Regulations (Preambles to Final Rules) - Table of Contents Regulations (Preambles to Final Rules) - Table of Contents
• Record Type: Occupational Exposure to Bloodborne Pathogens
• Section: 5
• Title: Section 5 - V. Quantitative Risk Assessment

V. Quantitative Risk Assessment

A. 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 ruled that 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 risk of health impairment under existing exposure 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 also stated "that the Act does limit the Secretary's power to require the elimination of significant risks" [448 U.S. 644].

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 standards, it reaffirmed its previous position in "benzene" 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 risk assessments should be put into quantitative terms to the extent possible.

OSHA has presented its views on risk assessment in detail in several proceedings such as the Supplemental Statement of Reasons for Final Rule for occupational exposure to Inorganic Arsenic (48 FR 1867), the Notice of Proposed Rulemaking for occupational exposure to Ethylene Dibromide (48 FR 45956), as well as in the rulemaking record for exposure to Bloodborne Pathogens, including the preamble to the proposed standard, the preliminary quantitative risk assessment for HBV infection, and the qualitative risk assessment for HIV infection in an occupational setting.

Quantifying the risk associated with exposure to bloodborne diseases such as HBV or HIV is different than quantifying the risk associated with exposure to toxic chemicals, the risks that OSHA has typically quantified. For most of these chemicals, response, in the form of adverse health effects, is associated with cumulative dose, and workers risk chronic health effects from long term exposure to airborne concentrations of the chemical. The response associated with exposure to bloodborne pathogens does not depend on cumulative dose acquired through years of exposure. With each exposure, either infection occurs or it does not occur. While repeated exposure increases the cumulative risk of infection within a specified time period, each exposure is associated with a unique risk which is the same for anyone exposed to the virus and depends upon the virulence of the pathogen, the size of the delivered dose, the route of exposure, among other factors, and not upon any prior exposure. Thus, in the case of bloodborne diseases, the best way to reduce the risk of transmission is by reducing exposure.

HBV is a bloodborne pathogen for which there are sufficient data to quantify the risk of infection from occupational exposure to blood or other potentially infectious material (hereafter referred to as occupational exposure) for an entire population of workers. Further, healthcare workers are the only occupational group for which data on the risk of HBV infection in an occupational setting are available to OSHA. A healthcare worker is defined as anyone employed in the healthcare industry. It includes persons working in medical and dental labs, nursing homes, dialysis centers, housekeeping staff as well as doctors and nurses (for a more extensive listing of occupations see TABLE VII-4.) OSHA will use the available data to estimate the annual and lifetime occupational HBV infection risk to healthcare workers with occupational exposure (approximately 4.9 million employees). From this OSHA will extrapolate the HBV risk estimate to non-healthcare employees with occupational exposure to blood or other potentially infectious material such as law enforcement officers and fire fighters (approximately 1.2 million.) OSHA believes and the record supports it, that it is the exposure to blood or other potentially infectious materials that places the employees at risk for hepatitis B and not some other factor unique to healthcare employment. This conclusion is supported by the epidemiological studies reviewed for this rulemaking as well (Exs. 4-13; 4-14; 4-16; 6-65). Therefore, OSHA will assume that the risk to non-healthcare workers with occupational exposure is similar to the risk of healthcare workers with equivalent occupational exposure. However, the record does not contain any usable quantitative data on non-healthcare workers and therefore, OSHA's extrapolated estimate may be higher or lower than the true occupational risk to non-healthcare workers.

A number of epidemiological studies demonstrate an increased prevalence of hepatitis B markers in the blood of healthcare workers with blood exposure, and a brief review of some of these studies is presented below, followed by OSHA's final assessment of HBV risk posed by occupational exposure to bloodborne pathogens including a summary and evaluation of comments submitted to the record. Finally, OSHA presents a qualitative risk assessment for infection from occupational exposure to HIV.

B. Review of the Epidemiology of HBV Infection in Healthcare Workers

Numerous epidemiological studies have measured the prevalence of HBV infection among healthcare workers. These studies determined what proportion of healthcare workers had ever been infected with HBV and measured prevalence as the proportion of workers with any serological marker of past or present HBV infection. Most of the studies relied upon the voluntary cooperation of the study population, so there is some chance for bias to be introduced into any estimate of HBV prevalence. Healthcare workers who know they are infected with HBV at the time of study or who know they are HBV carriers may decline to participate in a study which they may feel could jeopardize their careers. This would lead to an underestimate of the prevalence of HBV infection among healthcare workers. However, the inclusion of healthcare workers who engage in non-occupational high risk behaviors could potentially lead to an overestimate of the annual HBV infection risk.

Jovanovich et al. did not rely upon voluntary participation in their study of HBV prevalence among workers at a 1000-bed community hospital in Detroit (Ex. 4-14). The authors reported a high prevalence of HBV among employees in worksites where blood and other potentially infectious materials are present (Ex. 4-14). All new employees were screened for HBV markers at the time of hire, and the blood tests were repeated every six months thereafter for all employees designated as being at high risk for HBV infection. In the hemodialysis unit, these tests were repeated monthly. This design allowed investigators to determine not only the HBV prevalence but also the conversion rate to HBV seropositivity per 100 employee-years. Jovanovich et al. reported the highest prevalence of HBV among the emergency room staff (27.9%), followed by the operating room staff (25.2%), the hemodialysis unit staff (17.2%), the dental staff and oral surgery (15.4%), and the staff of the intensive care unit (12.7%) (Ex. 4-14). The authors did not state what proportions of the study subjects were in specific occupations (e.g., physicians, nurses, etc.). The emergency room staff experienced the highest rate of conversion to HBV seropositivity with a conversion rate of 11.7 per 100 employee-years (Ex. 4-14).

Like Jovanovich et al., Dienstag and Ryan found the highest prevalence of serological markers for HBV among the emergency room staff, (specifically nurses), in a study of workers at an 1100 bed urban teaching hospital in Boston (Ex. 4-13). This study relied upon voluntary participation, and of 830 staff at the hospital, 624 or 75% agreed to participate. Among workers with frequent blood contact, the prevalence of HBV serological markers was 21.2% versus 8.6% for workers with occasional, rare, or no blood contact (p< .001). The highest rates of soro positivity positivity were found among emergency room nurses, pathology staff, blood bank staff, laboratory technicians, intravenous teams, and surgical house officers. The prevalence of HBV serological markers was 30% among emergency room nurses and was in excess of 15% in each of the other groups. Workers with less contact had HBV serological markers at rates between 5% and 10%. Four of thirty-two administrators, (16%), were found to have serological markers of HBV infection, but the authors stated that the high observed prevalence among this group may have been related to the inclusion of two persons known to be members of a high risk group. All of these groups were compared to a population of 462 volunteer blood donors, which had a 5% prevalence of HBV markers. Neither frequency of patient contact nor socioeconomic status (SES), as measured by years of education, were found to be associated with the prevalence of HBV serological markers. SES is often associated with prevalence of HBV infection but not among this cohort. Indeed, as demonstrated in TABLE V-1, among workers with a comparable level of education, frequency of blood contact was statistically significantly associated with HBV prevalence. Prevalence increased with age for all employees regardless of degree of blood contact, but prevalence was observed to increase with years in occupation only for workers with frequent blood contact.

                              TABLE  V - 1
            CORRELATION BETWEEN FREQUENCY OF BLOOD CONTACT
        AND HBV PREVALENCE IN HOSPITAL WORKERS WITH UNIFORM
          SOCIOECONOMIC STATUS MEASURED BY YEARS OF EDUCATION(a)

__________________________________________________________________________
                Exposure                # with HBV    Odds    Chi-square
Personnel       to Blood        N      Markers (%)    Ratio    (p-value)
__________________________________________________________________________

Physicians(b)  Frequent        81        17 (21%)     3.11        6.02
              Infrequent       89         7 (8%)               (p< .02)

Nurses (c)     Frequent       104        22 (21%)     2.80        7.16
              Infrequent      126        11 (9%)               (p< .01)
_________________________________________________________________________
  Footnote(a)  Data from Table 2 of Dienstag and Ryan (Ex. 4-13).
  Footnote(b)  Median level of education for the physicians was 20 years.
  Footnote(c)  Median level of education for the nurses was 17 years.

