in the east wing (34/48, 71 %) was significantly higher than the reaction rates of residents living in the north and front wings (30/87, 34 %). No baseline skin test information was presented for the residents to determine the level of conversion. However, it was noted that half of the nursing home residents were former residents of a state institution for the developmentally disabled. A 1970 tuberculin skin test survey of that institution had shown a low rate of positive reactions.

In addition to the nursing home residents, nursing home employees were also skin tested. Of the 91 employees tested, 61 (67 %) were negative and 30 (33 %) were positive. Similar to results observed among the residents, positive reaction rates were higher for employees who had ever worked in the east wing (50 %) than for those who had never worked in the east wing (23 %). Retesting of the employees 3 months later revealed 3 conversions. These results suggested that there may have been occupational transmission of TB in this facility.

Occupational transmission has also been observed in a retrospective study of residents and employees who lived or worked in an Arkansas nursing home between 1972 and 1981 (Ex. 7-83). In this retrospective study, investigators reviewed the skin testing and medical chart data collected over a 10-year period at an Arkansas nursing home. Among the nursing home residents who were admitted between 1972 and 1982, 32 of 226 residents (17 %) who were initially tuberculin negative upon admittance became infected while in the home, based on conversion to positive after at least two previous negative tests. Twenty-four (63 %) of these conversions were infected in 1975, following exposure to one infectious resident. This resident, who had negative skin tests on three previous occasions during his stay in the home, was not diagnosed with TB until after he was hospitalized because of fever, loss of weight and productive cough. The remaining 37 % converted in the absence of a known infectious case. Thus, the authors suggested that nosocomial infections are likely to result from persons unsuspected of having TB.

Skin testing was also reviewed for employees of the nursing home. Questionnaires were completed by 108 full-time employees. Eleven of 68 employees with follow-up skin tests converted to positive skin status during the study period. Ten of the 11 (91 %) converters reported that they had been in the nursing home in 1975, the same year in which many of the residents were also found to have converted from a single infectious case. In addition, employees working at least 10 years in the home had a higher percentage of conversions (9 of 22, 40 %) than employees working less than 10 years (2 of 46, 4.4 %). Based on the results of this study, the authors concluded that, in addition to occurrence of TB cases from the reactivation of latent infections among the elderly, TB can also be transmitted from one resident to another resident or staff. Consequently, TB must be considered as a potential nosocomial infection in nursing homes.

Thus, long-term care facilities for the elderly represent a high-risk situation for the transmission of TB. These types of facilities possess a number of characteristics that increase the likelihood that active disease may be present among the facility residents and may go undetected. Similar to other high-risk settings, the evidence shows that the primary factors in the transmission of TB among residents and staff have been the failure to promptly identify residents with infectious TB and initiate and adequately implement appropriate exposure control measures.

Drug Treatment Centers

Another occupational setting that has been identified as a high-risk environment for the transmission of TB is drug treatment centers. Similar to other high-risk sites, drug treatment centers have a higher prevalence of TB infection than the general population. For example, in 1989 the CDC funded 25 state and city health departments to support tuberculin testing and administration of preventive therapy in conjunction with HIV counseling and testing. In this project, 28,586 clients from 114 drug treatment centers were given tuberculin skin tests. Of those, 2,645 (9.7 %) were positive (Ex. 6-8). When persons with previously documented positive tests were included, 4167 (13.3 %) were positive.

There is also evidence to suggest that drug dependence is a risk factor for TB disease. For example, Reichman et al. (Ex. 7-85) evaluated the prevalence of TB disease among different drug-dependent populations in New York: (1) An in-hospital population, (2) a population in a local drug treatment center, and (3) a city-wide population in methadone clinics. For the in-hospital population of 1,283 patients discharged with drug dependence, 48 (3.74 %) had active disease, for a prevalence rate of 3,740 per 100,000. In comparison, the TB prevalence rate for the total inpatient population was 584 per 100,000 and for New York City as a whole was 86.7 per 100,000. Screening of clients at a local drug treatment center in Harlem revealed a TB prevalence of 3750 per 100,000 in the drug-dependent population. Similarly, in the New York methadone program, the city-wide TB prevalence was 1,372 per 100,000. The authors also reported that although estimates of TB infection rates for both drug-dependent and non-drug dependent people were similar, the prevalence of TB disease among the drug-dependent was higher, thus suggesting that drug dependency may be a risk factor for disease.

Clients of drug treatment centers not only have a high prevalence of TB infection, a majority of them are intravenous drug users. Of the estimated 645,000 clients discharged each year from drug treatment centers, approximately 265,000 are intravenous drug users who either have or are at risk for HIV infection. In the Northeastern U.S., HIV seroprevalence rates of up to 49 % have been reported (Ex. 6-8). These individuals are at increased risk of developing active TB disease.

To determine the risk of active TB associated with HIV infection, Selwyn et al. (Ex. 5-6) prospectively studied 520 intravenous drug users enrolled in a methadone maintenance program. In this study, 217 HIV seropositive and 303 seronegative intravenous drug users, who had complete medical records documenting their history of TB and skin test status, were followed from June 1985 to January 1988. On admission to the methadone program, and at yearly intervals, all patients were given tuberculin skin tests.

Forty-nine (23 %) of the seropositive patients and 62 (20 %) of the seronegative patients had positive reactions to the skin test before entry into the study. Among the patients who initially had negative skin tests, 15 of 131 (11 %) seropositive patients and 62 of 303 (13 %) seronegative patients converted to positive tuberculin status. While the prevalence and incidence rates of TB infection were similar for the two groups of patients, seropositive patients showed a higher incidence of developing active disease. Active TB developed in 8 of the seropositive subjects with TB infection (4 %), whereas none of the seronegative patients with TB infection developed active TB during the study period.

Among individuals who are infected with HIV or who have AIDS, TB disease may be difficult to diagnosis because of the atypical radiographic appearance that TB may present in these individuals. In these individuals, TB may go undiagnosed and present an unsuspected source of exposure. Clients of drug treatment centers also may be more likely to discontinue or inadequately adhere to TB therapy regimens in instances where they develop active disease. As in other instances, this increases the likelihood of relapse to active disease or possibly the development of MDR-TB, both of which result in additional or even prolonged periods of infectiousness during which other clients or staff can be exposed.

There is evidence showing the transmission of TB in drug treatment facilities among both the clients and the staff. In a CDC case study (Ex. 5-6), a Michigan man who was living in a residential substance abuse treatment facility and was undergoing therapy for a previously diagnosed case of TB disease, was discovered by the local health department to have MDR-TB. As a result, a contact investigation was initiated at the drug treatment facility in which he resided.

Of the 160 clients and staff who were identified as potential contacts, 146 were tested and given tuberculin skin tests in November. No health screening program had been in place at the facility. The following March repeat skin tests were given. Of the 70 persons who were initially tuberculin negative and were still present in the facility, 15 (21 %) had converted to positive status (14 clients and 1 staff member). The investigators noted that the number of converters may have been underestimated for two reasons. Many of the clients were at risk for HIV infection and thus may have been anergic and not responded to the tuberculin skin tests. In addition, nearly half of the clients who were initially negative were not available for repeat skin testing.

Several factors may have contributed to the observed conversions in this facility. For example, no health screening program was in place. Therefore, individuals with TB would go unidentified. In addition, the clients were housed in a building with crowded dormitories for sleeping. The only ventilation in this building was provided by opening windows and doors. Thus, environmental conditions were ideal for the transmission of TB.

Consequently, the high-risk characteristics of clients who frequent these centers (e.g., high prevalence of infection and factors increasing the likelihood of developing active disease) and environmental characteristics of the center (e.g., crowding and poor ventilation), lead to drug treatment centers being considered a high-risk setting for the transmission of TB. The available evidence shows that the failure to promptly identify clients with infectious TB and to initiate and properly implement exposure control methods (e.g., proper ventilation) resulted in the infection of clients and staff at these facilities.

Conclusion

The available evidence clearly demonstrates that the transmission of TB represents an occupational hazard in work settings where employees can reasonably be anticipated to have contact with individuals with infectious TB or air that may reasonably be anticipated to contain aerosolized M. tuberculosis as a part of their job duties. Epidemiological studies, case reports, and outbreak investigations have shown that in various work settings where there has been an increased likelihood of encountering individuals with active TB or where high-hazard procedures are performed, employees have become infected with TB and in some cases developed active disease. While some infections were a result of more direct and more prolonged exposures, other infections resulted from non-direct and brief or intermittent exposures. Because of the variability in the infectiousness of individuals with active TB, one exposure may be sufficient to initiate infection.

Several factors, common to many of these work settings, were identified as contributing to the transmission of TB: (1) Failure or delayed recognition of individuals with active TB within the facility, and (2) failure to initiate or adequately implement appropriate infection control measures (e.g., performance of high-hazard procedures under uncontrolled conditions, lack of negative pressure ventilation, recirculation of unfiltered air, and lack of appropriate respiratory protection). Thus, in work settings where employees can reasonably be anticipated to have contact with individuals with infectious TB or air that may contain aerosolized M. tuberculosis and where appropriate infection control programs are not in place, employees are at increased risk of becoming infected with TB.

Infection with TB is a material impairment of the worker's health. Even though not all infections progress to active disease, infection marks a significant change in an individual's health status. Once infected, the individual is infected for his or her entire life and carries a lifetime risk of developing active disease, a risk they would not have had they not been infected. In addition, many individuals with infection undergo preventive therapy to stop the progression of infection to active disease. Preventive therapy consists of very toxic drugs that can cause serious adverse health effects and, in some cases, may be fatal.

Although treatable, active disease is also a serious adverse health effect. Some TB cases, even though cured, may result in long-term damage to the organ that is infected. Individuals with active disease may need to be hospitalized while they are infectious and they must take toxic drugs to stop the progressive destruction of the infected tissue. These drugs, as noted above, are toxic and may have serious side effects. Moreover, even with advancements in treating TB, individuals still die from TB disease. This problem is compounded by the emergence of multidrug-resistant strains of TB. In these cases, due to the inability to find adequate drug regimens which can treat the disease, individuals remain infectious longer, allowing the disease to progress further and cause more progressive destruction of the infected tissue. This increases the likelihood of long-term damage and death.

V. Preliminary Risk Assessment for Occupational Exposure to Tuberculosis
Introduction

The United States Supreme Court, in the "benzene" decision (Industrial Union Department, AFL-CIO v. American Petroleum Institute, 448 U.S. 607 (1980)), has stated the OSH Act requires that, prior to the issuance of a new standard, a determination must be made, based on substantial evidence in the record considered as a whole, that there is a significant health risk under existing conditions and that issuance of a new standard will significantly reduce or eliminate that risk. The Court stated that

"before he can promulgate any permanent health or safety standard, the Secretary is required to make a threshold finding that a place of employment is unsafe in the sense that significant risks are present and can be eliminated or lessened by a change in practices" (448 U.S. 642).

The Court in the Cotton Dust case (American Textile Manufacturers Institute v. Donovan, 452 U.S. 490 (1981)), rejected the use of cost-benefit analysis in setting OSHA health standards. However, the Court reaffirmed its previous position in the "benzene" case that a risk assessment is not only appropriate, but also required to identify significant health risk in workers and to determine if a proposed standard will achieve a reduction in that risk. Although the Court did not require OSHA to perform a quantitative risk assessment in every case, the Court implied, and OSHA as a matter of policy agrees, that assessments should be put into quantitative terms to the extent possible. The following paragraphs present an overall description of OSHA's preliminary quantitative risk assessment for occupational exposure to tuberculosis (TB).

An earlier version of this risk assessment was reviewed by a group of four experts in the fields of TB epidemiology and mathematical modeling. The reviewers were George Comstock, MD, MPH, DPH, Alumni Centennial Professor of Epidemiology, The Johns Hopkins University; Neil Graham MBBS, MD, MPH, Associate Professor of Epidemiology, The Johns Hopkins University; Bahjat Qaqish, MD, PhD, Assistant Professor of Biostatistics, University of North Carolina; and Patricia M. Simone, MD, Chief, Program Services Branch, Division of Tuberculosis Elimination, CDC. The reader is referred to the peer review report in the docket for additional details (Ex. 7-911). The revised version of OSHA's risk assessment, as published in this proposed rule, includes OSHA's response to the reviewers' comments as well as updated risk estimates based on recent purified protein derivative (PPD) skin testing data made available to the Agency since the peer review was performed and is generally supported by the reviewers or is consistent with reviewers' comments. (Note: PPD skin test and tuberculin skin test (TST) are synonymous terms.)

The CDC estimates that, once infected with M. tuberculosis, an untreated individual has a 10 % lifetime probability of developing active TB and that approximately half of those cases will develop within the first or second year after infection occurs. Individuals with active TB represent a pool from which the disease may spread. Based on data from the CDC, OSHA estimates that every index case (i.e., a person with infectious TB) results in at least 2 other infections (Ex. 7-269). For some percentage of active cases, a more severe clinical course can develop which can be attributed to various factors such as the presence of MDR-TB, an allergic response to treatment, or the synergistic effects of other health conditions an individual might have. Further, OSHA estimates that for 7.78 % of active TB cases, TB is expected to be the cause of death. Section 6(b)(5) of the OSH Act states that,

The Secretary, in promulgating standards dealing with toxic materials or harmful physical agents under this subsection, shall set the standard which most adequately assures, to the extent feasible, on the basis of the best available evidence, that no employee will suffer material impairment of health or functional capacity even if such employee has regular exposure to the hazard dealt with by such standard for the period of his working life.

For this rulemaking, OSHA defines TB infection as a "material impairment of health", for several reasons. First, once infected with TB, an individual has a 10 % lifetime likelihood of developing active disease and approximately 1 % likelihood of developing more serious complications leading to death. Second, allergic reaction and hepatic toxicity due to chemoprophylaxis with isoniazid, which is one of the drugs used in the recommended course of preventive treatment, pose a serious threat to a large number of workers. Third, defining infection with M. tuberculosis as material impairment of health is consistent with OSHA's position in the Bloodborne Pathogens standard and is supported by CDC and several stakeholders who participated in the pre-proposal meetings, as well as Dr. Neil Graham, one of the peer reviewers of this risk assessment. In his comments to OSHA, Dr. Graham stated,

The focus of OSHA on risk of TB infection rather than TB disease is appropriate. TB infection is a potentially adverse event, particularly if exposure is from a MDR-TB patient, or if the health-care or institutional worker is HIV seropositive. In addition, a skin test conversion will in most cases mandate use of chemoprophylaxis for > 6 months which is at least inconvenient and at worst may involve adverse drug reactions. (Ex. 7-271)

The approach taken in this risk assessment is similar to the approach OSHA took in its risk assessment for the Bloodborne Pathogens standard. As with bloodborne pathogens, the health response (i.e., infection) associated with exposure to the pathogenic agent does not depend on a cumulative level of exposure; instead, it is a function of intensity and frequency of each exposure incident. However, unlike hepatitis B, where the likelihood of infection once an exposure incident occurs is known with some degree of certainty, the likelihood of becoming infected with TB after an exposure incident is not as well characterized. With TB, the likelihood of infection depends on the potency of an exposure incident and the susceptibility of the exposed individual (which is a function of the person's natural resistance to TB and his or her health status). Further, the potency of a given exposure incident is highly dependent on several factors, such as the concentration of droplet nuclei in the air, the duration of exposure, and the virulence of the pathogen (e.g., pulmonary and laryngeal TB are considered more infectious than other types).

The Agency has sufficient data to quantify the risk associated with occupational exposure to TB among health care workers in hospitals on a state-by-state basis. In addition to hospital employee data, OSHA has obtained data on selected health care employee groups from the TB Control Office of the Washington State Health Department. These groups include workers employed in long-term health care, home health care, and home care. Small entities are encouraged to comment and submit any data or studies on TB infection rates relevant to their business.

Because it is exposure to aerosolized M. tuberculosis that places workers at risk of infection, and not some factor unique to the health care profession, the Agency concluded that the experience of these groups of health care workers is representative of that of the other "high-risk" workers covered by this proposal. This means that the risk estimates calculated for these groups of workers are appropriate to use as the basis for describing the potential range of risks for workers in other work settings where workers can be expected to come into close and frequent contact with individuals with infectious TB (or with other sources of aerosolized M. tuberculosis) as an integral part of their job duties. As discussed in section IV (Health Effects), epidemiological studies, case reports, and outbreak investigations have shown that workers in various work settings, including but not limited to hospitals, have become infected with tuberculosis as a result of occupational exposure to aerosolized M. tuberculosis when appropriate infection control programs for tuberculosis were not in place.

In this preliminary risk assessment, OSHA presents risk estimates for TB infections, cases of active disease, and TB-related deaths (i.e., where TB is considered the cause or a major contributing cause of death) for workers with occupational exposure to tuberculosis.

A number of epidemiological studies demonstrate an increased risk of TB infection among health care workers in hospitals and other work settings. A brief review of a selection of these studies is presented below, followed by OSHA's estimates of excess risk due to occupational exposure. Finally, OSHA presents a qualitative assessment of the risk of TB infection caused by occupational exposure to tuberculosis in correctional facilities, homeless shelters, drug treatment centers, medical laboratories, and other high-risk work groups.

Review of the Epidemiology of TB Infection in Exposed Workers

There are several studies in the published scientific literature demonstrating the occupational transmission of infectious TB. Reports of TB outbreaks and epidemiologic surveillance studies have shown that health care and certain other workers are, as a result of their job duties, at significantly higher risk of becoming infected than the average person.

OSHA conducted a thorough search of the published literature and reviewed all studies addressing occupational exposure to tuberculosis and TB infection in hospitals and other work settings. All published studies show positive results (i.e., workers exposed to infectious individuals have a high likelihood of becoming infected with TB). Because there are so many studies, OSHA selected a representative subset of the more recent studies conducted in the U.S. to include in this section. These studies were chosen because they show occupational exposure in various work settings, under various working conditions, and under various scientific study designs.

OSHA's summary of the studies is presented in Table V-1(a) and Table V-1(b). These studies represent a wide range of occupational settings in hospitals, ranging from TB and HIV wards in high prevalence areas, such as New York City and Miami, to hospitals with no known TB patients located in low prevalence areas such as the state of Washington. The studies include prospective studies of entire hospitals or groups of hospitals, retrospective surveys of well-controlled clinical environments, such as an HIV ward in a hospital, and case studies of single-source infection (i.e., outbreak investigations).

Outbreak investigations describe occupational exposure to tuberculosis from single index patients or a well-defined group of patients. Such investigations are more likely to demonstrate an upper limit of occupational risk in different settings, usually under conditions of suboptimal environmental and infection controls. Although outbreak investigations demonstrate the existence of occupational risk under certain conditions and the importance of the early identification of suspect TB patients quite well, these studies do not provide information conducive to risk assessment estimations. Limitations of outbreak investigations include the frequent absence of baseline PPD test results, the difficulty of extrapolating the results to non-outbreak conditions of TB exposure, and, often, small sample sizes. Table V-1(a) lists some of the published outbreak investigations and shows the risks posed to health care workers by such outbreaks, as well as the failures in control programs contributing to these episodes.

Prospective and/or retrospective surveillance studies are used to estimate conversion rates from negative to positive in PPD skin testing programs. These conversion rates can be used to estimate the excess incidence of TB infection. Surveillance studies among health care workers lend themselves to a more systematic evaluation of the risk of TB infection than outbreak investigations, for several reasons. First, these studies better reflect the risk of TB experienced by workers under routine conditions of exposure. Second, these studies are usually based on a larger group of workers and therefore yield more precise and accurate estimates of the actual risk of infection. However, the extent to which results from surveillance studies can be generalized depends on a careful evaluation of the study population. Some studies report skin test conversion rates for all workers in the hospital(s) under study. Such studies often include large groups of employees with little or no exposure to TB. Results from such studies may reflect an overall estimate of risk in that environment, but may underestimate the occupational risk of those with frequent exposure.

Other surveillance studies report PPD conversion rates of more narrowly-defined groups of workers, usually those working in "high-risk" areas within a hospital such as the HIV or TB wards. Some of these studies have internal control groups (i.e., they compare PPD conversion rates between a group of workers with extensive exposure to TB and a group of workers with minimal or no exposure to TB), thus making it possible to more precisely quantify the magnitude of excess risk due to occupational exposure. However, these studies are also limited in their usefulness for risk assessment purposes. They usually have small sample sizes, making it more difficult to observe statistically significant differences. More importantly, internal control groups may overestimate background risk, and thus underestimate excess occupational risk, unless painstaking efforts are made to eliminate from the control group those individuals with the potential for occupational exposure, a difficult task in some hospital environments. Table V-1(b) contains a selected list of published surveillance studies.

In reviewing Table V-1(a) and Table V-1(b), the reader should bear in mind that these tables are not intended to present an exhaustive list of epidemiologic studies with TB conversion rates in occupational settings. Instead, these tables present brief summaries of some of the epidemiologic evidence of occupational TB transmission found in the published literature; they are intended to convey the seriousness of the risk posed to health care workers and to illustrate how failures in control programs contribute to this risk. Upon reviewing these studies, a consistent pattern emerges: these work settings are associated with a high likelihood for occupational exposure to tuberculosis, and high rates of TB infection are being observed among health care workers.

Quantitative Assessment of Risk

Data availability usually dictates the direction and analytical approach OSHA's risk assessment can take. For this rulemaking, three health endpoints will be used: (1) TB infection, which is ``material impairment of health'' for this proposed standard; (2) Active disease following infection; and, (3) Risk of death from active TB.

In order to account for regional variability in TB prevalence and therefore to account for expected variability in the risk of TB infection in different areas, the Agency chose to develop occupational risk estimates on a state-by-state basis. This approach was criticized by Dr. Neil Graham as being too broad and "* * * insufficient in light of the tremendous variability * * * that can occur within a state." (Ex. 7-911). The Agency recognizes that risk estimates on a county-by-county basis would be preferable; however, the unavailability of comprehensive county data has prevented the Agency from conducting such analysis.

The annual excess risk of TB infection due to occupational exposure is defined as a multiplicative function of the background rate of infection and is expressed as:

where:

p is the annual excess risk due to occupational exposure,
R_b is the background rate of TB infection, and
ERR_o is a multiplicative factor denoting the excess relative risk due to occupational exposure
(ERR_o).

Estimates of ERR_o are derived from surveillance studies of workers with occupational exposure to TB. ERR_o is defined as the relative difference between the overall exposed worker risk and the background (population) risk and is calculated as the difference between overall worker and background risk divided by the background risk.

The annual excess risk due to occupational exposure is defined as a function of the background risk because of data limitations. If data on overall worker risk were available for each state, then the excess risk due to occupational exposure would simply be the difference between overall worker risk and background risk. Instead, the annual excess risk due to occupational exposure (i.e., p) is estimated using a multiplicative model because data on overall worker risk (i.e., R_w) were available only for the states of Washington, and North Carolina and for Jackson Memorial Hospital located in Miami, Florida. Therefore, the annual excess risk due to occupational exposure in state i (p_i) is expressed as:

where:

R_wj is the overall worker risk estimated from surveillance studies (study j),

R_bj is the study control group risk (i.e., study background risk), and

R_bi is the background rate for state i.

When i=j (i.e., Washington State or North Carolina), the excess risk due to occupational exposure, is expressed as the straight difference between overall worker risk and background risk.

OSHA calculated estimates of ERR_o based on three occupational studies: the Washington State study, the North Carolina study, and the Jackson Memorial Hospital study (Exs. 7-263, 7-7, 7-108). These estimates were expressed as percent change above each study's background. The derivation of these estimates is described in section 2.

In order to estimate an overall range of occupational risk of TB infection, taking into account regional differences in TB prevalence in the U.S., OSHA: (1) Estimated background TB infection rates by state (R_bi), and (2) applied estimates of ERR_o, derived from the occupational studies, to the state background rates to calculate estimates of excess risk due to occupational exposure by state.

OSHA used a multiplicative function of each state's background infection rate to estimate excess risk of TB infection because the probability of occupational infection can be viewed as a function of the number of contacts and frequency of contacts with infectious individuals. Thus, estimates of expected relative increase in risk above background due to occupational exposure are calculated for the three available studies and these relative increases (i.e. ERR_o) are multiplied by background rates for each state to derive estimates of excess occupational risk by state. These state estimates are then used to derive a national estimate of occupational risk.

The CDC compiles and publishes national statistics on the incidence of active TB in the U.S. by state based on reported cases. OSHA relied on these data to estimate TB infection background rates through the use of a mathematical model because information on TB infection is not being collected nationwide by CDC. A more detailed discussion on the methodology and derivation of background risk estimates by state is found in section 3, and discussion on the estimation of occupational risk estimates by state is found in section 4 of this risk assessment.

Because section 6(b)(5) of the OSH Act requires OSHA to assess lifetime risks, OSHA has converted the annual excess risk due to occupational exposure into an excess lifetime risk based on a 45-year working lifetime. The formula used to calculate lifetime occupational risk estimates of the probability of at least one occurrence of TB infection due to occupational exposure in 45 years is expressed as { 1-(1-p)⁴⁵ }, where p is the annual excess risk due to occupational exposure. Two assumptions are critical in defining lifetime risk: (1) the exposure period is 45 years, and (2) the annual excess risk remains constant. The implication of the second assumption is that the worker's exposure profile and working conditions, which may affect the level and intensity of exposure, and the virulence of the pathogen, remain unchanged throughout a working lifetime. The merit of this assumption was questioned by Dr. Graham, because, as he states "* * * patient contact may vary greatly throughout a career for many HCWs [health care workers]." and " * * * physicians (and nurses) often do not have extensive patient contact until [their] mid-twenties, while other workers increasingly retire early." Dr. Graham recommends that OSHA's risk assessment be adjusted to account for variable exposure levels and variable working lifetimes. Although accounting for variable exposure levels could result in more precise risk estimates, the unavailability of comprehensive information on lifetime TB exposure scenarios by occupational group prevented the Agency from developing a more complex risk model.

OSHA has customarily assumed a 45 year working lifetime in setting health standards. The Agency believes that this assumption is reasonable and consistent with the Act. The Act requires the Secretary to set a standard for toxic substances that would assure "no employee * * * suffer material impairment of health or functional capacity even if such employee has regular exposures to the hazard for the period of his working lifetime." 29 U.S.C. § 655(b)(5) (emphasis added). The U.S. Court of Appeals for the District of Columbia upheld the use of a 45-year lifetime in the asbestos standard against an assertion by the Asbestos Information Association that the average duration of employment was five years. Building and Construction Trades Department, AFL-CIO v. Brock, 838 F.2d 1258, 1264, 1265 (D.C. Cir. 1988). The Court said that OSHA's assumption "appears to conform to the intent of Congress" as the standard must protect even the rare employee who would have 45 years of exposure. Id. at 1264. In addition, while working lifetimes will vary, risk is significant for some who work as little as one year and, at any rate, individual and population risks are likely to remain the same so long as employees who leave one job are replaced by others, and those who change jobs remain within a covered sector. Nevertheless, the Agency solicits information regarding the likelihood of exposure to active TB in the workplace and duration of employment in various occupational groups. Lifetime risk estimates of TB infection by state are described in section 4.

Lifetime risk estimates of developing active TB are calculated from lifetime risk estimates of TB infection assuming that, once infected, there is a 10 % likelihood of progressing to active TB. These estimates are discussed in section 4. Further, the number of deaths caused by TB is calculated from the lifetime estimates of active TB using OSHA's estimate of TB case fatality rate, also discussed in section 4.

1. Definitions

For the purpose of estimating incidence rates, TB infection rate is defined as the annual probability of an individual converting from negative to positive in the tuberculin skin test. Annual occupational risk is defined as the annual excess risk of becoming infected with TB due to occupational exposure, and is estimated as a function of the background risk. Lifetime occupational risk is defined as the excess probability of becoming infected with TB due to exposure in the workplace, at least once, in the course of a 45-year working lifetime and is estimated as { 1-(1-p)⁴⁵ } where p is the annual occupational risk of TB infection.

2. Data Sources for Estimating Occupational Risk

The quantitative data needed to develop an overall national estimate of risk for TB infection due to occupational exposure are not available. The CDC does not publish occupational data associated with TB infection incidence and active TB on a nationwide basis. There has been some effort to include occupational information on the TB reporting forms, but only a limited number of states are currently using the new forms that capture occupational information in a systematic way.

However, there are a number of sources that permit the risk in occupational settings to be reasonably estimated and, with the aid of mathematical models, to develop estimates of excess relative occupational risk (ERR_o), which can then be multiplied by the state-specific background rates to yield estimates of excess occupational risk. OSHA has identified three data sources that are suitable for assessing the excess risk of TB infection in health care workers with occupational exposure. These include: (1) A 1994 survey of tuberculin skin testing in all health care facilities in Washington State; (2) A state-wide survey of hospitals in North Carolina, conducted in 1984-1985, which addressed TB skin testing practices, TB infection prevalence, and TB infection incidence among hospital employees in that state; and (3) the employee tuberculin skin test conversion database from Jackson Memorial Hospital in Miami, Florida. In addition to these hospital employee data, the Agency has obtained data on selected other work groups from the state of Washington. These groups include workers employed in long-term health care, home health care, and home care.

On the issue of data availability for this risk assessment, Dr. Graham agrees with OSHA that there are no comprehensive data available with respect to occupational risk of TB infection in health care and other institutions in the U.S. Instead of relying on two state specific studies, Dr. Graham recommends, though with serious reservations, the use of a review study by Menzies et al. (Ex. 7-130). Dr. Graham admits that the "validity of the estimates in these reports [reviewed in the Menzies et al. study] must be open to serious question * * *" for the following reasons, which were pointed out by Dr. Graham: several of the studies reviewed are very old and not relevant to TB risk in the 1990s; four studies use tine tests and self-reports of skin test results, which are not useful for estimation of risk of TB infection; the studies were not consistent in the inclusion of high and low risk workers; two-step testing was not done; and the participation rates were extremely low or unreported in many of the studies included in this review.

OSHA has chosen not to rely on the Menzies et al. review study, because, in addition to Dr. Graham's reservations (which the Agency shares), OSHA is also concerned about the inclusion in the Menzies et al. review article of studies conducted outside the U.S. Factors known to affect the epidemiology of TB, such as environmental conditions, socio-economic status, and work practices, are expected to differ greatly from one country to another, and are not controlled for in the statistical analyses of these studies. For all of these reasons, the Agency has chosen to rely solely on U.S. studies for its quantitative risk estimations.