Pattison et al. reported similar findings of the relationship between frequency of blood contact and the prevalence of HBV in an earlier study conducted between 1972 and 1974 at a 495 bed urban hospital in Arizona (Ex. 6-65). The study population was selected from consecutive employees undergoing yearly physical examination on the anniversary of their initial employment examination. Except for physicians, study participants had been affiliated with the hospital for at least two years. Over 99% of the eligible employees, excluding physicians, representing 40% of all hospital personnel participated in the study (Ex. 6-65). The overall prevalence of HBV serological markers was 14.4% (Ex. 6-65). No association was observed between frequency of patient contact and prevalence of HBV, but the association between frequency of blood contact and prevalence of HBV was statistically significant (p< .05) (Ex. 6-65). Among workers with frequent blood contact, the seroprevalence of HBV markers was 18.9%; for workers with occasional blood contact, it was 13.4%; and for workers with no blood contact, it was 11.4%. Socioeconomic status, as measured by the Hollingshead Index derived from educational level attained and category of employment (highest socioeconomic level corresponding to Hollingshead Index 1; lowest socioeconomic level corresponding to Hollingshead Index 5), was statistically significantly associated with HBV prevalence but only when categories 1 through 4 were combined and compared to category 5. Among workers with similar Hollingshead indices (i.e. controlling for socioeconomic status), workers with frequent or occasional blood contact were twice as likely to have serological markers for HBV as were workers with no blood contact (Ex. 6-65). This is demonstrated in TABLE V-2.

                            TABLE  V - 2
            CORRELATION BETWEEN FREQUENCY OF BLOOD CONTACT
          AND HBV PREVALENCE IN HOSPITAL WORKERS WITH SIMILAR
        SOCIOECONOMIC STATUS MEASURED BY THE HOLLINGSHEAD INDEX(a)
___________________________________________________________________

Hollingshead   Exposure            # with HBV    Odds   Chi-square
Index (b)      to Blood(c)    N    Markers (%)   Ratio  (p-value)
___________________________________________________________________

1 and 2         Freq/Occ     136    18 (13.2%)   2.21       1.09
                 Never        31     2 (6.5%)            (p< .25)

3 and 4         Freq/Occ     125    20 (16.0%)   1.97       2.56
                 Never       102     9 (8.8%)            (p< .10)

5               Freq/Occ      41    13 (31.7%)   2.55       4.34
                 Never        78    12 (15.4%)           (p< .05)

Total           Freq/Occ     302    51 (16.9%)   1.66       3.57
                 Never       211    23 (10.9%)           (p< .10)
___________________________________________________________________
  Footnote(a)  Data from Table 3 of Pattison et al. (Ex. 6-65).
  Footnote(b)  The Hollingshead Index is a measure of socioeconomic status
derived from educational level attained and category of employment. The
highest socioeconomic level corresponds to Hollingshead Index 1; the
lowest socioeconomic level corresponds to Hollingshead Index 5.
  Footnote(c)  Pattison categorized blood exposure as frequent or
occasional (Freq/Occ) versus never.

In a more recent study by Hadler et al., frequency of blood contact but not frequency of patient contact was again shown to be strongly related to HBV prevalence (Ex. 4-16). Of all employees at three urban teaching hospitals and two midwest community hospitals, 5,697 (36%) participated in this study. Serological markers of past or present HBV infection were found in 14.2% of the study population (Ex. 4-16). For workers with frequent blood contact, the prevalence of HBV markers increased with duration in occupation at a rate of 1.05 infection per 100 person-years (R=.95; p< .01), and for workers with occasional blood contact, the prevalence increased at a rate of .71 infections per 100 person-years (R=.85; p=.05) (Ex. 4-16). Among workers with no blood contact, HBV prevalence was constant over the duration of employment. Hadler et al. also found that frequency of needle accidents was related to HBV prevalence. Among workers with frequent or occasional needle accidents, HBV prevalence increased with duration in occupation at a rate of .80 infections per 100 person-years, and among workers with rare needle accidents, prevalence increased at a rate of .72 infections per 100 person-years (Ex. 4-16). Among workers who reported no needle accidents, the increase in HBV prevalence with duration in occupation was much lower (.24 infections per 100 person-years). When subjects were stratified into groups by degree of blood contact, frequency of needle contact was positively associated with HBV infection rates only in persons with frequent blood contact and not in persons having occasional or no blood contact (Ex. 4-16).

Needlesticks and cuts with sharp objects are by no means the only way workers with exposure to blood or other potentially infectious material can be exposed to the hepatitis B virus. In a study of the transmission of HBV in clinical laboratory areas, Lauer et al. found that 26 of 76 (34%) environmental surfaces sampled were positive for hepatitis B surface antigen (HBsAg) (Ex. 6-56). Samples were taken in a dialysis room specifically used for patients who had HBV infections at the time of dialysis. In addition, samples were collected in the clinical laboratory where tests were done on blood samples drawn from HBV-infected dialysis patients. The HBsAg was found on the outside of 6 of 11 (55%) of the sampled blood-specimen containers and 4 of 9 (44%) of the sampled serum-specimen containers (Ex. 6-60). The gloves and bare hands of personnel who had contact with the blood- and serum-specimen containers were also sampled, and two of the three samples taken, including one from a bare hand, were positive for HBsAg (Ex. 6-50). Other contaminated surfaces included the handle portion of pipetting aids, marking devices, and an assay instrument for complete determination of blood cell counts. The authors stated that their "data indicate that transmission of HBV in the clinical laboratory is subtle and mainly via hand contact with contaminated items during the various steps of blood processing. These data support the concept that the portal of entry of HBV is through inapparent breaks in skin and mucous membranes." (Ex. 6-56, p. 513).

C. Quantitative Assessment of HBV Risk

OSHA's quantitative risk assessment focuses on HBV infection in healthcare workers because healthcare workers with occupational exposure to blood or other potentially infectious material constitute the only occupational group with such exposure for which OSHA has sufficient data to quantitatively estimate the occupational risk of HBV infection (for a listing of occupations included see Section VII, TABLE VII-4.) OSHA believes, and the record supports it, that it is the exposure to blood or other potentially infectious material that places these workers at risk for HBV and not some other factor unique to healthcare workers. This conclusion is borne out by the epidemiological studies reviewed in the previous section. Further, OSHA believes that the risk to non-healthcare workers with occupational exposure is similar to the risk of healthcare workers. Therefore, OSHA will use the data available for healthcare workers with occupational exposure to predict the HBV infection risk to any worker with occupational exposure to blood or other potentially infectious material.

Estimates of the incidence of HBV infection in the U.S. population in general and among healthcare workers in particular come from the Hepatitis Branch of the Center for Infectious Disease, U.S. Public Health Service's Centers for Disease Control (CDC). There are two systems for collecting information on hepatitis: the CDC National Morbidity Reporting System and the Viral Hepatitis Surveillance Program (VHSP). The National Morbidity Reporting System collects data on the number and type of hepatitis infections as well as the patients' ages in reported cases. The VHSP collects serological and epidemiological data pertaining to risk factors for the disease (Ex. 6-217). Based on the 1988 national hepatitis surveillance data, the CDC estimates that there were 280,000 HBV infections in 1988 in the U.S. Of these, it is estimated that 8,700 were in persons whose only source of infection was related to healthcare employment (Ex. 298). "This estimate is derived from cases of hepatitis B reported to the Viral Hepatitis B Surveillance Program (VHSP) in which employment as a health care worker was the only source of infection and from surveillance data in the Sentinel Counties Study of Viral Hepatitis." (Ex. 298). Only a fraction of the 280,000 estimated infections are actually reported to the CDC because most infections produce no symptoms and people are unaware that they have contracted hepatitis B. Furthermore, even when people become ill enough to seek medical help, the disease is not always correctly diagnosed or faithfully reported. CDC estimates that approximately 1 in 12 cases of hepatitis B is actually reported (Stephen C. Hadler, M.D., Tr. 9/18/89, p. 12). For its risk assessment, OSHA will use the latest available data, as reported by CDC, and assume that exactly 280,000 HBV infections occur each year.

OSHA estimates that there are approximately 4.9 (4,897,595) million healthcare workers with occupational exposure putting them at risk for bloodborne diseases including HBV (see Benefits in Section VII). A portion of the 4.9 million workers are not at risk for HBV infection because of immunity. OSHA estimates that approximately 2.6 million (2,568,974) adults have received either the plasma-derived or the yeast hepatitis B vaccine of which approximately 2.0 million (2,029,189) are estimated to be healthcare workers (See section VII). Further, 96% of the vaccinated workers are considered to have achieved immunity to further infection (Ex.292). In addition, CDC estimates that between 15% and 30% of healthcare workers with occupational exposure (734,700 to 1,469,400) have already been infected with HBV and are now immune to further infection (Ex. 6-199).