Estimates of excess risk due to occupational exposure are expressed as the percent increase above background based on relative risk estimates derived from occupational studies. Internal control groups provided estimates of background risk for the Washington state and Jackson Memorial data sets. In the absence of a suitable internal control group, the estimated annual state-wide TB infection rate, as calculated in Section 3, was used as the background rate in the North Carolina study.

(a) Washington State Data: Initially, OSHA relied on a three-year prospective study, conducted between 1982 and 1984 in the state of Washington, to derive an estimate of excess risk for TB infection as a result of occupational exposure (Ex. 7-42). OSHA received several objections to the use of this study. The study used hospitals with no known TB cases as "controls" based on the assumption that in those hospitals the risk of TB infection to employees may be the same as for the general population. Dr. Qaqish noted that this assumption is highly questionable and that the use of such controls is not appropriate. Dr. Graham and Dr. Qaqish pointed out that the published results did not include conversions identified through contact investigations, which means that the conversion rate reported in that study was likely to be an underestimate of the true risk. In addition, the commenters noted that the study was designed to estimate the effectiveness of the TB screening program and may have produced skin testing results biased toward the null; the study is relatively old; and, the study was conducted prior to the AIDS epidemic and therefore the results may not be relevant to the occupational risk at present because the relationship between HIV and TB is not reflected in this study.

In an effort to respond to reviewers' comments, the Agency chose to update the analysis by relying on a data set of tuberculin skin testing results from a survey of the state's tuberculin skin testing program in 1994. This survey is conducted by the TB Control Office in the Washington State Health Department and it covers all hospitals in the state, as well as long-term care, home health care, and home care facilities. OSHA was given access to the database for the 1994 survey as well as data on conversions identified through contact investigations for the same year (Ex. 7-263). Table V-2 summarizes the results of the 1994 survey. Of the 335 health care establishments in the state of Washington, 273 responded to the survey, for an overall response rate of 81.5 %. Of those, 76 were hospitals, 142 were long-term care, 47 were home health care, and 8 were home care facilities. Hospitals had the highest survey response rate (85 %) and home health care had the lowest (77 %). Every employee at risk for TB infection (i.e., who was known to be tuberculin skin test negative at the start of the study period) in the participating hospitals and long-term care facilities was given a tuberculin skin test, including administrators, housekeepers, business office staff, and all part-time employees. Testing in home health care facilities was generally confined to those nursing staff who had direct client contact. Employees in home care are those who provide services to patients in home health care and include food handlers, cleaning aides, personal care-givers, and some social workers.

The overall rate of skin test conversion for workers in the health care system in Washington State in 1994 was 3.24 per 1,000 employees tested. This is greater than a 4-fold increase from the estimated state-wide background rate of 0.69 per 1,000 at risk, as calculated in section 3. The annual rate of TB conversion ranged from 1.27 per 1,000 tested for hospital employees to 9.80 per 1,000 tested for long-term care employees.

The annual rate of 9.8 per 1,000 for long-term care employees probably reflects the high potential for exposure to undiagnosed active TB in those facilities. As a rule, long-term facilities in Washington State do not have AFB isolation rooms. Therefore, residents with no obvious TB symptoms but who might be infectious spend most of their time in open spaces exposing other residents and workers to infectious droplet nuclei. However, once a resident has been identified as a suspect TB patient, that person is transferred to a hospital until medically determined to be non-infectious.

Also, since employees who were 35 years of age or younger were not given a two-step test at hiring, and a high percentage of employees are foreign born and therefore most likely to have been vaccinated during childhood with the BCG vaccine, some of the conversions observed might be late boosting because of BCG. However, an almost two-fold increase in risk for long-term care workers even as compared to the significant excess risk among home health care workers clearly indicates that the risk of TB infection for workers in long-term care is high and not likely to be fully explained by late boosting. Beginning in 1995, two-step testing has been done on all new hires in Washington State. Thus, tuberculin skin testing data for 1995 are not expected to be influenced by possible late boosting; OSHA will place the 1995 data in the rulemaking record as they become available.

Hospital workers had the lowest overall rate of conversion (overall rate of 1.27 per 1,000). This, in part, can be attributed to the existence of extensive TB control measures in that environment in Washington State. Compliance with the CDC Guidelines and OSHA's TB Compliance Directive is quite high in Washington State because: (a) There is a strong emphasis on early identification of suspect TB patients; (b) there is a strong emphasis on employee training and regular tuberculin skin testing (although on a less-frequent basis than recommended in the Guidelines: All employees are tested at hire and annually thereafter); (c) the use of respirators is expected when entering an isolation room; and (d) all isolation rooms are under negative pressure, have UV lights, and exhaust to the outside. In addition, conversion data in hospitals are more likely to represent true TB infections than in the other health care settings, because hospitals are more likely to re-test converters in an effort to eliminate false-positive cases.

A more thorough analysis of the hospital data is presented in table V-3. Because the Washington State survey was not designed to compare exposed persons with matched controls who have had no exposure, several alternative definitions of an internal control (unexposed) group were used in analyzing this data set. Three different analyses, shown in table V-3, produced estimates of annual occupational infection rates ranging from 0.4 to 0.6 per 1,000 above control (i.e., ranging from a 47 % to an 84 % increase above control). In order to minimize the likelihood of contaminating the control group with persons having significant occupational exposure, OSHA defined the control group as workers in hospitals located in zero-TB counties and with no known TB patients. This analysis is summarized in table V-3 as Definition 1. If potential for occupational exposure is defined as either working in a hospital in a county that has active TB or in a hospital that has had TB patients, then the annual risk due to occupational exposure is 47 % above background. The excess annual risk due to occupational exposure appears to be approximately 60 % above background, if workers in hospitals with a transfer-out policy for TB patients are considered to be the control group, shown as Definition 2 in table V-3. A 60 % increase above background is not statistically significantly different from a 47 % increase and therefore these two "control" groups can be viewed as producing "statistically" equivalent results. However, the Agency believes that Definition 1 is more appropriate, though the risk estimates are higher if the control group is defined based on Definition 2, because there is a higher likelihood of potential for exposure to a patient with undiagnosed TB under Definition 2 conditions. Comparisons of all hospital TST data to the state-wide estimate of TB infection rate resulted in an estimate of the annual excess occupational risk of approximately 84 % above background, shown in table V-3 as Definition 3. Estimates of the annual and lifetime occupational risk of TB infection for the average health care worker in hospitals by state, extrapolated from this study and using Definition 1 as the control group, are presented and summarized in section 4.

Annual rates of excess risk due to occupational exposure were estimated for long-term care, home health care, and home care and are presented in Section 4. The same control group used in the hospital data analysis, Definition 1 (i.e., 0.876/1,000 workers at risk) was used to estimate the background risk among workers in long-term care, health care, and home care facilities and settings. Using 0.876 as the background infection rate for workers in these settings (a) provided a level of consistency among the Washington data analyses, and (b) resulted in a lower estimate of occupational risk for the non-hospital health care workplaces than would have resulted had the state-wide background risk estimate (i.e., 0.67/1,000 see Section 3) been used. When industry-specific risk data are used, there is approximately a 10-fold increase in annual risk for workers in long-term care, a 5-fold increase in annual risk for workers in home health care, and a 1-fold increase in annual risk for workers in home care (see Section 4).

Estimates of the range of annual and lifetime occupational risk for the average health care worker in long-term care, home health care, and home care by state, extrapolated from the Washington State study, are presented in Section 4.

(b) North Carolina Study: A state-wide survey of all hospitals in North Carolina (NC) was conducted in 1984-1985 (Ex. 7-7). The survey's questionnaire was designed to address three main areas of concern affecting hospital employees: (1) Tuberculin skin testing practices; (2) TB infection prevalence; and (3) TB infection incidence. The incidence of new infections among hospital personnel was assessed over a five-year period by reviewing tuberculin skin test conversion data during calendar years 1980 through 1984 and was calculated as the number of TB skin test conversions divided by the number of skin tests administered. (Since most employees were only given annual testing, the number of tests administered is a very close estimate of the total number of people tested within a year and thus can be used as the denominator in estimating infection incidence.) Only 56 out of 167 hospitals reported information on TB conversion rates (34 % response rate). The authors estimated a state-wide TB infection rate of 11.9 per 1,000 per year for hospital employees in 1984 and a five-year mean annual infection rate of 11.4 per 1,000, with a range of 0-89 per 1000 employees at risk for TB infection. An analysis of the data by region (i.e., eastern, central, western) showed that the eastern region had consistently higher rates (with an average infection rate of 18.0 per 1,000) followed by the central region (7.0 per 1,000) and the western region (6.1 per 1000). Results of this study are shown in table V-4.

Use of this study's overall results for risk estimates was criticized by the peer reviewers because of design flaws in the study (e.g., high non-response rate, inconsistent skin testing practices, and limited two-step testing) and, most importantly, the presence of atypical mycobacteria (contributing to false positive results) in the eastern part of the state. Based on further input from Dr. Comstock, the Agency chose to rely on the study results from the western region only, because they are considered to be more representative of the "true" risk of infection and are expected to be less confounded by cross-reactions to atypical mycobacteria. Further, the Agency chose to rely on the conversion rate estimated for 1984 because it was the most recent data reported in the study. Therefore, the western region conversion rate of 7.2 per 1,000, estimated based on responses to the survey from eight hospitals in 1984, was used as an overall worker conversion rate. Further, the 1984 rate was adjusted by the percent decrease of active TB between 1984 and 1994 in North Carolina so that the final worker conversion rate for 1994 based on the western region rates reported in this study was estimated to be 5.98 (7.2 * 532/641 = 5.98) per 1,000 employees at risk for TB infection.

The North Carolina study did not have an internal control group to use as the basis for estimating excess risk due to occupational exposure because the conversion rates presented in this study were based on TST results for the entire hospital employee population. In the absence of an internal control group, the Agency used the estimated state-wide background rate of 1.20 per 1,000 as the background rate of infection for the western region in North Carolina (see Section 3) to estimate excess risk due to occupational exposure.⁽¹⁾ Based on this study, annual occupational risk is approximately four times greater than background [(5.98-1.2)/1.2 = 3.98]. Estimates of the annual and lifetime occupational risk of TB infection based on this study by state are presented in Section 4.

(c) Jackson Memorial Hospital Study: Jackson Memorial Hospital (JMH) is a 1500-bed general facility located in Miami, Florida, employing more than 8,000 employees. It is considered one of the busiest hospitals in the U.S. It is the primary public hospital for Dade County and the main teaching hospital for the University of Miami School of Medicine. JMH treats most of the TB and HIV cases in Dade County and, consequently, there is a higher likelihood of occupational exposure to TB in this facility than in the average hospital in the U.S. From March 1988 to September 1990, an outbreak of multidrug-resistant TB (MDR-TB) occurred among patients and an increased number of TST conversions was observed among health care workers on the HIV ward. This prompted a re-evaluation of the hospital's infection control practices and the installation of engineering controls to minimize exposure to TB. As part of the evaluation of the outbreak, NIOSH did a Health Hazard Evaluation and issued a report (Ex. 7-108). In addition, NIOSH conducted a retrospective cohort study of JMH to determine whether the risk of TB infection was significantly greater for health care workers who work on wards having patients with infectious TB than those who work on wards without TB patients.

For the data analysis of this study, "potential for occupational exposure" was defined based on whether an employee worked on a ward that had records of 15 or more positive cultures for pulmonary or laryngeal TB during 1988-1989. In other words, positive culture was taken as a surrogate for exposure to infectious TB. The authors restricted the "exposed" group to employees on wards with exposures to pulmonary or laryngeal TB because they intended to restrict the study to hospital workers with exposure to patients with the highest potential for being infectious. There were 37 wards at JMH that had submitted at least one positive culture during 1988-1989. Seven wards met the criteria of 15 or more and were therefore included in the "exposed" group. These were the medical intensive care unit, five medical wards, and the emergency room. The "control" group was defined as hospital workers assigned to wards with no TB patients (i.e., wards with no records of positive cultures during 1988-89). The "control" wards were post-partum, labor and delivery, newborn intensive care unit, newborn intermediate care unit, and well newborn unit. The results of this analysis are presented in Table V-5.

Table V-5 shows a substantially elevated risk for those workers with potential exposure to patients with infectious TB. The relative risk ranges from 9 to 11.7 between 1989 and 1991 and is statistically significant for all of those years. This suggests that the excess risk due to occupational exposure is approximately 8-fold above background; this is an overall risk estimate that reflects the occupational risk of TB infection for JMH employees with patient contact, because this analysis included everyone tested in the "exposed" and "control" group, regardless of his or her specific job duties or length of patient contact.

An analysis of various occupational groups within this cohort showed that nurses and ward clerks in the "exposed" groups had the highest conversion rates: 182 and 156 conversions per 1,000 workers tested, respectively. Other studies have shown that health care workers who provide direct patient care are at greater risk for infection than workers who do not provide direct patient care. The high risk seen in ward clerks was unexpected since these workers are not involved in direct patient care. However, in the emergency room, the risk for TST conversion for the ward clerks was almost three times higher than for the nurses, 222 and 83 per 1,000, respectively. Ward clerks in the emergency room are responsible for clerical processing of patients after triage, handling specimens for the laboratory, and gathering clothing and valuables from admitted patients. During these interactions, there may have been less strict adherence to infection control measures, and this could explain the high conversion rate.

OSHA used the results from the 1991 analysis of the data in the JMH study to estimate occupational risk of TB infection in hospital workers with a relatively high likelihood of occupational exposure, for the following reasons: (a) 1991 represents the most recent year for which conversion data are available prior to the time when TB infection control measures were fully implemented at JMH; and (b) The higher conversion rates reported for 1990 and 1989 (75.5 and 62.2 per 1,000 respectively) may be atypical, i.e., they may to some extent reflect the effect of the outbreak and not the long-term occupational risk.

Based on the results of this study, OSHA estimates that the annual excess risk of TB infection due to occupational exposure is 7.95 times greater than background. Estimates of annual and lifetime occupational risk of TB infection for the average health care worker in hospitals by state, extrapolated from this study, are presented and summarized in section 4.

3. Estimation of Background Risk of TB Infection

OSHA's methodology for estimating population (background) TB infection rates relies on the assumption that TB infection occurring in an area can be expressed as a numerical function of active TB cases reported in the same area. If the likelihood of observing any infection in a population is minimal, then the likelihood of observing active disease diminishes. Conversely, the presence of active TB implies the presence of infection, since active disease can only progress from infection. Therefore, there is a functional relationship linking TB infections to active disease being observed in a particular area during a specified time period.