The number of people vaccinated, as estimated by OSHA, differ slightly from those reported by Merck Sharp and Dohme (Merck) (Ex. 292). In their post-hearing comment Merck reported sales of approximately 5.6 million doses of the three-dose series of HEPTAVAX-B and 3.0 million doses of the three-dose series of RECOMBIVAX-B (Ex. 292). The company then extrapolated from the sales figures that, at a minimum, 2.9 million people have received the three-dose series of the hepatitis B vaccine (Ex. 292). Further, Merck estimated 85% of 2.9 million, or 2.5 million are currently in the work force covered by the proposed standard (Ronald W. Ellis, M.D., Tr. 9/18/89, p. 77). The discrepancy between Merck's and OSHA's numbers may be explained by the fact that Merck's estimate reflects company sales and not necessarily amount of vaccine actually used, where OSHA's estimate reflects the minimum number of vaccinated people who have received the three-dose regimen as derived from the Agency's survey (Exs. 264; 266). For the risk assessment, the Agency will use the results from OSHA's survey as being more reflective of the actual number of healthcare employees who are fully protected against HBV infection. However, the survey most likely underestimates the number of vaccinated individuals because not everyone will receive the three-dose regimen.

There is abundant testimony and other evidence in the record demonstrating the effectiveness of the vaccine. Protective HB antibody levels were present in over 96% of healthy adults who have received the series (Ronald W. Ellis, M.D., Tr. 9/18/89, p. 86; Ex. 292). Therefore, OSHA estimates that (2,029,189 x .96)=1,948,021 workers, are immune to HBV as the result of vaccination. Prior infection and vaccination, remove between 2,833,000 and 3,390,500 from the pool of 4.9 million healthcare workers with occupational exposure, leaving between 2,065,000 and 2,507,500 workers at risk for HBV infection. To estimate the number of healthcare workers at risk OSHA did the following: The number of healthcare workers immune from vaccination (1,948,000) was first subtracted from the total number at risk (4,898,000). The remaining pool was further reduced by either 15% or 30% to account for the range of people who are immune to hepatitis B because of previous infections.

Of the 280,000 HBV infections each year (based on 1988 Hepatitis Surveillance data), CDC estimates that 8,700 cases occur in health care workers with occupational exposure (Ex. 298). If between 2,065,000 and 2,507,500 healthcare workers are at risk, then the annual HBV infection rate for these workers is between 3.47 and 4.21 per 1,000 exposed workers (See TABLE V-3). OSHA's estimate of the annual HBV infection rate is an empirical estimate of the probability of HBV infection for healthcare workers exposed to blood or other potentially infectious materials who lack immunity either because of prior infection or vaccination. This estimate of the annual HBV infection risk applies to the population of healthcare workers with occupational exposure and not to a specific healthcare worker picked at random. The estimate of the annual HBV infection risk for a healthcare worker who has been vaccinated or has been previously infected (i.e. is immune to HBV) is zero. The annual HBV infection risk for any healthcare worker with occupational exposure randomly selected will depend entirely upon the immune status of that worker.

Clearly it is possible for workers with exposure to blood to become infected with HBV by means other than occupational exposure. The virus can be transmitted sexually and by non-occupational exposure to blood. In addition, over 50% of all cases of HBV reported to the Centers for Disease Control in 1985 had no known risk factors (Ex. 6-217).

Several commenters viewed OSHA's estimates as being overestimates of the true risk by stating that OSHA did not appropriately consider the fact that most healthcare workers who are infected with hepatitis B probably contracted their infection due to factors outside the workplace (Ex. 20-2879C). In fact, OSHA took measures to exclude the effect of high risk behaviors in healthcare workers by estimating the risk attributable to occupational exposure. The risk attributable to occupational exposure is the difference between the risk faced by exposed workers and the background risk faced by the general population. In order to remove that portion of HBV cases in healthcare workers that might be due to IV drug use or other known risk factors, the Agency subtracted from the healthcare worker risk the background (population) risk of HBV infection. Dr. Stephen Hadler, an expert on viral hepatitis from the Hepatitis Branch of the CDC, supported this methodology of calculating the risk of HBV infection attributable to occupational exposure in his testimony (Stephen C. Hadler, M.D., Tr. 9/18/89, p. 36).

There were 193,220,000 residents, over the age of 15, in the U.S. in 1988 (Ex. L6-665) (note: Exhibit L6-665 is an updated version of Exhibit 6-389. Reliance on the old data would not have changed the results of the annual HBV infection rate). Of these, it is estimated that 4.8% (approximately 9.3 million) have been infected with hepatitis B and, therefore, are immune (Ex. 6-390). In addition, we will assume that all of the 2,569,000 persons who have received the hepatitis B vaccine are adults and that 96% of them (2,466,000) are immune. Therefore, of the 193 million adults in the U.S., approximately 182 million are at risk of HBV infection. The number of adults at risk in the U.S. was estimated by first removing the number of adults immune from vaccination (2,466,000) from the total population and subsequently reducing the remaining pool by 4.8%. Given that there are 280,000 cases of infection each year, the annual infection rate is 1.54 infections per 1000 adults. This estimated infection rate for the entire adult population constitutes the background risk for HBV. In other words, OSHA estimates that the probability that an adult in the U.S. will be infected with HBV this year is .00154. Estimates of the populations at risk and their HBV infection rates are given in TABLE V-3.

                          TABLE  V - 3
        ESTIMATE OF POPULATIONS AT RISK FOR HBV INFECTION(a)

_________________________________________________________________________
                                   U.S. Adults         Healthcare workers
                                                       with occupational
                                                       exposure
_________________________________________________________________________
Number in population                193,220,000                 4,898,000
Percent immune(c)                          4.80                   15 - 30
Number immune                         9,275,000       734,700 - 1,469,400
Number vaccinated                     2,569,000                 2,029,000
Percent vaccinated                         1.33                     41.43
Number immune from vaccination(b)     2,466,000                 1,948,000
Number at risk                      181,598,000     2,065,000 - 2,507,500
Annual HBV infection rate per 1000       1.542              3.470 - 4.213
  exposed healthcare workers who
  lack immunity
_________________________________________________________________________
  Footnote(a) Most numbers have been rounded to the nearest thousand.
  Footnote(b) This assumes vaccination efficacy to be 96%.
  Footnote(c) Percent immune is the proportion of the population which has
already been infected with the HB virus. Previous infection confers
life-long immunity.

OSHA's estimate of the background risk of HBV infection is probably much higher than the actual risk faced by most adults. Certain behaviors are known to substantially increase the risk of HBV infection, but not all adults engage in these behaviors with equal probability. For example, a recent General Social Survey conducted in early 1988 recorded homosexual activity among 3.2% of 504 sexually active men in the previous 12 months, yet the proportion of HBV cases associated with homosexual activity in 1987 in the CDC's Sentinel County study was 9%, nearly three times as large as the percentage of homosexual activity reported (Exs. 6-342; 6-321). Intravenous drug users, who accounted for 28% of the HBV cases in 1987 in the same CDC study, are another group which are disproportionately represented in the number of HBV cases as compared to their number in the adult population. Removing the HBV cases associated with homosexual activity and IV drug use from the annual number of cases and removing adult men who engage in homosexual activity and IV drug users from the population at risk would substantially reduce OSHA's estimate of the background risk of infection because a greater proportion of cases would be removed from the number of HBV cases (i.e. the numerator) than the proportion of people removed from the population at risk (i.e. the denominator). Unfortunately, there are no reliable estimates of the number of people engaging in high risk behaviors such as homosexual activity or IV drug use. Therefore, OSHA must rely on its estimate of 1.54 HBV infections per 1000 adults as its estimate of the background risk, but the Agency is aware that the true risk for most adults in the U.S., and therefore the background risk for healthcare workers, is probably much lower.

As outlined in the discussion of the health effects of HBV, there are a number of possible outcomes following infection. Between two thirds and three fourths of all infections result in either no symptoms of infection or a relatively mild flu-like illness. Between 25% and 33% of the infections, however, take a much more severe clinical course. As noted above, the symptoms include jaundice, dark urine, extreme fatigue, anorexia, nausea, abdominal pain, and sometimes joint pain, rash, and fever. For its risk assessment, OSHA will use the lower estimate of 25% as the proportion of HBV infections which take a more severe clinical course. Hospitalization is required in about 20% of the more severe clinical cases.

CDC estimates that 2.225% of HBV infections lead to death (Ex. 6-392). Death from fulminant hepatitis occurs in 0.125% of cases (Ex. 6-392). Death from cirrhosis of the liver is estimated to occur in 1.7% of cases, and death from primary hepatocellular carcinoma is estimated to occur in 0.4% of cases (Ex. 6-392). Between 5% and 10% of individuals infected with HBV become chronic carriers of the virus (Ex. 6-392). These individuals represent a pool from which the disease may spread. About 25% of the chronic carriers suffer from chronic active hepatitis (Ex. 6-392). The estimated numbers of infections that result in any of these outcomes each year in both the adult population and in the population of healthcare workers is presented in TABLE V-4. Among the adult population, approximately 182 million persons are estimated to be at risk for HBV. As shown in TABLE V-3, there are between 2,065,000 and 2,507,500 healthcare workers annually at risk for HBV infection. Using the estimates of annual HBV infections from TABLE V-4 and the population estimates from TABLE V-4, the annual risk of HBV infection for the adult population and for any healthcare worker with occupational exposure have been calculated and are presented as rates per 1,000 exposed workers in TABLE V-5. A healthcare worker is defined as anyone employed in the healthcare industry. It includes persons working in medical and dental labs, nursing homes, dialysis centers, housekeeping staff as well as doctors and nurses. OSHA assumes that the annual risk of HBV infection for workers with occupational exposure to blood and other potentially infectious materials is similar to that of healthcare workers with equivalent exposures.

                               TABLE  V-4
     ESTIMATES OF THE NUMBER OF ANNUAL HBV INFECTIONS AND OUTCOMES
      IN THE U.S. POPULATION AND AMONG HEALTHCARE WORKERS EXPOSED
        TO BLOOD OR OTHER POTENTIALLY INFECTIOUS MATERIAL(a)
_____________________________________________________________________
                                      U.S. Adults        Healthcare
                                                         workers with
                                                         occupational
                                                         exposure
_____________________________________________________________________
HBV Infections                           280,000                8,700
Clinical Illness (25%)                    70,000                2,175
Hospitalized (5%)                         14,000                  435
HBV Carrier (5% - 10%)           14,000 - 28,000            435 - 870
Chronic HBV (25% Carriers)         3,500 - 7,000            109 - 218
Fulminant Death (.125%)                      350                   11
Death - Cirrhosis (1.7%)                   4,760                  148
Death - PHC(b) (0.4%)                      1,120                   35
All Deaths (2.225%)                        6,230                  194
______________________________________________________________________
  Footnote(a) Data from Ex. 298 and Ex. 6-392.
  Footnote(b) Primary Hepatocellular carcinoma.

TABLE V-6 presents the risk attributable to occupational exposure for HBV infection and its outcomes per 1000 exposed workers. The annual risk attributable to occupational exposure is simply the difference between the annual risk faced by exposed workers and the annual risk faced by the adult population, both given in TABLE V-5. Because Section (6)(b)(5) of the OSH Act states that no employee shall suffer "material impairment of health or functional capacity even if such an employee has regular exposure to the hazard dealt with . . . for the period of his working life", OSHA has converted the attributable annual risk into an attributable lifetime risk on the assumption that a worker is employed in his or her occupation for 45 years. TABLE V-6 shows that for every 1000 workers with occupational exposure to blood or other potentially infectious material, between 83 and 113 will become infected with HBV over the course of their working lifetime because of occupational exposure to the virus. Of these, 21 to 30 will suffer clinical illness and 4 to 6 will need hospitalization. Between 4 and 12 of the cases with clinical illness will become chronic carriers, and 1 to 3 of them will suffer from chronic hepatitis. HBV infection from occupational exposure will lead to the death of 2 to 3 of these 1000 exposed workers.

                             TABLE  V - 5
    ESTIMATES OF THE ANNUAL RISK FOR HBV INFECTION AND ITS OUTCOMES
   IN THE U.S. ADULT POPULATION AND AMONG HEALTHCARE WORKERS EXPOSED
         TO BLOOD OR OTHER POTENTIALLY INFECTIOUS MATERIAL(a)

____________________________________________________________________
                                U.S. Adults     Healthcare workers
                                                with occupational
                                                exposure(b)
____________________________________________________________________
HBV Infections                        1.542         3.470 - 4.213
Clinical Illness (25%)                0.386         0.868 - 1.053
Hospitalized (5%)                     0.077         0.174 - 0.211
HBV Carrier (5% - 10%)        0.077 - 0.154         0.174 - 0.421
Chronic HBV (25% Carriers)    0.019 - 0.039         0.043 - 0.105
Fulminant Death (.125%)               0.002         0.004 - 0.005
Death - Cirrhosis (1.7%)              0.026         0.059 - 0.072
Death - PHC(c) (0.4%)                 0.006         0.014 - 0.017
All Deaths (2.225%)                   0.034         0.077 - 0.094
____________________________________________________________________
  Footnote(a) Risks are expressed as the number of events per 1,000
exposed healthcare workers who lack immunity.
  Footnote(b) Risks for exposed workers are estimated assuming 15%
and 30% of the workers had a previous infection and are thus immune.
  Footnote(c) Primary Hepatocellular carcinoma.
                          TABLE  V - 6
  HBV RISK ATTRIBUTABLE TO OCCUPATIONAL EXPOSURE FOR HEALTHCARE WORKERS
   EXPOSED TO BLOOD OR OTHER POTENTIALLY INFECTIOUS MATERIAL(a),(b)
_____________________________________________________________________
                                                         Lifetime
                                     Annual Risk(c)      occupational
                                                          risk(d)
_____________________________________________________________________
HBV Infections                      1.928 - 2.671      83.18 - 113.40
Clinical Illness (25%)              0.482 - 0.668       21.46 - 29.61
Hospitalized (5%)                   0.096 - 0.134         4.33 - 5.99
HBV Carrier (5% - 10%)              0.096 - 0.267        4.33 - 11.95
Chronic HBV (25% Carriers)          0.024 - 0.067         1.08 - 3.00
Fulminant Death (.125%)             0.002 - 0.003         0.11 - 0.15
Death - Cirrhosis (1.7%)            0.033 - 0.045         1.47 - 2.04
Death - PHC(b) (0.4%)               0.008 - 0.011         0.35 - 0.48
All Deaths (2.225%)                 0.043 - 0.059         1.93 - 2.67
_____________________________________________________________________
  Footnote(a)  Risks are expressed as the number of events per 1,000
exposed healthcare workers who lack immunity.
  Footnote(b)  The risk attributable to occupational exposure is the
difference between the annual risk faced by exposed workers and the
annual risk faced by the adult population, both given in TABLE V-5.
  Footnote(c)  Risks for exposed workers are estimated assuming 15%
and 30% of the workers had a previous infections and are thus immune.
  Footnote(d)  Assumes 45 years of occupational exposure and is
calculated as [1 - (1 - p)(45)], where p is the annual risk divided
by 1,000. Assumes p is constant through time.

OSHA's estimate of the risk of HBV infection attributable to occupational exposure to blood or other potentially infectious material is most likely an underestimate of the true risk. As noted above, the true risk of HBV infection among the majority of U.S. adults is probably much lower than OSHA's estimate of the background risk since the majority of adults do not engage in the high risk behaviors associated with a large proportion of HBV infections. By overestimating the background risk, OSHA has probably underestimated the risk attributable to occupational exposure. In addition, OSHA's estimate of the number of people with immunity to hepatitis B because they have been vaccinated reflects the minimum number of vaccinated people who have received the three-dose regimen. However, OSHA's survey most likely underestimates the number of vaccinated individuals because not everyone receives the three-dose regimen. In doing so, OSHA's estimate of HBV risk attributable to occupational exposure may be slightly underestimated.

Nonetheless, OSHA's calculations show that workers with occupational exposure are at a substantially increased risk of infection, clinical illness, hospitalization, chronic hepatitis, and death over the course of their working lifetimes. These workers are at an increased risk of becoming HBV carriers which is frequently associated with serious chronic illness and of transmitting the infection sexually and perinatally.