Peer reviewer comments on this assumption varied. Neil Graham states in his comment "Although factors such as migration and distribution of the population may influence this relationship it seems probable that this assumption is largely correct and justifiable." (Ex. 7-271). On the other hand, Dr. Simone expresses concern over this assumption and states "It is not necessarily true that a change in cases now reflects the risk of infection now." Dr. Qaqish demonstrates in his comment that the net effect of assuming a proportional relationship between the number of active cases and the number of new infections is to introduce a possible bias into the estimate of background risk of TB infection, although such a bias could work in either direction, i.e., toward increasing or decreasing the estimate of risk. Dr. Qaqish further states that in the absence of more "relevant data," it is not possible to determine the actual net effect in magnitude and direction of the bias and "without obtaining additional data, it would be impossible for the Agency to improve on the accuracy of the risk estimates * * * " OSHA has considered all of the reviewer comments and is aware of the inherent uncertainty and the potential for bias associated with the use of this assumption; however, in the absence of the additional "relevant" data to which Dr. Qaqish refers, the Agency believes this approach to be justifiable.

In defining the model used to estimate the annual infection rates occurring in a geographical area based on data on active disease cases reported for the same area, infections progressing to active disease are assigned to one of three distinct groups: those occurring this year, last year, and in previous years.

TB cases reported to CDC each year are a combination of new and old infections that have, for various reasons, progressed to active disease. Until recently, it was believed that most of the active cases were the product of old infections. However, with the use of DNA fingerprinting techniques, researchers have reported that a larger percentage of active cases may be attributed to new or recent infections. Small et al. reported, in an article on tracing TB through DNA fingerprinting, that as many as 30 % of the active cases reviewed in the study may be the result of recent infections (Ex. 7-196).

In this risk assessment, the Agency assumes the lifetime risk that an infection will progress to active TB to be approximately 10 %. This estimate is supported by CDC and in her comment, Dr. Simone states that: "The assumption * * * is generally agreed upon." Dr. Comstock and Dr. Qaqish both questioned the validity and accuracy of CDC's estimate. Their comments suggest that the true lifetime rate of progression from infection to active disease for adults may be less than 10 percent. However, as Dr. Graham points out, the 10 % assumption is a widely accepted "rule of thumb" and is also in relative agreement with data from the unvaccinated control group of the British Medical Research Council (MRC) vaccination trial in adolescents (Ex. 7-266).

In the MRC study, 1,338 adolescents' skin tests converted following TB exposure where the precise date of conversion was known. Of these, 108 (8.1 %) individuals developed active TB during follow-up. Of these, 54 % developed active TB within one year and 78 % within 2 years. This results in a risk of approximately 4 % at one year, 6 % at two years, and an overall risk of 8 %. Given that the risk of TB reactivation increases with age, the lifetime risk is expected to be higher than the 8 % attained in this study and, as Dr. Graham points out, a 10 % overall lifetime risk seems reasonable.

Based on Dr. Graham's recommendation to rely on the progression rates from the MRC study, OSHA changed the assumption on the progression parameters from 2.5 % (first year), 2.5 % (second year), and 5 % (remaining lifetime) to 4 %, 2 % and 4 %, respectively. Therefore the total 10 % progression from infection to active disease is partitioned into 3 groups: progression during the first year after infection (40 % of all infections that eventually progress, for a net probability of 4 %), progression during the second year (20 % of all infections that eventually progress, for a net probability of 2 %), and progression during all subsequent years (the remaining 40 % of progressing infections). This last probability (4 %) is assumed to be uniformly distributed across the remaining lifespan.

TB rates vary considerably by geographic area, socio-economic status, and other factors. In an attempt to account for some of those factors, to the extent possible, background TB infection rates have been estimated separately for each state. The derivation of background infection rates involves several steps for which the process and formulae are presented below.

Step 1: Background rate of TB infection for state i in year j is defined as:

where:

B_i(j) is the background TB infection rate for state i in year j

I_i(j) is an estimate of the number of new infections that occurred in state i in year j

X_i(j) is the population at risk for TB infection in state i in year j.

Step 2: Estimation of I_i(j), the number of new TB infections:

Let:

A_i(j) be the total number of adult TB cases reported to CDC by state i in year j.

A_(j) be the total number of adult TB cases reported to CDC by all states in year j.

P_i(j) be the estimated prevalence of adult TB infection in state i during year j.

R_i be the ratio of the number of adult TB cases reported in 1993 to the number of adult cases reported in 1994 in state i.

The number of TB cases reported in 1994 can be expressed as a function of TB infections expected to have progressed to active disease, by the following formula:

where j ranges from 1919 to 1992. The quantity inside the summation symbol is the sum of all people who were infected with TB between 1919 and 1992 and are still alive in 1994. This summation can be approximated by the prevalence of TB infection in 1992, P_i(1992). Therefore, the number of active TB cases reported in 1994 can be expressed as:

Further, if we assume that the number of new infections is directly proportional to the number of active cases, then I_i(1993) can be expressed as follows:

and (2) can be expressed as:

then solving for I_i(1994) becomes: ⁽²⁾

Step 3: Estimation of X_i(1994):

Where:

N_i is the adult population for state i as reported by U.S. Census in 1994.

P_i(1993) is the estimated number of infected adults in state i in 1993 (i.e., prevalence of TB infection in state i among adults).

To estimate the number of adults currently at risk for TB infection in each state, the number of already infected adults (i.e., prevalence of TB infection P_i in 1993) is subtracted from the adult population in 1994.

Step 4: Estimation of population currently infected as of 1993 by state, P_i(1993):

The prevalence of TB infection in each state is estimated as a function of TB infection prevalence in the U.S. in 1993 and the percent TB case rate for each state.

Where:

P₍₁₉₉₃₎ is the prevalence of TB infections in the U.S. in 1993 (Ex. 7-66) and

A₍₁₉₉₃₎ is the total number of adult TB cases reported in 1993.

Estimates of TB infection prevalence in the U.S. were developed for OSHA by Dr. Christopher Murray of the Harvard Center for Population and Development Studies and are presented in Table V-6 (Ex. 7-267). The mathematical model used by Dr. Murray to estimate TB infection prevalence has been designed to capture the transmission dynamics of TB by modeling transfers between a series of age-stratified compartments using a system of differential equations. The model adjusts for various epidemiological factors known to influence the course of active TB, such as onset of infection (i.e., old vs. new infections) and the impact of immigration rates and the HIV epidemic. However, it does not differentiate among gender or race categories. The model has been successfully validated using actual epidemiological data on active TB from 1965 to 1994. The estimates of TB prevalence rates presented here are specific for adults (i.e., older than 18 years of age), which make them more appropriate for estimating risk of transmission in an occupational setting.

To estimate the number of previously infected adults in each state (P_i), the estimated national TB prevalence figure was multiplied by the active cases for each state and divided by the total number of active cases reported [see equation (7)] (i.e., the national prevalence estimate was apportioned among the states based on each state's percent contribution to active TB reported for 1993). To estimate the number of adults at risk of TB infection, (X_i), the number of already infected adults was subtracted from the adult population estimate for each state (see equation (6)). The number of new infections expected to have occurred in 1994 was estimated using equation (5).

The background rate of TB infection for 1994 was then estimated by dividing the number of new infections (I_i) by the number of susceptible adults in each state (X_i) (see equation (1)).

Results on estimated TB background annual infection rates for each state are presented in Table V-7(a) -- Table V-7(c). In Table V-7(a) TB infection rates are based on an average value of TB infection prevalence, as estimated by Dr. Murray, in the U.S. (i.e., 12,667,062). In Table V-7(b), infection rates are based on the minimum value of TB infection prevalence in the U.S. (i.e., 12,037,524). In Table V-7(c), infection rates are based on the maximum value of TB infection prevalence in the U.S. (i.e., 13,296,599). An overall range of background annual TB infection rates was constructed by combining all three sets of infection rates and was estimated to be between 0.194 and 3.542 per 1,000 individuals at risk of TB infection, with a weighted average of 1.46 per 1,000 using state population size as weights.

Step 5 Model validation:

An alternative, but less sophisticated, way to estimate annual risk of infection, if prevalence is known in a specific age group, is to use the following formula:

Annual Rate of Infection = -ln(1-P)/d (8)

Where:

P is the percent prevalence of infection and

d is the average age of the population (Ex. 7-265).

In order to validate the model used by OSHA to estimate background infection rates, estimates of TB infection prevalence for 1994 were used to calculate predicted infection rates using equation (8). Based on Murray's model, TB infection prevalence is expected to range from 6.31 % to 6.97 % in 1994 among adults (18+). Using these figures and assuming the average age to be 45 years, formula (8) predicts that infection rates can range from 1.45 to 1.61 per 1,000. These results are in close agreement with OSHA's weighted average estimate of the national TB infection rate, which is 1.46 per 1,000.

4. Occupational Risk Estimations

OSHA used the three different data sources to obtain estimates of risk of TB infection for health care employees: the Washington State data, the North Carolina study, and the NIOSH Health Hazard Evaluation (HHE) from Jackson Memorial Hospital (Exs. 7-263, 7-7, 7-108). The Washington State data represent workplaces located in low TB prevalence areas, where TB infection control measures and engineering controls are required by state health regulations. The North Carolina data represent workplaces located in areas with moderate TB prevalence and inadequate TB infection control programs. Finally, the Jackson Memorial Hospital data are representative of county hospitals serving high-risk patients whose employees have a high frequency of exposure to infectious TB. These data sources provide information on the magnitude of the expected excess risk in three different environments, and are used to provide a range of possible values of excess risk.

Based on the Washington State data, the annual risk is expected to be 1.5 times the background rate for hospital employees, approximately 11 times the background rate for long-term care employees, 6 times the background rate for home health care workers, and double the background rate for home care employees. Based on the North Carolina data, the annual risk is expected to be approximately 5 times the background rate. Based on the Jackson Memorial Hospital data, the annual risk is expected to be approximately 9 times the background.

Estimates of expected excess risk of TB infection for workers with occupational exposure by state are calculated by applying the excess relative risk ratios, derived from the three occupational studies, to the overall background rate of infection for each state and are presented in table V-8(a) -- table V-8(c). A range of excess risk of TB infection due to occupational exposure is constructed by using the minimum and maximum estimates of excess risk among all states for each data source. These results are presented in table V-9 and table V-10 for workers in hospitals and for workers in other work settings, respectively.

Lifetime estimates of the excess risk of TB infection were estimated based on the annual excess risk by using the formula {1-(1-p) ⁴⁵}, where p is the annual excess risk. Lifetime excess estimates of TB infection are presented in table V-9 and table V-10. Lifetime risk estimates of developing active TB are calculated from lifetime risk estimates of TB infection assuming that, once infected, there is a 10 % likelihood of progressing to active TB; these estimates are presented in table V-11 and table V-12. Further, the risk of death caused by TB is calculated from the lifetime estimates of active TB using OSHA's estimate of the TB case fatality rate (also presented in table V-11 and table V-12). The methodology used to estimate a TB case fatality rate is presented below.

As outlined in the Health Effects section, several possible outcomes are possible following an infection. Approximately 90 % of all infections never progress to active disease. An estimated 10 % of infections is expected to progress to active disease; most of these cases are successfully treated. However, a percentage of active TB cases develop further complications. Approximately 7.8 % of active TB cases may take a more severe clinical course and lead to death. The TB case fatality rate was estimated using information on reported deaths caused by TB from table 8-5 of the Vital Statistics for the U.S. and cases of TB reported in CDC's TB Surveillance system for 1989 through 1991 (Exs. 7-270, 7-264). As shown in table V-13, the TB case death rate ranged from 69.94 to 89.18 per 1,000 with a 3-year average of 77.85 per 1,000 TB cases. The Agency used the 3-year average (77.85 per 1,000) for its estimate of deaths caused by TB. This estimate is in close agreement with published results from a retrospective cohort study conducted in Los Angeles County on TB cases in 1990 (Ex. 7-268). In this study, all confirmed TB cases reported in the county in 1990 were tracked and the number of deaths where TB was the direct or contributing cause was ascertained. "Contributing cause" was defined as a case of TB of such severity that it would have caused the death of the patient had the primary illness not caused death earlier. Of the 1,724 cases included in the study, TB was considered the cause of death or the contributing cause of death in 135 cases (78.31 per 1,000).

National estimates of annual and lifetime risk for TB infection, active disease and death caused by TB due to occupational exposure are computed as weighted averages of the state estimates and are presented in table V-14.

(a) Risk Estimates for Hospital Employees: Logistic regression analysis of the Washington state hospital data indicated an increase in annual risk (47 % above background) for employees with potential exposure to TB. For this particular analysis the control group was defined as those hospitals with no-known TB patients that are located in counties that did not report any active TB cases in 1994. However, an increased risk of 47 % above background in the annual infection rate is expected to produce a range of 4 to 72 TB infections per 1000 exposed workers in a working lifetime, which could result in as many as 7 cases of active TB and approximately 1 death per 1,000 exposed workers.

Based on the survey of hospitals in North Carolina's western region, the expected overall risk due to occupational exposure is estimated to be 4 times the background rate. This results in an expected range of lifetime risk between 34 and 472 infections per 1,000 employees at risk for TB infection. Lifetime estimates of active TB cases resulting from these infections are expected to range between 3 and 47, resulting in as many as 4 deaths per 1,000 exposed employees at risk of TB infection. As done previously, the North Carolina study results were adjusted to reflect 1994 TB disease trends.

Based on the data from Jackson Memorial Hospital, the overall risk due to occupational exposure is estimated to be 8 times the background rate. This results in an expected range of lifetime risk between 67 and 723 infections per 1,000 employees at risk. Lifetime estimates of the number of active TB case per 100 exposed workers are expected to range between 7 and 72, resulting in as many as 6 deaths per 1,000 exposed employees at risk for TB infection.

In summary, table V-9 and table V-14 show that the annual occupational risk of infection is expected to range:

(a) From .09 to 1.66 with a weighted average of 0.68 per 1,000 for workplaces located in relatively low TB prevalence areas, and where TB infection measures and engineering controls are required;

(b) From 0.77 to 14.1 with a weighted average of 5.7 per 1,000 for workplaces located in areas with moderate TB prevalence and inadequate TB control programs; and

(c) From 1.54 to 28 with a weighted average of 11.8 per 1,000 for workplaces located in high TB prevalence areas, serving high risk patients, with high frequency of exposure to infectious TB.

Similarly, the lifetime occupational risk is expected to range:

(a) From 4 to 72 with a weighted average of 30 per 1,000 for workplaces located in relatively low TB prevalence areas, and where TB infection measures and engineering controls are required;

(b) From 34 to 472 with a weighted average of 219 per 1,000 for workplaces located in areas with moderate TB prevalence and inadequate TB control programs; and

(c) From 67 to 723 with a weighted average of 386 per 1,000 for workplaces located in high TB prevalence areas, serving high risk patients, with high frequency of exposure to infectious TB.

Risk estimates derived from either study (Washington State or North Carolina) represent an overall rate of occupational risk, because both studies include PPD skin testing results from the entire hospital employee population, whereas the Jackson Memorial study addresses the occupational risk to workers where exposure to infectious TB is highly probable.

Although the exact compliance rate is not known, hospitals in Washington State have been required to implement the CDC TB guidelines with respect to engineering controls (requiring isolation rooms with negative pressure) and infection control measures (advocating early patient identification, employee training, respiratory protection, and PPD testing).