Since 1982, a plasma-derived hepatitis B vaccine, HEPTAVAX B, has been available. In July of 1986, a genetically engineered hepatitis B vaccine, RECOMBIVAX HB, manufactured by Merck Sharp & Dohme, was licensed by the U.S. Food and Drug Administration. In August of 1989, a second recombinant DNA hepatitis B vaccine, ENGERIX B, manufactured by SmithKline and Beecham, was approved for marketing in the United States by the U. S. Food and Drug Administration (Jerome A. Boscia, M.D., Tr. 12/19/89, p.985). "Due to the increasing popularity of Merck's alternative recombinant vaccine, as well as the increasing difficulty of obtaining plasma suitable for the manufacturing process, HEPTAVAX B is no longer in production" (Ronald W. Ellis, M.D., Tr. 9/18/89, p. 74). All vaccines have proven to be highly effective in preventing hepatitis B infection in high risk populations. When given in the recommended three dose series, Merck Sharp & Dohme reports that RECOMBIVAX HB has been found to induce protective antibodies in over 95% of healthy adults 20-39 years of age, but like the plasma-derived vaccine, the new vaccine induced a somewhat lower antibody response in older adults (Ex. 6-176). Although it was stated during testimony that large-scale studies directly measuring the efficacy of the recombinant hepatitis B vaccine in adults have not been done, Merck has sponsored numerous clinical studies of vaccine immunogenicity. The vaccine's immunogenicity is impacted by the age of the recipient with response rates lower in older adults than in younger adults. When a weighted average was taken to account for the age distribution of the recipients, it was determined that 96.4% of the healthy adults who receive the recombinant vaccine developed protective levels of antibody. However, when the data were adjusted to account for the difference between the age distribution of employees covered by this standard and the age distribution of recipients of the vaccine in the Merck study the immunogenicity rate was adjusted down to 92.7 percent (Ronald W. Ellis, M.D., Tr. 9/18/89, pp.85-86). Based on Merck's latest data, the seroconversion rate using RECOMBIVAX HB is expected to reach 99% when given in the recommended three-dose series to healthy adults between 20 and 29 years old (Ex. 292). To estimate the remaining occupational risk after vaccination, OSHA will assume a 96% vaccine efficacy rate instead of 92.7%, which is Merck's estimate of vaccine efficacy adjusted for the age distribution of those covered by this standard. Although the final estimate of 92.7% was provided to OSHA, Merck did not provide the basic data and underlying methodology based upon which this estimate was derived. Therefore, the Agency is unable to duplicate these results and determine the accuracy of these figures. By assuming 96%, OSHA may be underestimating the remaining risk to workers with occupational exposure. OSHA believes that administration of the hepatitis B vaccine will lead to a significant reduction in the HBV infection risk faced by workers with occupational exposure to blood or other potentially infectious material.

OSHA estimates that there are between 2,517,649 and 3,057,145 healthcare and other workers with occupational exposure who are both at risk for HBV and covered by this standard (see TABLE VII-4.) If all of these workers were vaccinated with a 96% effective hepatitis B vaccine, then over 45 years (a working lifetime under the Act), OSHA estimates that between 244,000 and 274,000 HBV infections would be prevented, between 61,000 and 68,500 cases of clinical illness would be prevented, and between 5,400 and 6,100 deaths would be prevented. The estimated number of HBV infections prevented is calculated as follows: the number of workers (at risk and covered by the standard) is multiplied by the lifetime occupational exposure risk (given in TABLE V-6) and by 0.96 to account for the vaccine efficacy. The estimated number of HBV infections and their outcomes which would be prevented by this provision are presented in TABLE V-7.

                         TABLE  V - 7
 INFECTIONS AND OUTCOMES PREVENTED IN HEALTHCARE WORKERS WITH 45 YEAR
     WORKING LIFETIME OF OCCUPATIONAL EXPOSURE TO BLOOD OR OTHER
 POTENTIALLY INFECTIOUS MATERIALS AFTER ADMINISTRATION OF HEPATITIS B
               VACCINE WITH 96% EFFICACY(a)
_____________________________________________________________________
                                               Number Prevented(b)
_____________________________________________________________________
HBV Infections                                   244,122 - 274,081
Clinical Illness (25%)                             61,031 - 68,520
Hospitalized (5%)                                  12,206 - 13,704
HBV Carrier (5% - 10%)                          12,206 - 27,408(c)
Chronic HBV (25% Carriers)                           3,052 - 6,852
Fulminant Death (.125%)                                  305 - 343
Death - Cirrhosis (1.7%)                             4,150 - 4,659
Death - PHC(b) (0.4%)                                  976 - 1,096
All Deaths (2.225%)                                  5,432 - 6,098
_____________________________________________________________________
  Footnote(a)  Numbers are calculated assuming that vaccine is given to
all workers with occupational exposure who are covered by this standard
and who have not been vaccinated or had a prior HBV infection. The
estimate of healthcare workers at risk and covered by the standard is
between 2,517,649 and 3,057,145 (see TABLE VII-4). Benefits are estimated
by applying 96% of the lifetime HBV risk attributable to occupational
exposure given in TABLE V-6 to the estimates of healthcare workers at risk
and covered by the standard.
  Footnote(b)  Risks for all exposed workers are estimated assuming 15%
and 30% had a previous infection and are thus immune.
  Footnote(c)  Smaller number assumes that 30% of the workers are immune
due to prior infection and 5% of the workers infected will become HBV
carriers. Larger number assumes that 15% of the workers are immune and 10%
of the workers infected will become HBV carriers.

TABLE V-8 presents the lifetime risk of HBV infection and its outcomes attributable to occupational exposure after administration of a 96% efficacious hepatitis B vaccine to employees at risk. TABLE V-8 shows that even if all employees are vaccinated and assuming the vaccine is 96% effective, the remaining risk of HBV infection to workers with occupational exposure is greater than 3 per thousand. After vaccination, the lifetime risk of HBV infection is between 3 and 5 per 1000, and the risk of clinical illness is approximately 1 per 1000. In constructing TABLE V-8, OSHA assumed that all workers at risk will agree to be vaccinated. However the record indicates that this has not been the case in the past and, although education of employees on the benefit of vaccination should increase acceptance, it is unlikely all workers will be willing to be vaccinated. OSHA witnesses testified that ". . . This [vaccination] program yielded about a 65 percent vaccination rate. . ." (Kathleen F. Gordon, M.S., Tr. 9/19/89, p. 9). ". . . The vaccination rates from 1982-1985 were 36-55 percent in the dental school and 23-47 percent in the medical school. The vaccination rate for the nursing school has always been below 10 percent which is attributed to lack of mandatory education, advocacy, and follow-up procedures. . ." (James A. Cottone, M.S., Tr. 9/19/89, p. 56). Dr. Joseph H. Coggin reported that the acceptance rate ranged from 20-25 percent primarily in blood laboratories, to 50 percent in the hospital because of a fairly active program, to 90 percent in laboratories where they work with hepatitis B or with HIV, to 100 percent where employees are required to be vaccinated if they want to work there (Joseph H. Coggin, Ph.D., Tr. 9/12/89, pp. 56-58). Dr. Campbell from The Baptist Medical Center testified that ". . . Eighty percent have accepted the offer and have been vaccinated. . . ." (Dr. L. L. Campbell, Tr. 9/19/89, p. 72). Angelica Corporation Health Services group reported approximately 40 percent of the employees at one plant accepted the vaccine, (Jill Witter, Esq., Tr. 9/18/89, p. 163). Baylor University Medical Center reported that their latest data, January through June of 1989, indicated an average acceptance rate of 92 percent, while the overall rate for the years since the vaccine has been offered was in the 70 percent range. (Dr. W. L. Sutker, Tr. 9/27/89, p. 14). The National Funeral Directors Association reported a 40 percent acceptance rate of their vaccination program among members, (T. Ryan, Tr. 9/27/89, p. 293). The Methodist Hospital of Dallas reported a 90 percent rate amongst those offered the vaccine, (Dr. J. A. Barnett, Tr. 9/27/89, p. 221).