Neither the facilities in North Carolina nor Jackson Memorial had engineering controls fully implemented at the time these data were collected. Early identification of suspect TB patients has always been recommended in North Carolina. However, engineering controls in isolation rooms were either not present or did not function properly because of modifications in the physical structure of the building (i.e., isolation rooms had been subdivided using partitions, air ducts had been re-directed because of remodeling, etc.). Tuberculin skin testing was very inconsistent and sporadic. In addition, employee training and use of respiratory protection were not emphasized.

By 1991, Jackson Memorial had most of the engineering controls in place in the HIV ward (where the first outbreak took place) and in selected areas with high TB exposure, but not in the entire hospital. However, the staff training program was still being developed and respiratory protection was not always adequate. Although exposures had been greatly reduced, "high risk" procedures were still being performed in certain areas of the hospital without adequate engineering controls, such as the Special Immunology clinic where HIV-TB patients received pentamidine treatments. Like the hospitals in the North Carolina study, Jackson Memorial represents a working environment that serves a patient population known to have high TB prevalence. In addition, Jackson Memorial only tested employees with patient contact in areas where active TB had been detected.

(b) Risk Estimates for Workers in Other Work Settings: In long-term care facilities for the elderly there is also a significantly increased likelihood that employees will encounter individuals with infectious TB. Persons over the age of 65 constitute a large proportion of the TB cases in the United States. In 1987, CDC reported that persons aged 65 and over accounted for 27 % (6150) of the reported cases of active TB in the U.S., although they account for only 12 % of the U.S. population. Many of these individuals were infected in the past and advancing age and decreasing immunocompetence have caused them to develop active disease. In 1990 the CDC estimated that approximately 10 million people were infected with TB. As the U.S. population steadily ages, many of these latent infections may progress to active disease. Because elderly persons represent a large proportion of the nation's nursing home residents and because the elderly represent a large proportion of the active cases of TB, there is an increased likelihood that employees at long-term care facilities for the elderly will encounter individuals with infectious TB.

Similarly, there are other occupational settings that serve high-risk client populations and thus have an increased likelihood of encountering individuals with infectious TB. For example, hospices, emergency medical services, and home-health care services provide services to client populations similar to those in hospitals and thus are likely to experience similar risks.

OSHA used information from the 1994 Washington state PPD skin testing survey to estimate occupational risk for workers in long-term care, home health care, and home care. Annual estimates of excess risk for TB infection are presented in TABLE V-10 and lifetime estimates for TB infection, active TB, and death caused by occupational TB are presented in TABLE V-12.

Based on the Washington State data, the overall annual excess risk for TB infection is estimated to be 10-fold over background for workers in long-term care. This results in an expected range of lifetime risk of between 85 and 800 infections per 1,000 employees at risk for TB infection. Lifetime estimates of the number of active TB cases resulting from these infections range from 9 to 81 and are projected to cause as many as 6 deaths per 1,000 exposed employees at risk of TB infection. Similarly, the overall annual excess risk of TB infection for workers in home health care is estimated to be approximately 500 % above background. This results in an expected range of lifetime risk of between 41 and 536 infections per 1,000 employees at risk for TB infection. Lifetime estimates of the number of active TB cases range from 4 to 54 per 1,000, and are projected to cause as many as 4 deaths per 1,000 exposed employees at risk of TB infection. Similarly, the overall annual excess risk of TB infection for workers in home care is estimated to be approximately 100 % above background. This results in an expected range of lifetime risk of between 10 and 164 infections per 1,000 employees at risk for TB infection. Lifetime estimates of the number of active TB cases range from 1 to 16, and are expected to result in approximately 1 death per 1,000 exposed employees at risk of TB infection.

Clearly, employees in all three groups (long-term care for the elderly, home health care, and home care) have higher risks than hospital employees in Washington. This could be attributed, in part, to the lack of engineering controls in these work settings. That respirators may be used only intermittently may also play a role. Although workers in these three groups are encouraged by local health authorities to use respiratory protection while tending to a suspect TB patient, the actual rate of respirator usage is difficult to ascertain. A third factor that may contribute to higher risk in these work settings is delayed identification of suspect TB patients due to confounding symptoms presented by the individuals. For example, many long-term care residents exhibit symptoms of persistent coughing from decades of smoking. Consequently, an individual in long-term care with a persistent cough may be infectious for several days before he or she is identified as having suspected infectious TB.

Qualitative Assessment of Risk for Other Occupational Settings

The quantitative estimates of the risk of TB infection discussed above are based primarily upon data from hospitals and selected other health care settings. Data from hospitals and certain health care settings were selected because OSHA believes that these data represent the best information available to the Agency for purposes of quantifying the occupational risks of TB infection and disease. However, as discussed above, it is their exposure to aerosolized M. tuberculosis that places these workers at risk of infection and not factors unique to these particular kinds of health care activities. Thus, OSHA believes that the risk estimates derived from hospitals and selected other work settings can be used to describe the potential range of risks for other health care and other occupational settings in which workers can reasonably anticipate frequent and substantial exposure to aerosolized M. tuberculosis.

In order to extrapolate the quantitative risk estimates calculated for hospital employees and other selected health care settings, OSHA, as a first step, identified risk factors that place employees at risk of exposure. Some amount of exposure to TB could occur in any workplace in the United States. TB is an infectious disease that occurs in the community and thus, individuals may bring the disease into their own workplace or to other businesses or work settings that they may visit. However, there are particular kinds of work settings where risk factors are present that substantially increase the likelihood that employees will be frequently exposed to aerosolized M. tuberculosis. First among these factors is the increased likelihood of exposure to individuals with active, infectious TB. Individuals who are infected with TB have a higher risk of developing active TB if they are (1) immunocompromised (e.g., elderly, undergoing chemotherapy, HIV positive), (2) intravenous drug users, or (3) medically underserved and of generally poor health status (Exs. 6-93 and 7-50). Thus, in work settings in which the client population is composed of a high proportion of individuals who are infected with TB, are immunocompromised, are intravenous drug users or are of poor general health status, there is a greatly increased likelihood that employees will routinely encounter individuals with infectious TB and be exposed to aerosolized M. tuberculosis. A second factor that places employees at high risk of exposure to aerosolized M. tuberculosis is the performance of high-hazard procedures, i.e., procedures performed on individuals with suspected or confirmed infectious TB where there is a high likelihood of the generation of droplet nuclei. A third factor that places employees at risk of exposure is the environmental conditions at the work setting. Work settings that have overcrowded conditions or poor ventilation will facilitate the transmission of TB. Thus, given that a case of infectious TB does occur, the conditions at the work setting itself may promote the transmission of disease to employees who share airspace with the individual(s) with infectious TB.

The second step in extrapolating the quantitative risks is to identify the types of work settings which have some or all of the risk factors outlined above. Once these work settings have been identified, OSHA believes that it is reasonable to assume that the quantitative risk estimates calculated for hospitals and other selected health care settings can be used to describe the risks in the identified work settings.

Correctional Facilities

Employees in correctional facilities or other facilities that house inmates or detainees have an increased likelihood of frequent exposure to individuals with infectious TB. Many correctional facilities have a higher incidence of TB cases in comparison to the incidence in the general population. In 1985, the CDC estimated that the incidence of TB among inmates of correctional facilities was more than three times higher than that for nonincarcerated adults aged 15-64 (Ex. 3-33). In particular, in states such as New Jersey, New York, and California, the increased incidence of annual TB cases in correctional facilities ranged from 6 to 11 times greater than that of the general population for their respective states (Exs. 7-80 and 3-33). A major factor in the increased incidence of TB cases in correctional facilities is the fact that the population of correctional facilities is over-represented by individuals who are at greater risk of developing active disease, e.g., persons from poor and minority groups who may suffer from poor nutritional status and poor health care, intravenous drug users, and persons infected with HIV. Similarly, certain types of correctional facilities, such as holding facilities associated with the Immigration and Naturalization Service, may have inmates/detainees from countries with a high incidence of TB. For foreign-born persons arriving in the U.S., the case rate of TB in 1989 was estimated to be 124 per 100,000, compared to an overall TB case rate of 9.5 per 100,000 for the U.S. (Ex. 6-26). Moreover, in the period from 1986 to 1989, 22 % of all reported cases of TB disease occurred in the foreign-born population. Given the increased prevalence of individuals at risk for developing active TB, there is an increased likelihood that employees working in these facilities will encounter individuals with infectious TB. In addition, environmental factors such as overcrowding and poor ventilation facilitate the transmission of TB. Thus, given that a case of infectious TB does occur, the conditions in the facility itself promote the transmission of the disease to other inmates and employees in the facility who share airspace.

As discussed in the Health Effects section, a number of outbreak investigations (Exs. 6-5, 6-6) have shown that where there has been exposure to aerosolized M. tuberculosis in correctional facilities, the failure to promptly identify individuals with infectious TB and provide appropriate infection control measures has resulted in employees being infected with TB. These studies demonstrate that, as in hospitals or health care settings, where there is exposure to aerosolized TB bacilli and where effective control measures are not implemented, exposed employees are at risk of infection. Thus, estimates based on the risk observed among employees in hospitals and in selected other work settings that involve an increased likelihood of exposure can be appropriately applied to employees in correctional facilities.

Recently, scientists at NIOSH have completed a prospective study of the incidence of TB infection among New York State correctional facilities employees (Ex. 7-288). This study is the first prospective study of TB infection among employees in correctional facilities in an entire state. Other studies have reported on contact investigations, which seek to identify recent close contacts with an index case and determine who might subsequently have been infected. Studies based on contact investigations have the advantage of a good definition of potential for exposure and they serve to identify infected persons for public health purposes. On the other hand, prospective studies of an entire working group have the advantage of covering the entire population potentially at risk, of considering all inmate cases simultaneously as potential sources of infection, and, most importantly, of permitting the calculation of incidence rates and risk attributable to occupational exposure.

Following an outbreak of active TB among inmates that resulted in transmission to employees in 1991, the state of New York instituted a mandatory annual tuberculin skin testing program to detect TB infection among employees. The authors used data from the first two years of testing to estimate the incidence of TB infection among 24,487 employees of the NY Department of Corrections. Subjects included in the study had to have two sequential PPD skin tests, have a negative test the first year, and have complete demographic information. The overall conversion rate was estimated to be 1.9 %. Preliminary results show that after controlling for age, ethnicity, gender, and residence in New York City, corrections offices and medical personnel, working in prisons with inmate active TB cases, had odds ratios of TB infection of 1.64 and 2.39, respectively, compared to maintenance and clerical personnel who had little opportunity for prisoner contact. Based on these results, the annual excess risk due to occupational exposure is estimated to be 1.22 % and 2.64 % for corrections officers and medical personnel, respectively. This translates into lifetime occupational risks of 423 and 700 per 1,000 exposed employees, respectively. In prisons with no known inmate TB cases, there were no significant differences in TB infection rates among employees in different job categories.

Homeless Shelters

Employees in homeless shelters also have a significantly increased likelihood of frequent exposure. A high prevalence of TB infection and disease is common in many homeless shelters. Screening in selected shelters has shown the prevalence of TB infection to range from 18 to 51 % (Ex. 6-15). Many shelter residents also possess characteristics that impair their immunity and thus place them at greater risk of developing active disease. For example, homeless persons often suffer from poor nutrition and poor overall health status, and they also have poor access to health care. In addition, they may suffer from alcoholism, drug abuse and infection with HIV. Screening of selected shelters has shown the prevalence of active TB disease to range from 1.6 to 6.8 % (Ex. 6-15). Thus, there is an increased likelihood that employees at homeless shelters will frequently encounter individuals with infectious TB in the course of their work.

In addition, as in the case for correctional facilities, homeless shelters also tend to be overcrowded and have poor ventilation, factors that promote the transmission of disease and place shelter residents and employees at risk of infection. Outbreaks reported among homeless shelters (Exs. 7-51, 7-75, 7-73, 6-25) also provide evidence that where there is exposure to individuals with infectious TB and effective infection control measures are not implemented, employees are at risk of infection. It is reasonable to assume, therefore, that risk estimates calculated for hospital employees who have an increased likelihood of exposure to individuals with infectious TB can be used to estimate the risks for homeless shelter employees.

Facilities That Provide Treatment for Drug Abuse

Employees in facilities that provide treatment for drug abuse have an increased likelihood of frequent exposure to individuals with infectious TB. Surveys of selected U.S. cities by the CDC have shown the prevalence of TB infection among the clients of drug treatment centers to range from approximately 10 % to 13 % (Ex. 6-8). Clients of these centers are also generally at higher risk of developing active disease. The clients typically come from medically underserved populations and may suffer from poor overall health status. As discussed in the Health Effects section, drug dependence has also been shown to be a possible risk factor in the development of active TB. Moreover, many of the drug treatment center clients are intravenous drug users and are infected with HIV, placing these individuals at an increased risk of developing active TB. Given these risk factors for the clients served at drug treatment centers, there is an increased likelihood that employees in these work settings will be exposed frequently to individuals with infectious TB.

Medical Laboratories

Medical laboratory work is a recognized source of occupational hazards. CDC considers workers in medical laboratories that handle M. tuberculosis to be at high risk for occupational transmission of TB either because of the volume of material handled by routine diagnostic laboratories or the high concentrations of pathogenic agents often handled in research laboratories.

Few surveys of laboratory-acquired infections have been undertaken; most reports are of small outbreaks in specific laboratories. Sulkin and Pike's study of 5,000 laboratories suggested that brucellosis, tuberculosis, hepatitis, and enteric diseases are among the most common laboratory-acquired infections (Ex. 7-289). In 1957, Reid noted that British medical laboratory workers had a risk of acquiring tuberculosis two to nine times that of the general population (Ex. 7-289). This result was validated in 1971 by Harrington and Channon in their study of medical laboratories (Ex. 7-289). A retrospective postal survey of approximately 21,000 medical laboratory workers in England and Wales showed a five-times increased risk of developing active TB among these workers as compared with the general population. Technicians were at greater risk, especially if they worked in anatomy departments. A similar survey carried out in 1973 of 3,000 Scottish medical laboratory workers corroborates the results from England and Wales. Three cases, one doctor and two technicians, were noted in the 1973 survey, which resulted in an overall incidence rate of 109 per 100,000 person-years. The general population incidence rate for active TB was 26 per 100,000 person-years, giving a risk ratio of 4.2 (Ex. 7-289).

The studies reviewed in this section indicate that workers in medical laboratories with potential for exposure to M. tuberculosis during the course of their work have a several-fold (ranging from 2- to 9-fold) increased risk of developing active disease compared with the risk to the general population. Although these studies were conducted over two decades ago, they represent the most recent data available to the Agency, and OSHA has no reason to believe that the conditions giving rise to the risk of infection at that time have changed substantially in the interim. The Agency is not aware of any more current data on transmission rates in medical laboratories. OSHA solicits information on additional studies addressing occupational exposure to active TB in laboratories; such studies would then be considered by OSHA in the development of a final rule.

Other Work Settings and Activities

In addition to the information available for correctional facilities, homeless shelters, and facilities that provide treatment for drug abuse, there are other work settings and activities where there is an increased likelihood of frequent exposure to aerosolized M. tuberculosis. For example, hospices serve client populations similar to those of hospitals and perform similar services for these individuals. Individuals who receive care in hospices are likely to suffer from medical conditions (e.g., HIV disease, end-stage renal disease, certain cancers) that increase their likelihood of developing active TB disease once infected. Thus, employees providing hospice care have an increased likelihood of being exposed to aerosolized M. tuberculosis. CDC has recommended that hospices follow the same guidelines for controlling TB that hospitals follow.