                         TABLE  V - 8
  ESTIMATE OF HBV INFECTION AND ITS OUTCOMES AMONG HEALTHCARE WORKERS
    EXPOSED TO BLOOD OR OTHER POTENTIALLY INFECTIOUS MATERIAL AFTER
       ADMINISTRATION OF HEPATITIS B VACCINE WITH 96% EFFICACY(a)
_______________________________________________________________________
                                                    Lifetime
                                           Occupational Risk(b),(c)
_______________________________________________________________________
HBV Infections                                  3.3272 - 4.5360
Clinical Illness (25%)                          0.8318 - 1.1340
Hospitalized (5%)                               0.1664 - 0.2268
HBV Carrier (5% - 10%)                          0.1664 - 0.4536
Chronic HBV (25% Carriers)                      0.0416 - 0.1134
Fulminant Death (.125%)                         0.0042 - 0.0057
Death - Cirrhosis (1.7%)                        0.0566 - 0.0771
Death - PHC(d) (0.4%)                           0.0133 - 0.0181
All Deaths (2.225%)                             0.0740 - 0.1009
______________________________________________________________________
  Footnote(a)  Risks are expressed as the number of events per 1,000
exposed healthcare workers who lack immunity.
  Footnote(b)  Risks for exposed workers are estimated assuming 15% and
30% of the workers had a previous infection and are thus immune.
  Footnote(c)  Assumes 45 years of occupational exposure and is calculated
by multiplying the lifetime occupational risk from Table V-6 by 0.04.
  Footnote(d)  Primary Hepatocellular carcinoma.

In general, acceptance rates of various hepatitis B vaccination programs implemented throughout the country ranged from over 95 percent to below 10 percent. This clearly demonstrates that, even in the presence of a well organized and supported vaccination program, not everyone is willing to accept the vaccine. A descriptive analysis of the reported compliance rates revealed a distribution with a mean of 55.9, a median of 56.5, a first quartile of 40, and a third quartile of 73 percent. This indicates that, based on evidence in the record, three fourths of the vaccination programs in existence reported compliance rates less than 75 percent. In addition, OSHA estimated an average compliance rate of 50 percent (see Section VII). OSHA's average compliance rate was derived as a weighted average of data obtained from two surveys conducted by the Agency. Data from the surveys were collected for nineteen industry groupings (e.g., dentists' offices, dialysis centers, and home health care) and four occupational categories of employees. The four occupational categories represented were doctors, dentists and nurses (Category A); laboratory workers, emergency responders and fire fighters (Category B); housekeepers (Category C); and service workers (Category D). Vaccination acceptance rates were calculated for each occupational category found in a particular industry group. These acceptance rates were then weighted by the number of affected workers and yielded a 50 percent average acceptance rate. OSHA constructed a scenario where 50 percent of those offered the vaccine would actually agree to be vaccinated and estimated the remaining lifetime occupational risk assuming a 96% vaccine efficacy. These numbers are found in TABLE V-9. Under an assumption of 50 percent compliance to a vaccination program and 96 percent efficacy rate the remaining lifetime occupational risk is highly significant. A range of 43 to 60 HBV infections are expected to occur per one thousand exposed workers per working lifetime. This will result in 11 to 15 clinical illnesses and approximately one death per thousand exposed workers.

                         TABLE  V - 9
   ESTIMATE OF LIFETIME OCCUPATIONAL RISK AMONG HEALTHCARE WORKERS
     EXPOSED TO BLOOD OR OTHER POTENTIALLY INFECTIOUS MATERIAL(a)
____________________________________________________________________________
                                  Lifetime  Occupational Risk(b)
____________________________________________________________________________
                                100%              50%             0%
                            Vaccination       Vaccination      Vacination
                               Rate             Rate(c)           Rate
____________________________________________________________________________
HBV Infections             3.465 - 4.797    43.25 - 58.97   83.18 - 113.40
Clinical Illness (25%)     0.867 - 1.201    10.81 - 14.74    21.46 - 29.61
Hospitalized (5%)          0.174 - 0.240      2.16 - 2.95      4.33 - 5.99
HBV Carrier (5% - 10%)     0.174 - 0.481      2.16 - 5.90     4.33 - 11.95
Chronic HBV (25% Carriers) 0.043 - 0.120      0.54 - 1.47      1.08 - 3.00
Fulminant Death (.125%)    0.004 - 0.006      0.05 - 0.07      0.11 - 0.15
Death - Cirrhosis (1.7%)   0.059 - 0.082      0.74 - 1.00      1.47 - 2.04
Death - PHC(0.4%)          0.014 - 0.019      0.17 - 0.24      0.35 - 0.48
All Deaths (2.225%)        0.077 - 0.107      0.96 - 1.31      1.93 - 2.67
___________________________________________________________________________
  Footnote(a)  Risks are expressed as the number of events per 1,000
exposed healthcare workers who lack immunity.
  Footnote(b)  Risks for exposed workers are estimated assuming 15% and
30% of the workers had a previous infection and are thus immune.
  Footnote(c)  Assumes 45 years of occupational exposure and is calculated
by multiplying the lifetime occupational risk from TABLE V-6 by 0.52
(0.50*0.04+.50=0.52).

In reviewing the record, a number of commenters expressed concern over certain aspects of the OSHA risk assessment. Specifically, two chapters of the Association for Practitioners in Infection Control (APIC), Dade County and Greater Omaha, the Joint Committee on Health Care Laundry Guidelines and the Presbyterian-University Hospital of Pittsburgh, among others, argued that without disease incidence data specific to particular occupations it is impossible for OSHA to make an accurate determination of the risk of HBV infection in an occupational setting (Exs. 20-371; 20-943; 20-1113; 20-1101). Further, they assert that the use of sero-prevalence data significantly overestimates current incidence of HBV infections in hospital personnel in most settings. Presbyterian-University Hospital stated that ". . . More appropriate data would be provided by real incidence data that control for non-occupational etiologies which can significantly confound data like that used [by OSHA] to arrive at this occupational risk assessment" (Ex. 20-1101). Further, Dr. W. L. Sutker of Baylor University Medical Center, in his testimony, disputed the validity of using an estimate of the number of healthcare workers infected with HBV as opposed to actual incidence data, even though he admitted he did not have a way to prove or disprove the accuracy of CDC's estimates which were used by OSHA (W. L. Sutker, M.D., Tr. 9/27/89, pp. 103-105). OSHA agrees that the use of incidence data adjusted for non-occupational etiologies would be ideal to use in determining the actual risk of HBV infection attributable to occupational exposure. However, actual incidence data simply do not exist at a national level. A limited amount of incidence data was submitted to the record, specific to certain hospitals or certain regions (Exs. 272; 20-1366). Regional data such as those cited above as well as those data presented in testimony by Baylor University Medical Center and the Presbyterian Health Care system can not be used for OSHA's risk assessment (W. L. Sutker, M.D., Tr. 9/27/89, pp. 102-122). While they are appropriate to show incidence of infection for an institution or a region, they are not appropriate for OSHA's risk assessment. OSHA's goal is to estimate the nationwide risk of HBV infection to workers with occupational exposure from national incidence rates of hepatitis B in healthcare workers. To do so, OSHA needs to rely on data that are not affected by regional differences and are adjusted for non-occupational etiologies. It would be inaccurate to use data from a low-incidence region to determine an overall occupational risk just as it would be inaccurate to use data from a high-incidence region. In our attempt to estimate the risk of infection OSHA needs data from a representative sample of the nation as a whole. The Centers for Disease Control is the only source of reliable national estimates of the numbers needed for OSHA's risk assessments. When faced with a choice of using regional data that are not adjusted for regional differences or data derived from non-representative samples (these are samples selected subjectively) as opposed to data estimated from national surveys, such as CDC's estimates, OSHA believes the only reasonable approach is to use the estimates provided by CDC. These estimates have been adjusted for factors such as under-reporting and are less affected by regional factors.

Another issue of concern to several commenters was OSHA's use of 45 years to estimate lifetime occupational risk. Dr. Sutker of Baylor University Medical Center believes that OSHA has overestimated the risk by a factor of 11 based on Baylor's historical data demonstrating that the average occupational tenure of a healthcare worker at Baylor is approximately 4 years (W. L. Sutker, M.D., Tr, 9/27/89, p. 21). Likewise, Dr. Goodman of the Presbyterian Hospital of Dallas argued in his testimony that OSHA's extrapolated figures of a lifetime occupational risk are probably inaccurate by a factor of at least 13, which he asserted vastly overestimated the true medical risk based on Presbyterian Hospital's historical figures indicating that the average tenure of a healthcare worker at Presbyterian is approximately 3.3 years (E. L. Goodman, M.D., Tr. 9/27/89, pp. 142-143). Turnover at an individual hospital is not the same as turnover in a particular occupation whether it be at the support staff level, among paraprofessionals or among professionals. A physician, for example, may leave a particular hospital, but would likely not leave the healthcare field. In any case, section 6(b)(5) of the OSH Act mandates that the Secretary "... 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 standard for the period of his working life." For the purposes of OSHA health standards, this working lifetime is 45 years (Asbestos, 51 FR 22612), (Benzene, 52 FR 34460), (Ethylene Oxide, 53 FR 11414). The record contains no evidence that would indicate that another time period for a working lifetime for healthcare employees would be more appropriate.

Another point of concern to several commenters was OSHA's methodology for estimating the risk of HBV infection. Dr. Sutker, representing Baylor University Medical Center and the Presbyterian Health Care system, suggested that ". . . [a] better method of accurately assessing healthcare worker risk to hepatitis B virus, is by utilizing officially reported probability figures . . . [to estimate the risk of infection due to a single needlestick exposure in a hospital setting] . . . " (W. L. Sutker, M.D., Tr. 9/27/89, pp. 11-14). The same argument was repeated by Dr. Goodman of the Presbyterian Hospital of Dallas in his testimony and in a post-hearing comment (E. L. Goodman, M.D., Tr. 9/27/89, pp. 139-140; Ex. 272). OSHA has considered Dr. Goodman's recommendation, but has concluded that his approach does not provide an accurate estimate of the risk of HBV infection because it completely disregards non-percutaneous exposures and exposures in non-hospital settings. Given the fact that healthcare workers may also be infected by coming in contact through breaks in the skin, and contact with mucous membranes as with needlesticks, measuring the HBV infection risk purely by percutaneous exposures will result in an underestimate of the true risk. OSHA's estimate of risk applies to employees with occupational exposure to blood or other potentially infectious materials regardless of occupational environment, whether it is a hospital, a dental office, or a funeral home and the estimate of HBV risk from a single incident does not vary with frequency or route of exposure or infectivity of the source individual. What the Agency has attempted to do is estimate an overall risk to HBV infection from occupational exposure regardless of the environment and not restrict its estimate to the risk from a percutaneous exposure from any single needlestick in a hospital setting.

In the preamble to the proposed standard, OSHA outlined its approach to the quantitative estimate of risk from exposure to HBV, including the selection of data sources and methodology used. On the basis of the preliminary quantitative risk assessment, OSHA concluded that the lifetime occupational risk from exposure to HBV was 75 to 119 HBV infections and 2 to 3 deaths per 1000 exposed healthcare workers who lack immunity. This figure was used to support OSHA's finding that exposure to bloodborne pathogens, and specifically HBV, represented a significant risk to workers exposed to blood or other potentially infectious materials. In this final risk assessment OSHA estimates that the lifetime occupational risk from exposure to HBV is 83 to 113 HBV infections and 2 to 3 deaths per 1000 exposed healthcare workers who are not immune. Even though the Agency used the same methodology to estimate the lifetime occupational risk as in the preamble, the number of expected infections changed in the final risk assessment for reasons outlined below. For its final risk assessment OSHA used the latest available data in the record. Based on CDC's 1988 reported data, the number of HBV infections in persons whose only source of infection was related to healthcare employment was reduced from 12,000 to 8,700. In addition, the number of healthcare workers with occupational exposure was reduced from 5.3 million to approximately 4.9 million. The reduction in the number of HBV infections in the later years is primarily due to the introduction of the hepatitis B vaccine. OSHA used 1988 Census population values instead of the 1985 Census figures used in the proposal (this changed the number of U.S. adults from 180 million to 193 million). In addition, the Agency used estimates of the number of HBV infections in U.S. adults and healthcare workers with occupational exposure based on the 1988 national hepatitis surveillance data instead of 1987 data used in the proposal (this changed the number of infections from 300,000 to 280,000). The above changes affected the estimates of the annual HBV infection rate and the lifetime occupational risk per 1000 healthcare workers; the numbers changed from 3.50-4.56 to 3.47-4.21 and from 75.38-118.54 to 83.18-113.40, respectively. Finally, in the proposal, 90 percent was used as an estimate of the vaccine's efficacy rate, whereas in the final, based on the latest data submitted to the record, the vaccine efficacy rate was estimated to be 96 percent.

D. Qualitative Assessment of HIV Risk

The CDC estimates that there are between 1 million and 1.5 million HIV-infected persons in the U.S. (Ex. 6-356). As of September 30, 1990 occupational information was available for 122,159 of the AIDS cases reported to CDC. Of these, 5,815 or 4.8% were identified as healthcare workers (Ex. L6-666) (note: Exhibit L6-666 is an updated version of Exhibit 6-378.). This proportion is similar to the proportion of the labor force employed in the healthcare field.

Most healthcare workers with AIDS also belong to some other group which places them at high risk for HIV infection (e.g. homosexual men, intravenous drug users, etc.). There is, however, a statistically significantly larger proportion of healthcare workers with no known risk factors (6%), than the proportion of other AIDS cases (i.e. individuals with AIDS not in the healthcare field) with no known risk factors (3%). As of September 30 1990, there were 337 reported cases of healthcare workers with AIDS with no known risk factors. These cases are being studied further. CDC reports that 69 could not be assigned to a risk group after follow-up, 65 had either died or refused to be interviewed, and 203 were still under investigation (Ex. L6-666).

Because the prevalence of HIV infection among healthcare and other workers with occupational exposure to blood or other potentially infectious material is unknown, it is not possible to estimate an "observed" infection rate. Therefore, it is not possible to quantify the risk as was done for occupational exposure to HBV. Certain deductions, however, can be made. It is known that the virus is present only in blood or certain body fluids and that exposure to these fluids from an HIV-infected person puts one at risk for HIV-infection. Therefore, workers who have occupational exposure to blood or certain body fluids are at risk.

No case of infection due to casual contact with these fluids has been documented. Rather, infection can occur only if infectious fluids enter the body either through a percutaneous or mucosal route, although exposure by either of these routes does not mean that infection will occur. In several prospective studies of healthcare workers with HIV exposures, seroconversions have been observed. Although the rate of infection is low, it is not insignificant.

The most recent report from the CDC Cooperative Needlestick Surveillance Group authored by Marcus and colleagues shows that of 860 healthcare workers with an exposure to HIV-infected blood through needlestick or cut from sharp instruments, 4 workers became infected with the virus yielding a seroprevalence rate of 4/860 = 0.47 (Ex. 6-372). One of the four was first tested for HIV antibody 10 months after sustaining a needlestick exposure to blood of an HIV-infected patient. As there was no available acute blood specimen collected within 30 days after exposure this case cannot by definition be considered a seroconversion. The remaining 3 HIV-seropositive subjects had HIV-seronegative acute blood specimens and were thus considered seroconversions, yielding a seroconversion rate of 3/860 = 3.5 per 1000 exposures to infected blood through needlestick or cut. Gerberding et al. recently reported that of 180 workers with 215 exposures to HIV-infected blood through needlesticks, 1 worker became infected with the virus (Ex. 6-375). This leads to a seroconversion rate of 4.7 per 1000 exposures to infected blood through needlestick. The HIV infection rates reported by both of these studies are very close.

Both the CDC and the Gerberding et al. studies provide estimates of the risk of infection given parenteral exposure to HIV-infected blood. Neither study, however, provides estimates of the risk of all occupational exposure. One approach to this problem has been suggested by Wormser et al. who estimated the probability of HIV infection in terms of HIV-infected patient-days for hospital staff caring for HIV-infected patients (Ex. 6-388). For the 18 month period from January of 1986 to June of 1987, the authors observed a needlestick rate of 1.9 per 1000 HIV-infected patient-days among staff caring for HIV-infected patients (Ex. 6-388). This rate was substantially lower than the needlestick rates of 4.3 and 4.6 per 1000 HIV-infected patient-days reported at the hospital for 1985 and 1984, respectively (Ex. 6-388).

Using the observed rate of 1.9 needlesticks per 1000 HIV-infected patient-days, Wormser et al. estimated the expected number of needlesticks for different numbers of HIV-infected patient-days. For example, for 45,000 HIV-infected patient-days, (750 HIV-infected patients hospitalized for 60 days, 1500 HIV-infected patients hospitalized for 30 days, etc.), the expected number of needlesticks is 86 (1.9/1000 x 45,000) (Ex. 6- 388). Wormser et al. then estimated the probability of at least one exposed worker becoming infected with HIV as 1 - [(1-p)n], where n is the number of needlesticks and p is the probability of becoming infected with HIV given needlestick exposure to HIV-infected blood, which the authors assumed to be 0.0035 (Ex. 6-388). The estimated probabilities, which are expressed per expected number of needlesticks or per HIV-infected patient-days, are presented in Table 1 (Ex. 6-388). In addition, OSHA has calculated these probabilities using Gerberding et al.'s estimate of 4.7 infections per 1000 needlestick exposures to HIV-infected blood and has included them in TABLE V-10.