Emergency medical service employees also have an increased likelihood of encountering individuals with infectious TB. Like hospices, emergency medical services cater to the same high risk client populations as hospitals. Moreover, emergency medical services are often used to transport individuals identified with suspected or confirmed infectious TB from various types of health care settings to facilities with isolation capabilities.

In addition, other types of services (e.g., social services, legal counsel, education) are provided to individuals who have been identified as having suspected or confirmed infectious TB and have been placed in isolation or confined to their homes. Employees who provide social welfare services, teaching, law enforcement or legal services to those individuals who are in AFB isolation are exposed to aerosolized M. tuberculosis. In particular, employees performing high-hazard procedures are likely to generate aerosolized M. tuberculosis by virtue of the procedure itself. Thus, employees providing these types of services also have an increased likelihood of exposure to aerosolized M. tuberculosis and are therefore likely to experience risks similar to those described above for hospital workers.

Although they do not have contact with individuals with infectious TB, employees who repair and maintain ventilation systems which carry air contaminated with M. tuberculosis and employees in laboratories who manipulate tissue samples or cultures contaminated with M. tuberculosis also have an increased likelihood of being exposed to aerosolized M. tuberculosis. Like employees in the work settings discussed above, these employees have an increased risk of frequent exposure to aerosolized M. tuberculosis.

Therefore, OSHA believes that the quantitative risk estimates derived from data observed among health care workers in the hospital setting can be generally used to describe the potential range of risks for workers in other occupational settings where there is a reasonable anticipation of exposure to aerosolized M. tuberculosis. The reasonableness of this assumption is supported by the overall weight of evidence of the available health data. As discussed in the Health Effects section, epidemiological studies, case reports and outbreak investigations have shown that in correctional facilities, homeless shelters, long-term care facilities for the elderly, drug treatment centers, and laboratories where appropriate TB infection control programs have not been implemented, employees have become infected with TB as a result of occupational exposure to individuals with infectious TB or to other sources of aerosolized M. tuberculosis. Thus, although the data on employee conversion rates in other work settings cannot be used to directly quantify the occupational risk of infection for those work settings, there is strong evidence that employees in various work settings other than hospitals can reasonably be anticipated to have exposure to aerosolized M. tuberculosis and that TB can be transmitted in these workplaces when appropriate TB infection control programs are not implemented.

VI. Significance of Risk

Section 6(b)(5) of the OSH Act vests authority in the Secretary of Labor to issue health standards. This section provides, in part, that:

The Secretary, in promulgating standards dealing with toxic materials or harmful physical agents under this subsection, shall set the standard which most adequately assures, to the extent feasible, on the basis of the best available evidence, that no employee will suffer impairment of health or functional capacity even if such employee has regular exposure to the hazard dealt with by such standard for the period of his working life.

OSHA's overall analytical approach to making a determination that workplace exposure to certain hazardous conditions presents a significant risk of material impairment of health is a four step process consistent with interpretations of the OSH Act and rational, objective policy formulation. In the first step, a quantitative risk assessment is performed where possible and considered with other relevant information to determine whether the substance to be regulated poses a significant risk to workers. In the second step, OSHA considers which, if any, of the regulatory alternatives being considered will substantially reduce the risk. In the third step, OSHA examines the body of "best available evidence" on the effects of the substance to be regulated to set the most protective requirements that are both technologically and economically feasible. In the fourth and final step, OSHA considers the most cost-effective way to achieve the objective.

In the Benzene decision, the Supreme Court indicated when a reasonable person might consider the risk significant and take steps to decrease it. The Court stated:

It is the Agency's responsibility to determine in the first instance what it considers to be "significant" risk. Some risks are plainly acceptable and others are plainly unacceptable. If, for example, the odds are one in a billion that a person will die from cancer by taking a drink of chlorinated water, the risk could not be considered significant. On the other hand, if the odds are one in a thousand that regular inhalation of gasoline vapors that are 2 % benzene will be fatal, a reasonable person might well consider the risk significant and take the appropriate steps to decrease or eliminate it. (I.U.D. v. A.P.I.), 448 U.S. at 655).

The Court indicated that "while the Agency must support its findings that a certain level of risk exists with substantial evidence, we recognize that its determination that a particular level of risk is `significant' will be based largely on policy considerations." The Court added that the significant risk determination required by the OSH Act is "not a mathematical straitjacket" and that "OSHA is not required to support its findings with anything approaching scientific certainty." The Court ruled that "a reviewing court (is) to give OSHA some leeway where its findings must be made on the frontiers of scientific knowledge and that the Agency is free to use conservative assumptions in interpreting the data with respect to carcinogens, risking error on the side of overprotection rather than underprotection." (448 U.S. at 655, 656).

As a part of the overall significant risk determination, OSHA considers a number of factors. These include the type of risk presented, the quality of the underlying data, the reasonableness of the risk assessments, and the statistical significance of the findings.

The hazards presented by the transmission of tuberculosis, such as infection, active disease, and death are very serious, as detailed above in the section on health effects. If untreated, 40-60 % of TB cases have been estimated to result in death (Exs. 5-80, 7-50, 7-66). Fortunately, TB is a treatable disease. The introduction of antibiotic drugs for TB has helped to reduce the mortality rate by 94 % since 1953 (Ex. 5-80). However, TB is still a fatal disease in some cases. From 1989-1991 CDC reported 5,452 deaths among adults from TB (see TABLE V-13, Risk Assessment section). In addition, there has been an increase in certain forms of drug-resistant TB, such as MDR-TB, in which the tuberculosis bacilli are resistant to one or more of the front line drugs such as isoniazid and rifampin, two of the most effective anti-TB drugs. The information available today is not adequate to estimate the future course of MDR-TB, but the reduction in the potential of transmitting this deadly form of the disease is itself another benefit of this standard. The current data indicate that among MDR-TB cases, the risk of death is increased compared to drug-susceptible forms of the disease. A CDC investigation of 8 outbreaks of MDR-TB revealed that among 253 people infected with MDR-TB, 75 % died within a period 4 to 16 weeks after the time of diagnosis (Ex. 38-A). MDR-TB may be treated, but due to the difficulty in finding adequate therapy which will control the bacilli's growth, individuals with this form of the disease may remain infectious for longer periods of time, requiring longer periods of hospitalization, additional lost worktime, and an increased likelihood of spreading TB infection to others until treatment renders the patient non-infectious. Because of the difficulty in controlling these drug-resistant forms of the disease with antibiotics, progressive lung destruction may progress to the point where it is necessary to remove portions of the lung to treat the advance of the disease.

The OSH Act directs the Agency to set standards that will adequately assure, to the extent feasible, that no employee will suffer "material impairment of health or functional capacity." TB infection represents a material impairment of health that may lead to active disease, tissue and organ damage, and death. Although infected individuals may not present any signs or symptoms of active disease, being infected with TB bacilli is a serious threat to the health status of the infected individual. Individuals who are infected have a 10 % chance of developing active disease at some point in their life, a risk they would not have had without being infected. The risk of developing active disease is even greater for individuals who are immunocompromised, due to any of a large number of factors. For example, individuals infected with HIV have been estimated as having an 8-10 % risk per year of developing active disease (Ex. 4B).

In addition, since infected individuals commonly undergo treatment with anti-TB drugs to prevent the onset of active disease, they face the additional risk of serious side effects associated with the highly toxic drugs used to treat TB. Preventive treatment with isoniazid, one of the drugs commonly used to treat TB infection, has been shown in some cases to result in death from hepatitis or has damaged the infected person's liver to the extent that liver transplantation was performed (Ex. 6-10). Thus, the health hazards associated with TB infection clearly constitute material impairment of health.

Clinical illness, i.e., active disease, also clearly constitutes material impairment of health. Left untreated, 40 to 60 percent of active cases may lead to death (Exs. 7-50, 7-66, 7-80). Individuals with active disease may be infectious for various periods of time and often must be hospitalized. Active disease is marked by a chronic and progressive destruction of the tissues and organs infected with the bacteria. Active TB disease is usually found in the lungs (i.e., pulmonary tuberculosis). Long-term damage can result even when cases of TB are cured; a common result of TB is reduced lung function (impaired breathing) due to lung damage (Ex. 7-50, pp. 30-31). Inflammatory responses caused by the disease produce weakness, fever, chest pain, cough, and, when blood vessels are eroded, bloody sputum. Also, many individuals have drenching night sweats over the upper part of the body several times a week. The intensity of the disease varies, ranging from minimal symptoms of disease to massive involvement of many tissues, with extensive cavitation and debilitating constitutional and respiratory problems. Long-term damage can also result from extrapulmonary forms of active disease; such damage may include mental impairment from meningitis (infection of membranes surrounding the brain and spinal cord) and spinal deformity and leg weakness due to infection of the vertebrae (i.e., skeletal TB) (Ex. 7-50, p. 31). Active disease is treatable but it must be treated with potent drugs that have to be taken for long periods of time. The drugs currently used to treat active TB disease may be toxic to other parts of the body. Commonly reported side effects of anti-TB drugs include hepatitis, peripheral neuropathy, optic neuritis, ototoxicity and renal toxicity (Ex. 7-93). Active disease resulting from infection with MDR-TB is of even greater concern due to the inability to find adequate drug regimens. Although OSHA has not been able to precisely quantify the increase in incidence of MDR-TB, the number of cases of MDR-TB is clearly on the rise. In these cases, individuals may remain infectious for longer periods of time and may suffer more long-term damage from the chronic progression of the disease until adequate therapy can be identified.

In this standard, OSHA has presented quantitative estimates of the lifetime risk of TB infection, active disease and death from occupational exposure to M. tuberculosis. Qualitative evidence of occupational transmission is also included in OSHA's risk assessment.

In preparing its quantitative risk assessment, OSHA began by seeking out occupational data associated with TB infection incidence in order to calculate an estimate of risk for TB infection attributable to occupational exposure for all U.S. workers. Unfortunately, an overall national estimate of risk for TB infection attributable to occupational exposure is not available. CDC, which collects and publishes the number of active TB cases reported nationwide each year, does not publish occupational data associated with the incidence of TB infection and active TB on a nationwide basis. There has been some effort to include occupational information on the TB reporting forms, but only a limited number of states are currently using the new forms and capturing occupational information in a systematic way. In the absence of a national database, OSHA used two statewide studies, from North Carolina and Washington (Exs. 7-7, 7-263), and data from an individual hospital, Jackson Memorial Hospital (Ex. 7-108), on conversion rates of TB infection for workers in hospitals. The Washington State database also contained information on three additional occupational groups: long-term care, home health care and home care employees. OSHA used these data to model average TB infection rates and estimate the range of expected risks in the U.S. among workers with occupational exposure to TB.

The conversion rates in the selected studies were used to estimate the annual excess relative risk due to occupational exposure, which was expressed as the percent increase of infection above each study's control group. In order to estimate an overall range of occupational risk of TB infection, taking into account regional differences in TB prevalence in the U.S. and indirectly adjusting for factors such as socio-economic status, which might influence the rate of TB observed in different parts of the country, OSHA: (1) Estimated background rates of infection for each state by assuming that the number of new infections is functionally related to the number of active cases reported by the state each year (i.e., the distribution of new infections is directly proportional to the distribution of active cases), and 2) applied estimates of the annual excess relative risk, derived from the occupational studies, to the state background rates to calculate estimates of excess risk due to occupational exposure by state. Thus, the excess occupational risk estimates are actually calculated from the three available studies, on a relative increase basis, and these relative increases are multiplied by background rates for each state to derive estimates of excess occupational risk by state. The state estimates are then used to derive a national estimate of annual occupational risk of TB infection. Given an annual rate of infection, the lifetime risk of infection was calculated assuming that workers are exposed for 45 years and that the worker's exposure profile and working conditions remain constant throughout his or her working lifetime. Lifetime infection rates are then used to calculate the lifetime risk of developing active disease based on the estimate that 10 % of all infections result in active disease. Given a number of active cases of TB, the number of expected deaths can be calculated based on the estimated average TB case death rate (i.e., number of TB deaths per number of active TB cases averaged over 3 years as reported by CDC).

OSHA estimates that the risk of material impairment of health or functional capacity, that is, the average lifetime occupational risk of TB infection for hospital workers ranges from 30 to 386 infections per 1,000 workers who are occupationally exposed to TB. These are different national averages, each derived by calculating the risk in each state and weighting it by the state's population. The low end of this range is derived by using the Washington State data, and is likely to seriously underestimate the true risk to which workers are exposed. This is because the Washington data represent occupational exposures among employees in hospitals which are located in areas of the country with a low prevalence of active TB and which have implemented TB controls (e.g., early identification procedures, annual skin testing, and negative pressure in AFB isolation rooms). The high end of this range is derived by using the Jackson Memorial Hospital study, and represents occupational risk for workers in hospitals located in high TB prevalence areas, serving high risk patients, and with a high frequency of exposure to infectious TB.

OSHA also used information from the Washington State database to estimate national average estimates of lifetime risk for workers in long-term care (i.e., nursing homes), home health care, and home care. The national average lifetime risk of TB infection is estimated to be 448 per 1,000 for workers in long-term care facilities, 225 per 1,000 for workers in home health care (primarily nursing staff), and 69 per 1,000 for workers in home care. The higher likelihood of occupational exposure in long-term care facilities (early identification of suspect TB cases is often difficult among the elderly) and the presence of fewer engineering controls in these facilities may explain the high observed occupational risk in that work setting.

The national average lifetime risk of developing active disease ranges from approximately 3 to 39 cases per 1,000 exposed employees for workers in hospital settings. Similarly, the average lifetime risk of active disease is estimated to be approximately 45 per 1,000 for workers in long-term care, 26 per 1,000 in home health care, and 7 per 1,000 in home care. This range is based on the estimate that 10 % of infections will progress to active disease over one's lifetime. This risk may be greater for immunocompromised individuals.

The national average lifetime risk of death from TB ranges from 0.2 to approximately 3 deaths per 1,000 exposed employees for workers in hospital settings. Similarly, the average lifetime risk of death from TB is estimated to be approximately 3.5 per 1,000 for workers in long-term care, 2 per 1,000 for workers in home health care, and 0.5 per 1,000 in home care. The lower range of the national lifetime risk of deaths, 0.2 per 1,000, is based on the Washington State hospital data where the prevalence of TB is low and infection control measures have been implemented. Thus, this lower range of risk underestimates the risk of death from TB for other employees who work in settings where infection control measures, such as those outlined in this proposed standard, have not been implemented. The risk assessment data show that where infection control measures were not in place, the estimated risk of death from TB was as high as 6 deaths per 1,000 exposed employees.