In reviewing TABLE V-10, it is important to remember that the probabilities presented there do not represent an estimate of the number of exposed workers who will become infected with HIV. "Number of workers exposed" is not used in any of the calculations, and therefore an expected number of infections per some number of workers caring for HIV-infected patients can not be calculated. One worker may experience more than one needlestick. The probabilities in TABLE V-10 depend only upon the number of needlesticks which, in turn, depends only upon the number of HIV-infected patient-days and the assumption that needlesticks occur at a rate of 1.9 per 1000 HIV-infected patient-days.

                           TABLE  V - 10
                PROBABILITY OF AT LEAST ONE INFECTION
        DUE TO NEEDLESTICK EXPOSURE TO HIV-INFECTED BLOOD(a)

_________________________________________________________________
                                              Probability of
                     Estimated             At Least 1 Infection
 HIV-Infected       Number of         ___________________________
 Patient-Days       Needlesticks(b)   Wormser(c)   Gerberding(d)
__________________________________________________________________
     5,000                10             .03               .05
    20,000                38             .12               .16
    45,000                86             .26               .33
   105,000               200             .50               .61
   200,000               380             .74               .83
   440,000               836             .95               .98
___________________________________________________________________
  Footnote(a) Probabilities are binomial (n,p) and calculated as one
minus the probability of no infections.
  Footnote(b) Number of needlesticks is calculated based on the estimated
rate of 1.9 needlesticks per 1000 HIV-infected patient-days.
  Footnote(c) From Wormser et al., (Ex. 6-388). Assumes that the
probability of infection give needlestick exposure to HIV-infected blood
is 3.5 per 1000 exposures.
  Footnote(d) Probabilities calculated by OSHA using Gerberding et al.'s
estimate of 4.7 infections per 1000 needlestick exposures to HIV-infected
blood (Ex. 6-375)

TABLE V-10 shows that the probability of HIV infection for at least one healthcare worker caring for HIV-infected patients does not increase linearly as the number of HIV-infected patient-days increases. A ten-fold increase in HIV-infected patient-days from 20,000 to 200,000 leads only to a six-fold increase in the probability of at least one infection. If one were to assume that the needlestick rate were two times higher than the rate used in TABLE V-10 (i.e. 3.8 needlesticks per 1000 HIV-infected patient-days instead of 1.9 needlesticks per 1000 HIV-infected patient-days), the probability of at least one infection doubles at 5000 HIV-infected patient-days but increases only 6% at 440,000 HIV-infected patient-days. If one were to assume that the needlestick rate were half as high as the rate used in TABLE V-10, (i.e. .95 needlesticks per 1000 HIV-infected patient-days instead of 1.9 needlesticks per 1000 HIV-infected patient-days), the probability of at least one infection is one-third smaller at 5000 HIV-infected patient-days but only one-fifth smaller at 440,000 HIV-infected patient-days . This approach to estimating the risk of HIV infection would apply only to staff caring for HIV-infected patients because Wormser et al. used a needlestick rate per HIV-infected patient-days which was estimated from this population.

Clearly, reducing the risk of needlestick will reduce the probability of HIV infection. CDC reported that of 1,201 exposures to HIV-infected blood through needlesticks, cuts with sharp objects, contamination of open wounds, or contamination of mucous membrane, 37% of the exposures might have been prevented if recommended infection control precautions had been followed (Ex. 6-372). Recapping of needles by hand accounted for 17% of the 1,201 exposures, improper disposal of used needles or sharp objects accounted for 14%, and contamination of open wounds accounted for 6% (Ex. 6-372).

A study of needlestick injuries among hospital personnel by Jagger et al. found that the risk of injury depended upon the type of device used and that devices requiring disassembly had the highest risks (Ex.6-350). Jagger investigated 326 needlestick injuries over a 10 month period and found that 17% occurred during use of the device, and 13% occurred during or after disposal of the devices. The majority (70%), however, occurred after use but before disposal of the devices (Ex. 6-350). The single largest cause of injury was due to recapping. Workers missed the cap and stabbed themselves when attempting to cover a used needle in 17.8% of the injuries (Ex. 6-350). Other major causes of injury were needles piercing caps when recapped after use (12.3%), contacting needles on exposed surfaces after use (10.7%), and needles protruding from trash (8.9%) (Ex. 6-380). The largest number of injuries was associated with disposable syringes, but when the injury rate for various devices was adjusted for the number of each type of device purchased, disposable syringes had the lowest accident rate at 6.9 per 100,000 purchased (Ex. 6-350). All of the devices requiring disassembly had higher accident rates ranging from 8.3 per 100,000 purchased for prefilled cartridge injection syringes to 36.7 per 100,000 purchased for intravenous tubing and needle assemblies (Ex. 6-380).

While most of the epidemiological investigations have concentrated on assessing the risk of HIV infection to healthcare workers exposed to HIV-infected blood through needlesticks or cuts with sharp objects, there is evidence that workers in research and production laboratories routinely exposed to high concentrations of the virus are also at risk of infection. Weiss et al. prospectively studied 265 laboratory and affiliated workers and found one worker infected with the same strain of HIV as was used in the laboratory (Ex. 6-187). The infected worker reported occurrences of HIV contamination in the work area but could not recall any episode of direct skin exposure with the virus and denied any parenteral exposures. The worker reported that double gloves were worn whenever there were bandaged cuts on fingers or hands. An episode of nonspecific dermatitis on the arm was recalled, but the affected area was always covered by a cloth laboratory gown. There was no contact of potentially infectious material with these areas as has been reported for healthcare workers infected after clinical exposure to HIV-infected fluids (see Case Reports in the discussion of HIV health effects). For 99 workers who shared a work environment involving exposure to concentrated virus, the authors estimated the HIV infection rate to be .48 per 100 person-years with a 95% upper confidence limit of 2.30 infections per 100 person-years of exposure (Ex. 6-187). OSHA estimates that over a 45 year working lifetime, the HIV infection risk would be 195 per 1000 exposed workers in research and production laboratories. The lifetime risk is estimated by using {1-(1-p)45} where p is .0048.

Weiss et al. also reported a second incident of HIV infection in a research laboratory worker who was employed in the production of concentrated virus and who was cut on the hand with a potentially contaminated stainless steel needle used for cleaning an apparatus. The worker was not part of the Weiss et al. cohort, and it is not yet known whether the virus which infected this worker is the same (i.e. genetically identical) as was found in the laboratory. Weiss et al. noted that although the infected workers were careful, neither was fully conversant with or strictly adhered to biosafety guidelines in day to day procedures at all times. Weiss et al. concluded that "infection in the laboratory workers took place under prescribed Biosafety Level 3 containment suggests the need to review carefully all operations involving highly concentrated infectious material and to ensure proficiency in the conduct of recommended safeguards." (Ex. 6-187).

Although it is not possible to quantify the risk of HIV infection in healthcare or other workers with occupational exposure to blood or other potentially infectious material or with direct exposure to the virus itself, the data show that a risk does exist. As the number of people with HIV-associated illnesses increases, the probability that workers exposed to blood or other potentially infectious material will also be exposed to HIV also increases. Given needlestick exposure to HIV-infected blood, the risk of seroconversion is estimated to be between 3.5 and 4.7 per 1000 exposures. For research and production laboratory workers with occupational exposure to high concentrations of the virus, the risk of seroconversion is estimated to be 4.8 per 1000 person-years. Over a 45 year working lifetime, the risk would be 195 per 1000 exposed workers. By reducing the risk of exposure to blood and other potentially infectious material and by strictly adhering to biosafety procedures in handling the virus in laboratories, the risk of HIV infection can be reduced.

As described in the health effects discussions, there are other bloodborne pathogens, such as syphilis and malaria, which are present in blood during certain phases of infection. During these phases, the blood of infected individuals poses a risk to exposed workers. Although the risk of these infections has not been quantified, it does exist and will be minimized or eliminated by preventing occupational exposure to blood.

[56 FR 64004, Dec, 6, 1991; 57 FR 29206, July 1, 1992]

Regulations (Preambles to Final Rules) - Table of Contents Regulations (Preambles to Final Rules) - Table of Contents