The quantitative risk estimates are based primarily upon data from hospitals and selected other work settings. However, it is frequent exposure to aerosolized M. tuberculosis which places workers at substantially increased risk of infection and not factors unique to the health care profession or any job category therein. Qualitative evidence, such as that from the epidemiological studies, case reports and outbreak investigations reported for various types of work settings, as discussed earlier in the Health Effects section, clearly demonstrates that employees exposed to aerosolized M. tuberculosis have become infected with TB and have gone on to develop active disease. These work settings share risk factors that place employees at risk of transmission. For example, these work settings serve client populations that are composed of a high prevalence of individuals who are infected with TB, are immunocompromised, are injecting drug users or are medically underserved and of poor general health status. Therefore, there is an increased likelihood that employees in these work settings will encounter individuals with active TB. In addition, high-hazard procedures, such as bronchoscopies, are performed in some of these work settings, which greatly increases the likelihood of generating aerosolized M. tuberculosis. Moreover, some of the work settings have environmental conditions such as overcrowding and poor ventilation, factors that facilitate the transmission of disease. Therefore, OSHA believes that the quantitative risk estimates based on hospital data and other selected health care settings can be extrapolated to other occupational settings where there is a similar increased likelihood of exposure to aerosolized M. tuberculosis.

Having specific data for non-health care workers and workplace conditions would add more precision to the quantitative risk assessment, but that level of detail is not possible with the currently available information. However, the Agency believes that such a level of detail is not necessary to make its findings of significant risk because the risk of infection is based upon occupational exposure to aerosolized M. tuberculosis. Nevertheless, OSHA seeks information on conversion rates and the incidence of active disease among employees in non-health care work settings in order to give more precision to its estimates of risk.

OSHA's risk estimates for TB infection are comparable to other risks which OSHA has concluded are significant, and are substantially higher than the example presented by the Supreme Court in the Benzene Decision. After considering the magnitude of the risk as shown by the quantitative and qualitative data, OSHA preliminarily concludes that the risk of material impairment of health from TB infection is significant.

OSHA also preliminarily concludes that the proposed standard for occupational exposure to TB will result in a substantial reduction in that significant risk. The risk of infection is most efficiently reduced by implementing TB exposure control programs for the early identification and isolation of individuals with suspected or confirmed infectious TB. Engineering controls to maintain negative pressure in isolation rooms or areas where infectious individuals are being isolated will reduce the airborne spread of aerosolized M. tuberculosis and subsequent exposure of individuals, substantially reducing the risk of infection. In addition, for those employees who must enter isolation rooms or provide services to individuals with infectious TB, respiratory protection will reduce exposure to aerosolized M. tuberculosis and thus reduce the risk of infection.

Several studies have shown that the implementation of infection control measures such as those outlined in this proposed standard have resulted in a reduction in the number of skin test conversions among employees with occupational exposure to TB. For example, results of a survey conducted by the Society of Healthcare Epidemiology of America (SHEA) of its member hospitals (Exs. 7-147 and 7-148) revealed that among hospitals that treated 6 or more patients with infectious TB per year there were 68 % fewer tuberculin skin test conversions in hospitals that had AFB isolation rooms with one patient per room, negative pressure, exhaust air directed outside and six or more air changes per hour, compared to hospitals that did not have AFB isolation rooms with these same characteristics. Similarly, an 88 % reduction in tuberculin skin test conversions was observed in an Atlanta hospital after the implementation of infection control measures such as an expanded respiratory isolation policy, improved diagnostic and testing procedures, the hiring of an infection control coordinator, expanded education of health care workers, increased frequency of tuberculin skin tests, implementation of negative pressure, and use of submicron masks for health care workers entering isolation rooms (Ex. 7-173). Improvements in infection control measures in a Florida hospital after an outbreak of MDR-TB reduced tuberculin skin test conversions from 28 % to 18 % to 0 % over three years (Ex. 7-167). These improvements included improved early identification procedures, restriction of high-hazard procedures to AFB isolation rooms, increased skin testing, expansion of initial TB treatment regimens, and daily inspection of negative pressure in AFB isolation rooms. Thus, these investigations show that the implementation of infection control measures such as those included under OSHA's proposed standard for TB can result in substantial reductions in infections among exposed employees.

As discussed in further detail in the following section of the Preamble to this proposed standard, OSHA estimates that full implementation of the proposed standard for TB will result in avoiding approximately 21,400 to 25,800 work-related infections per year, 1,500 to 1,700 active cases of TB resulting from these infections and 115 to 136 deaths resulting from these active cases. In addition, because the proposed standard encourages the identification and isolation of active TB cases in the client populations served by workers in the affected industries, there will also be non-occupational TB infections that will be averted. OSHA estimates that implementation of the proposed standard will result in avoiding approximately 3,000 to 7,000 non-occupational TB infections, 300 to 700 active cases of TB resulting from these infections, and 23 to 54 deaths resulting from these active cases. OSHA preliminarily concludes that the proposed standard for TB will significantly reduce the risk of infection, active disease and death from exposure to TB and that the Agency is thus carrying out the Congressional intent and is not attempting to reduce insignificant risks.

Although the current OSHA enforcement program, which is based on the General Duty Clause of the Act, Section 5(a)(1), and the application of some general industry standards, such as 29 CFR 1910.134, Respiratory Protection, has reduced the risks of occupational exposure to tuberculosis to some extent, significant risks remain and it is the Agency's opinion that an occupational health standard promulgated under section 6(b) of the Act will much more effectively reduce these risks for the following reasons. First, because of the standard's specificity, employers and employees are given more guidance in reducing exposure to tuberculosis. Second, it is well known that a standard is more protective of employee health than an enforcement program based upon the general duty clause and general standards. Unlike the proposed standard, the general duty clause specifies no abatement methods and the general industry standards do not set forth abatement methods specifically addressing occupational exposure to TB. Third, the general duty clause imposes heavy litigation burdens on OSHA because the Agency must prove that a hazard exists at a particular workplace and that it is recognized by the industry or the cited employer. Since the proposed standard specifies both the conditions that trigger the application of the standard and the employer's abatement obligations, thereby establishing the existence of the hazard, no independent proof that the hazard exists in the particular workplace need be presented. The reduction in litigation burdens will mean that the Labor Department, as well as the employer, will save time and money in the investigation and litigation of occupational TB cases. Finally, the promulgation of this proposed standard will result in increased protection for employees in state-plan states which, although not required to adopt general duty clauses, must adopt standards at least as effective as Federal OSHA standard.

In summary, the institution of the enforcement guidelines has been fruitful, but it has not eliminated significant risks among occupationally exposed employees. Therefore, OSHA preliminarily concludes that a standard specifically addressing the risks of tuberculosis is necessary to further substantially reduce significant risk. OSHA's preliminary economic analysis and regulatory flexibility analysis indicate that the proposed standard is both technologically and economically feasible. OSHA's analysis of the technological and economic feasibility is discussed in the following section of the preamble.

VII. Summary of the Preliminary Economic Analysis and Regulatory Flexibility Analysis

OSHA is required by the Occupational Safety and Health Act of 1970 and several court cases pertaining to that Act to ensure that its rules are technologically and economically feasible for firms in the affected industries. Executive Order (EO) 12866 and the Regulatory Flexibility Act (as amended) also require Federal agencies to estimate the costs, assess the benefits, and analyze the impacts on the regulated community of the regulations they propose. The EO additionally requires agencies to explain the need for the rule and examine regulatory and non-regulatory alternatives that might achieve the objectives of the rule. The Regulatory Flexibility Act requires agencies to determine whether the proposed rule will have a significant economic impact on a substantial number of small entities, including small businesses and small government entities and jurisdictions. For proposed rules with such impacts, the agency must prepare an Initial Regulatory Flexibility Analysis that identifies those impacts and evaluates alternatives that will minimize such impacts on small entities. OSHA finds that the proposed rule is "significant" under Executive Order 12866 and "major" under Section 804(2) of the Small Business Regulatory Enforcement Fairness Act of 1996. Accordingly, the Occupational Safety and Health Administration (OSHA) has prepared this Preliminary Economic and Regulatory Flexibility Analysis (PERFA) to support the Agency's proposed standard for occupational exposure to tuberculosis (TB). The following is an executive summary of that analysis. The entire test of the PERFA can be found in the rulemaking docket as Exhibit 13. The complete PERFA is composed of various chapters that describe in detail the information summarized in the following section.

Statement of Need

TB is a communicable, potentially lethal disease caused by the inhalation of droplet nuclei containing the bacillus Mycobacterium tuberculosis ( M. tuberculosis). Persons exposed to these bacteria can respond in different ways: by overcoming the challenge without developing TB, by becoming infected with TB, or by developing active TB disease. Those who become infected harbor the infection for life, and have a 10 percent chance of having their infection progress to active disease at some point in their life. Those with active disease have the signs and symptoms of TB (e.g., prolonged, productive cough; fatigue; night sweats; weight loss) and have about an 8 percent risk of dying from their disease.

TB has been a worldwide health problem for centuries, causing millions of deaths worldwide. In the United States, however, there has been a decline in the number of active TB cases over the last four decades. Between 1953 and 1994, the number of active cases declined from 83,304 to 24,361, an annual rate of decline of 3.6 percent over the period as a whole (Figure VII-1). The 1988-1992 period, however, saw the first substantial increase in the number of active cases since 1953. A number of outbreaks of this disease have occurred among workers in health care settings, as well as other work settings, in recent years. To add to the seriousness of the problem, some of these outbreaks have involved the transmission of multi-drug resistant strains of M. tuberculosis, which are often fatal. Very recently, i.e., after 1992, this trend has reversed, and the number of such cases appears once again to have begun to decline. Nevertheless, TB remains a major health problem, with 22,813 active cases reported in 1995. Because active TB is endemic in many U.S. populations -- including groups in both urban and rural areas -- workers who come into contact with diseased individuals are at risk of contracting the disease themselves.

Many occupational groups, including workers in health care, nursing homes, homeless shelters, hospices, correctional facilities, laboratories, physicians' offices, and other settings are at risk of contracting TB on the job. These workers are at risk because they are exposed in the course of their work to patients and others with active TB disease, perform procedures that expose them to airborne concentrations of M. tuberculosis, or serve client populations where the incidence of active disease is unusually high.

The purpose of OSHA's standard is to reduce these risks in health care and other work settings where active TB cases are likely to be encountered by employees. To accomplish this goal, the proposed standard requires those employers who are responsible for the working conditions where such encounters occur to implement a program of infection prevention and infection control that is designed to prevent occupational infections in the first place, and to identify and treat any job-related infections that do occur. The approach taken in the proposed standard is similar to that adopted by OSHA in its 1991 bloodborne pathogens standard, which is given credit for achieving a dramatic reduction in the number of cases of hepatitis among health care and other workers since it was issued. OSHA predicts that, once implemented, the proposed TB standard will have similar results, achieving reductions on the order of 70 to 90 percent in the number of TB infections, active cases, and directly related deaths.

This Preliminary Economic and Regulatory Flexibility Analysis includes an introductory chapter that describes the major provisions of the standard. The proposal would apply to occupational exposure to TB occurring in, during, or through the provision of services by:

• Hospitals.
• Nursing homes.
• Correctional facilities.
• Immigration detainment facilities.
• Law enforcement facilities.
• Hospices.
• Substance abuse treatment centers.
• Homeless shelters.
• Medical examiners' offices.
• Home health care providers.
• Emergency medical services.
• Research and clinical laboratories handling TB.
• Contract work on ventilation systems or areas of buildings that may contain aerosolized M. tuberculosis.
• Physicians performing certain high hazard procedures.
• Social service workers providing services to individuals identified as having suspected or confirmed infectious TB.
• Personnel service agencies when providing workers to covered facilities.
• Attorneys visiting known or suspected infectious TB patients.

The groups, industries, and work settings covered by the standard have been included in its scope for specific reasons. For example, hospitals are included because they treat patients with active TB disease, while hospices, certain laboratories, pulmonary and certain other physicians, medical examiners, and contract HVAC workers are covered because employees in these settings/jobs are exposed to aerosolized M. tuberculosis during the performance of high-hazard procedures, such as bronchoscopies, sputum induction, autopsies, and during work on ventilation systems that may contain TB bacteria. Other work settings, such as homeless shelters and nursing homes, are covered because their employees serve a client population known to have a high incidence of TB infection. Another group of employees included within the scope of the standard are workers who must occasionally serve patients with active TB who are being treated in "isolation," i.e., a room or area specifically designed to contain the TB microorganism and prevent its spread to surrounding areas. Attorneys and social workers are typical of this group. Finally, the proposed standard covers personnel service agencies that provide temporary, seasonal, or "leased" personnel to hospitals and other covered work settings.

OSHA estimates that the standard would apply to approximately 102,000 establishments and provide protection to more than 5 million workers currently at risk of occupational exposure to TB. More than half of these workers -- almost 4 million -- work in the two industries most affected by the standard: hospitals and nursing homes. Other covered industries with large numbers of workers are home health care, emergency medical services, and correctional institutions.

Table VII-1 shows the number of affected establishments and the population at risk for each covered industry. (Table VII-1 does not include all sectors that might hypothetically be covered by the standard. For example, a chiropractor who engaged in high hazard procedures would be covered by the standard. However, this possibility is sufficiently rare for this activity not to have been included in this analysis. OSHA solicits comments on any affected job categories or industries it may have omitted.) Because the standard requires employers in the covered industries to make an initial determination that will identify which job classifications, employees, and activities within their workplace involve occupational exposure to TB, its requirements are narrowly targeted to those workers most at risk. Thus, for example, only approximately 57 percent of hospital workers are potentially affected by the standard; these workers would include those working on infectious disease floors or wards, radiology units, autopsy suites, and in other, similarly exposed locations.

Technological Feasibility

Chapter III of the analysis evaluates the technological feasibility of the proposed standard for affected establishments. OSHA preliminarily concludes that no provisions of the rule pose technological feasibility problems for any potentially affected entities. This is the case because the standard emphasizes administrative controls, such as the early identification of suspected or confirmed cases of TB and employee information and training, rather than engineering controls. In addition, the engineering controls that are required, such as AFB isolation rooms, biological safety cabinets, and temporary AFB isolation facilities, would be mandated only in those situations where individuals with suspected or confirmed infectious TB are admitted and isolated, where high hazard procedures are performed, and in situations where individuals cannot be placed into AFB isolation rooms within five hours of being identified as having suspected or confirmed infectious TB. All of the engineering controls required by the standard are currently available and in widespread use in many affected establishments.

Benefits of the Proposed Standard

Workers employed in the work settings covered by the standard are at significant risk of material impairment of health as a result of exposure to M. tuberculosis on the job. These workers will be the primary beneficiaries of the protection provided by the rule. However, because TB is a communicable disease, many other individuals will also benefit from the standard. Reducing the number of cases of TB among workers who are regularly in contact both with patients and infected members of client populations will reduce the incidence of TB infections and active cases in these client populations (since infected individuals spend the most time with other members of their group) and among members of the families of exposed workers. OSHA has expressed the benefits of the standard in terms of the numbers of TB infections, active cases, and TB-related deaths averted by the standard. In addition to reducing morbidity and mortality among workers, their families, and client populations, the standard will also generate readily quantifiable cost savings in the form of lower medical costs, less lost production, and reduced costs for administering workers' compensation claims and other private and social insurance system transactions.

OSHA's estimates of the potential benefits of the standard take into account the extent of current industry compliance with the provisions of the proposed standard, i.e., the benefits estimates do not include the benefits that employers in affected sectors are already garnering as a result of their voluntary efforts to provide protections to their TB-exposed employees. The benefits assessment presented in Chapter IV of the economic analysis is based on OSHA's Preliminary Risk Assessment (see that section of the preamble), which quantifies the occupational risk of TB infection among workers in hospitals, nursing homes, home health care work settings, and home care work settings. The estimates of risk are based on the rate of tuberculin skin test (TST) conversions among these populations. TST conversions are a widely used and well-documented index of TB infection; rates of conversion among the exposed populations are then compared with rates in unexposed or less-exposed "control" populations to obtain an estimate of the "excess" risk associated with occupational exposure. Table VII-2 shows the results of OSHA's estimates of the risks confronting workers in various work settings, based on statistical analyses and studies in the literature.

Where risk data of good quality were available for a specific industry, OSHA relied on that data. However, such data were available only for the hospital, nursing home, home health care, and home care industries. Accordingly, OSHA identified the best data to use to characterize the occupational risk of TB infection posed to workers in the other work settings covered by the proposed rule. After a careful review of the available data, OSHA chose to rely on data from western North Carolina that looked at occupational risk in a total of eight hospitals. These data were selected because they derived from hospitals that were relatively "uncontrolled," i.e., that had not yet implemented many of the controls that would be required by the proposed standard. Data from the other hospitals shown in Table VII-2 were judged to be less appropriate for the purpose of extrapolation because Washington State hospitals are already generally in compliance with the requirements of the proposed rule and Jackson Memorial Hospital had recently experienced an outbreak of multi-drug resistant TB among its patients at the time the risk data were gathered. OSHA believes that using occupational risk data from hospitals to characterize the risk in other occupational settings for which risk data are unavailable is appropriate because employees in these other settings serve client populations that have a high incidence of active TB cases, perform high-hazard procedures, or visit hospitalized TB patients. The use of a hospital-based risk estimate results in a lower estimate of risk than would be the case if OSHA had used risk data from nursing homes or home health care to characterize the risk in other settings, but a higher risk than if OSHA had used risk data from the home care industry to do so.

To predict the effectiveness of the proposed standard, OSHA evaluated the reduction in occupational risk that various control measures required by the standard can be expected to achieve. Effectiveness is measured as the percent reduction in TST conversions and in the TB infections, active cases, and deaths represented by those conversions. Based on a thorough review of the available literature on the effectiveness of control programs that have actually been implemented in a number of hospitals, OSHA believes that the proposed standard, once implemented, would reduce TB infections among occupationally exposed hospital workers by 90 percent, and would decrease such infections in the other work settings covered by the standard by 70 to 90 percent. OSHA also estimated the effectiveness and medical surveillance and follow-up in preventing infections from advancing to active cases of TB. OSHA found that such measures reduced the probability of an infection advancing to an active case by 35 to 47 percent, depending on the frequency of testing.

Using these effectiveness data, taking account of the current levels of compliance in various workplaces, and relying on the estimates of excess risk presented in OSHA's Preliminary Risk Assessment, OSHA predicts that the proposed standard will avert about 21,000 to 26,000 work-related TB infections per year, 1,500 to 1,750 active disease cases resulting directly from these infections, and 115 to 136 deaths directly related to the same infections. Preventing this number of infections among workers will, in turn, prevent about 3,000 to 7,000 infections, 300 to 700 active cases, and 23 to 54 deaths among the families, friends, clients, and contacts of these workers. In addition, the standard will annually generate cost savings of $89 to $116 million dollars in avoided medical costs, lost production caused by absence from work and other factors, and insurance administration costs. Table VII-3 shows the benefits of the proposed standard.

Chapter V of the economic analysis projects the costs employers in the various industries covered by the standard are estimated to incur to achieve compliance with the rule's requirements. OSHA estimated costs for each covered industry and for each provision of the standard. These costs take account of the baseline levels of compliance prevailing in each industry at the present time and are presented as annualized costs discounted at 7 percent. Annualized costs are the sum of annualized initial costs and recurring annual costs. For example, a temporary AFB isolation room costing $4,095 with annual maintenance costs of $50 would have annualized costs of $633 ($583 + $50).

The total estimated costs of compliance for the standard as a whole are $245 million per year. The most costly provisions of the standard are those requiring medical surveillance and training for occupationally exposed employees. Together, these two provisions account for 60 percent of the costs of compliance. The two industries projected to incur the highest costs are hospitals and nursing homes. Together, the costs incurred by these two industries are estimated to be $138 million per year. Tables VII-4 and VII-5 summarize the annualized costs of compliance, by provision and industry, respectively.

Chapter VI assesses the economic impacts of the proposed standard on the industries affected by the proposed standard and also analyzes the impacts on the small businesses within each of these industries. OSHA preliminarily concludes that the standard is economically feasible for affected firms. On average, annualized compliance costs for all entities amount only to 0.06 percent of revenues and only 1.8 percent of profits. For all industries, costs as a percentage of revenues are less than 1 percent. For two industries, costs as a percentage of profits exceed 5 percent; these industries are substance abuse treatment centers and personnel services. OSHA does not believe, however, that these profit impacts will actually be incurred by facilities in these two sectors. Only 18.5 percent of substance abuse treatment centers operate on a for-profit basis. If substance abuse treatment centers can increase their revenues by as little as 0.34 percent, they can completely offset their compliance costs. The revenue increases or reductions in services needed to achieve cost passthrough are not expected to represent significant impacts for these facilities. The situation for personnel service firms is similar; these firms would have to increase the prices charged to their customers by as little as 0.56 percent to completely offset the costs of compliance. It is likely that these agencies will be able to pass such a small increase in costs through to their customers, i.e., to facilities purchasing personnel services. Table VII-6 shows compliance costs as a percentage of revenues, by industry.

OSHA has preliminarily concluded that the proposed standard will have a significant impact on a substantial number of small entities and has therefore, as required by the Regulatory Flexibility Act Amendments of 1996, conducted an Initial Regulatory Flexibility Analysis (IRFA). This analysis has identified significant impacts on the small entity portion of the hospital, nursing home, correctional institution, homeless shelter, substance abuse treatment center, contract HVAC, and personnel services industries.

For the purposes of this analysis, OSHA defines small for-profit entities using the Small Business Administration's (SBA's) Table of Size Standards. For businesses affected by the proposed standard, the SBA classifies entities with annual revenues of less than $5 million as small for all industries, with the exception of contract HVAC firms, for which entities with less than $7 million in annual revenues are classified as small.

A small not-for-profit entity is defined as any nonprofit enterprise that is independently owned and operated and is not dominant in its field. Based on this definition, all not-for-profit entities affected by the proposed standard are considered small.

Many of the affected industries consist almost entirely of public sector facilities, such as correctional facilities, immigration detainment facilities, law enforcement facilities, medical examiners' offices, and social service organizations. Several other affected industries include some government-owned facilities, such as hospitals, nursing homes, and emergency medical services. Under the Regulatory Flexibility Act, "small governmental jurisdiction" refers to governments of cities, counties, towns, townships, villages, school districts, or special districts with populations of less than 50,000. For most of the affected industries, information on the number of such entities was not readily available. Where data were unavailable, the number of small publicly-owned entities was estimated based on the average number of people served per employee in each industry, from which OSHA estimated the average employment size of establishments serving populations of less than 50,000. These entities are considered small for the purposes of this analysis. OSHA requests information on size standards for public-sector entities.

OSHA requests comment on these definitions and estimates of the number of small entities. The complete IRFA is presented in Chapter VI of the economic analysis, and is also presented here.

Initial Regulatory Flexibility Analysis

The Regulatory Flexibility Act, as amended in 1996, requires that an Initial Regulatory Flexibility Analysis contain the following elements:

(1) A description of the reasons why action by the agency is being considered;

(2) A succinct statement of the objectives of, and legal basis for, the proposed rule;

(3) A description of, and, where feasible, an estimate of the number of small entities to which the proposed rule will apply;

(4) A description of the projected reporting, recordkeeping and other compliance requirements of the proposed rule, including an estimate of the classes of small entities that will be subject to the requirement and the type of professional skills necessary for preparation of the report or record; and

(5) An identification, to the extent practicable, of all relevant Federal rules that may duplicate, overlap or conflict with the proposed rule.

In addition, a regulatory flexibility analysis must contain a description of any significant alternatives to the proposed rule that accomplish the stated objectives of applicable statutes (in this case the OSH Act) and that minimize any significant economic impact of the proposed rule on small entities.⁽³⁾ This section of the analysis closes with a review of the recommendations of the SBREFA Panel concerning this proposed rule and discusses how OSHA has responded to these recommendations.

Reasons for the Proposed Rule

From 1985 to 1994, the number of active TB cases in the United States increased by 9.4 percent, reversing a 30-year downward trend. Although the number of cases reported to the CDC has declined over the past few years, TB remains a serious problem in the United States. In 1994, 24,361 active TB cases were reported to the Centers for Disease Control and Prevention (CDC), and TB was reported to have caused 1,590 deaths in that year alone (Ex. 7-283).

Transmission of M. tuberculosis is a recognized risk in several work settings. A number of outbreaks of this dreaded disease have occurred among workers in health care settings, as well as other work settings, in recent years. To add to the seriousness of the problem, some of these outbreaks have involved the transmission of multidrug-resistant strains of M. tuberculosis, a form of the disease that is often fatal.

Objectives of the Proposed Rule

The objective of this proposal is to reduce the risk of occupational exposure to M. tuberculosis in exposed working populations through the use of engineering controls, work practice controls, respiratory protection, medical surveillance, training, signs and labels, and recordkeeping. Implementation of these measures has been shown to minimize or eliminate occupational exposure to M. tuberculosis, and thus to reduce the risk of TB infection among workers. The legal authority for this proposed standard is the Occupational Safety and Health Act, 29 U.S.C. 655(b).

Description of the Number of Small Entities

The proposed rule would cover 80,400 establishments operated by 67,116 small entities, as defined above. Of the 67,116 small entities, about 49 percent (32,605 entities) are for-profit small entities, 20 percent (13,622 entities) are publicly-owned, and 31 percent (20,889 entities) are not-for-profit. About 79 percent of the total number of affected establishments are operated by small entities. The proposed rule covers 48,804 establishments operated by 48,044 very small entities, defined as entities of all kinds employing fewer than 20 workers. Almost 48 percent of the affected establishments are operated by very small entities.

Description of Proposed Reporting, Recordkeeping and Other Compliance Requirements

Avoiding a One-Size-Fits-All Standard. Occupational TB occurs in a wide variety of settings, which means that the risk varies substantially, and control measures differ, from one facility to another. OSHA's proposed TB standard has been tailored to recognize these differences. With respect to the background risk of exposure, the OSHA standard distinguishes between work settings in counties that have had no cases of TB in one of the past two years and fewer than 6 cases in the other of the past two years, work settings in counties with one or more cases of TB in both of the past two years or that have had 6 or more cases of TB in one of the past two years, and work settings that have encountered 6 or more cases of TB in the past 12 months. In addition, the OSHA standard treats different types of exposure to TB differently. For example, the standard has different requirements for employers who own facilities that treat TB patients, employers whose client populations have high TB rates, employers whose employees (such as attorneys and social service providers) visit patients who have been identified as having suspected or confirmed cases of TB, employers whose employees engage in various high hazard procedures, employers whose employees provide maintenance for ventilation systems serving confirmed or suspected TB patients, and employers who provide personnel to treat patients in their own homes. In part because of these many distinctions, the SBREFA Panel found that the regulation was difficult for many employers to understand (Ex. 12). To make the tailoring of the standard to specific situations easier to see, OSHA has developed tables showing which provisions of the standard are most likely to apply to employers in different circumstances and in various affected sectors (see the Scope paragraph discussion in Section X of the Preamble, "Summary and Explanation"). In addition, OSHA intends to provide extensive outreach when the standard is published in final form. OSHA solicits comments on other ways to avoid a "one-size-fits-all" standard while at the same time making the standard easier to follow. For example, would developing a flow chart and/or expert system that asks employers a series of questions and then directs employers to applicable requirements be an aid to affected small entities?

Description of the Proposed Standard. The proposed rule would require that employers develop and implement exposure control plans; institute work practice and engineering controls; provide respiratory protection in various situations; provide medical surveillance (e.g., tuberculin skin testing, medical histories, medical management, medical follow-up, medical removal); and communicate hazards through the use of signs, labels, and training. These proposed requirements are discussed in greater detail in the Introduction (Chapter I) of this analysis.

The proposed standard would also require that employers establish and maintain medical, training, illness/injury, and engineering control maintenance and performance monitoring records. All establishments affected by the proposed rule would be affected by these proposed requirements. However, only establishments with engineering controls would be required to maintain records of the maintenance and monitoring of engineering controls.

In estimating the cost of establishing and maintaining medical records, OSHA used the wage rate of a clerical worker with some knowledge of medical recordkeeping as the base wage. However, the knowledge required to perform such duties can be acquired by most clerical workers with little effort. All recordkeeping requirements included in the proposed rule could therefore be performed by the existing staff in any of the covered industries. A detailed description of the proposed requirements appears in the Introduction and in the Costs of Compliance chapters of this analysis.

Relevant Federal Rules That May Duplicate, Overlap, or Conflict With the Proposed Rule

On October 28, 1994, the Centers for Disease Control and Prevention (CDC) of the U.S. Department of Health and Human Services published "Guidelines for Preventing the Transmission of Mycobacterium tuberculosis in Health-Care Facilities," which recommends that facilities adopt many of the requirements included in this proposed standard. CDC has also published guidelines for the prevention of transmission of TB in homeless shelters, long-term care facilities for the elderly, and correctional institutions. OSHA has consulted with CDC in developing the proposed standard, and the basic elements of the standard correspond to the basic elements in the CDC guidelines. However, the CDC publication is only recommendatory and is therefore not enforceable. OSHA's studies (see chapters IV and V) show that few facilities are following all elements of these guidelines. Further, many portions of the CDC guidelines are written in language that does not lend itself to enforcement even if the guidelines were made mandatory. For example, portions of the CDC guidelines for health care facilities suggest that the employer "consider" adopting certain controls. A fuller discussion of the similarities and differences between OSHA's proposed rule and the CDC's recommendations is provided in Section III of the Preamble, which describes the events leading to the proposed standard. Although the U.S. Public Health Service has overall responsibility for the control of TB in the U.S. population, OSHA is the only agency specifically mandated to address the problem of TB transmission in occupational settings.

The Health Care Financing Administration (HCFA) of the U.S. Department of Health and Human Services requires that facilities undergo an initial accreditation inspection prior to receiving Medicare and Medicaid funding. Such facilities include hospitals, nursing homes and other long-term care facilities, and clinical laboratories. Hospitals are reinspected annually, nursing homes every 15 months, and laboratories every two years. One of the requirements of such accreditation is the implementation of an infection control program. However, unlike the OSHA proposed rule, HCFA's requirements do not specify the elements that must be included in such a program. HCFA may cite facilities with poor results for specific program deficiencies but does not have the authority to cite facilities for failing to include specific elements in their infection control programs, unless those program elements are specifically required by an OSHA standard. This means that in the absence of an OSHA TB standard, HCFA could not require implementation of specific controls. The proposed rule does not in any way conflict with HCFA requirements. Further, the existing HCFA requirements have not ensured that health care facilities adopt the elements of an effective infection control and have not prevented outbreaks of TB in this workforce.

next page