Methods

Top Methods Results Discussion Author & Article Info References

Selection of Blood Centers and Blood Collection Sites

We selected 6 of the 20 blood centers enrolled in the ongoing interview study of HIV-seropositive blood donors to participate in the related health care worker donor study: the American National Red Cross regions of Atlanta, Greater Chesapeake and Potomac, Los Angeles-Orange County, Pennsylvania/New Jersey (catchment did not include Newark), and Southeastern Michigan; and Irwin Memorial Blood Centers in San Francisco. Their selection was guided in part by the estimated level of HIV infection in the catchment population. A higher HIV seroprevalence rate theoretically would increase a health care worker's opportunity for occupational blood exposure to patients with HIV infection. We used the annual rate of reported acquired immunodeficiency syndrome (AIDS) cases per 100 000 persons for the major metropolitan area served by each of the centers to approximate HIV infection in the population. For the metropolitan areas served by the 20 centers, the reported AIDS case rate during 1991 ranged from 5.9 to 122.5 per 100 000 persons; for the 6 centers it ranged from 9.3 to 122.5 per 100 000 persons [3]. Four of the 6 centers were in metropolitan areas with case rates greater than 24 per 100 000 persons. In comparison, the 1991 AIDS case rate for the United States ranged from 4.9 per 100 000 persons for nonmetropolitan areas to 25.3 per 100 000 persons for metropolitan areas with populations of 500 000 or more persons. In addition, data from seroprevalence surveys in settings such as sexually transmitted disease clinics and drug treatment centers suggested that the catchment areas for the 6 blood centers included urban areas with increased rates of HIV infection [4]. We could not include several centers with AIDS case rates higher than those of the 6 centers because distribution and collection of the study questionnaires was logistically impossible or would have severely disrupted routine blood collection procedures.

Although health care workers can donate blood at any blood collection site, we expected that they would comprise most of the donors at medical facilities. Accordingly, we included only donors from hospitals and other medical facilities (that is, medical or nursing schools, skilled nursing facilities, and laboratories) that served as blood collection sites from March 1990 through August 1991 in the health care worker donor study.

HIV Blood Donor Screening Procedures

Routine blood donation procedures were followed at all blood centers [1, 2]. Briefly, information regarding AIDS, the routes of HIV infection, and the reasons persons should refrain from donation were provided to each prospective blood donor; each was told that he or she would be screened for antibody to HIV. The donor was questioned directly about medical symptoms and possible HIV risk factors, including injection drug use, transfusion, sexual contact with a person with HIV, and travel outside the United States. Depending on the responses to these questions, the prospective donor could be deferred from donation at this point. Finally, the donor was given a card detailing risk factors for HIV infection and was instructed to place one of two bar-coded labels on the card to indicate whether the blood should or should not be used for transfusion (that is, confidential unit exclusion). At each of the hospital or medical facility blood drives conducted at the six centers, blood center personnel tabulated the number of all deferrals due to exposure to HIV. However, they included exposures other than occupational (for example, sexual contact with an at-risk partner) in the tabulation. At two of the six centers, health care worker donors were asked routinely about possible occupational exposures to blood through percutaneous injuries and mucous membrane or cutaneous contact. At the remaining four centers, health care worker donors were asked about such exposures rarely or inconsistently.

Estimation of HIV Incidence and Prevalence Rates

Six Blood Centers

The numerator and denominator for HIV prevalence and incidence rates were measured directly at the six centers. We derived the numerator (that is, the number of seropositive health care worker donors who donated at a targeted hospital or medical facility) from the ongoing interview study at the six centers. Seropositive donors were invited to participate in a formal interviewer-directed, 45-minute interview that solicited demographic information, risk factors for HIV infection, and motives for donation. Interviewed donors who reported that they worked in a hospital, laboratory, or other medical facility in the last 10 years were also questioned about possible occupational exposures. We did not include persons who indicated that their jobs were administrative, clerical, or secretarial in the numerator because of the very low likelihood of occupational contact with blood.

We derived the denominator from a brief, anonymous, supplemental questionnaire distributed to donors at hospitals or other medical facilities. This questionnaire was necessary because occupational information was not collected during blood donation. To help ensure that an individual donor did not complete multiple questionnaires, the questionnaires were distributed at only one blood drive at each hospital or medical facility. We included in the denominator blood donors who reported that they had worked in a hospital, laboratory, or other medical facility in the last 10 years and who did not have administrative, clerical, or secretarial jobs. Because the voluntary supplemental questionnaire about work history in a health care setting was anonymous, we could not link an individual questionnaire with HIV antibody test results.

To estimate the incidence of HIV infection among health care worker donors in the six centers, we asked questionnaire respondents to indicate the month and year of their last previous blood donation. To assess the validity of this information before initiating the full-scale study, we did a pilot study in one blood center, during which we verified the donors' self-reported dates of donation by checking them against written blood bank records. More than 90% of donors correctly reported their most recent previous donation, which was not surprising because donors are given a wallet-size card that documents the dates of their donations. Person-years for an individual donor were calculated in the interview study and in the supplemental questionnaire survey as the number of years between the current and most recent previous donation. We excluded from the analysis persons whose previous donation was before March 1985, when screening for HIV antibody was instituted by blood banks. We used blood bank records to verify previous donation dates for any repeatedly seropositive health care worker donors who were interviewed.

Twenty Blood Centers

We measured directly the numerator for HIV prevalence from the ongoing study at the 20 centers. We included in the numerator all interviewed persons who were HIV seropositive and had donated blood at a hospital in the catchment area served by the 20 blood centers and who indicated during their interview that they had worked in a health care setting in the previous 10 years but not in administrative, clerical, or secretarial jobs.

We estimated the denominator (that is, the total number of health care worker donors tested for HIV) not directly but as follows. During the study period, blood was collected from 3 339 079 persons at the 20 centers. By reviewing blood collection logs, these centers estimated that 1.7% of their donations were collected from hospitals (estimates regarding the number of donations from other medical facilities were not available). Therefore, during the study period, 56 764 (0.017 x 3 339 079) persons donated blood at a hospital blood drive at the 20 centers. Using data from the questionnaire survey distributed at the 6 centers, we estimated how many of these 56 764 persons were health care workers. The proportion of blood donors who identified themselves as health care workers at the 6 centers ranged from 60% to 74% (mean, 64%); for 5 of the 6 centers, this proportion ranged from 60% to 65%. Accordingly, we derived the estimated number of health care worker donors at the 20 centers by multiplying 56 764 by 0.64; we calculated upper and lower estimates for the denominator by multiplying 56 764 by 0.65 and 0.60, respectively.

Results

Top Methods Results Discussion Author & Article Info References

HIV Seroprevalence and Incidence at Six Blood Centers

From March 1990 through August 1991, 11 829 (88.5%) of 13 365 persons who donated blood and completed the supplemental questionnaire at 272 hospitals or other medical facilities where the 6 centers had collected blood indicated that they had worked in a hospital, laboratory, or other medical facility in the previous 10 years. After we excluded 3310 donors who had administrative, clerical, or secretarial jobs, we classified 8519 persons as health care worker donors. The number of health care worker donors enrolled by the blood centers ranged from 263 to 3394. Most (68%) health care worker donors were women. The median age for health care worker donors was 34 years (range, 17 to 79 years); 85.2% were white; 9.4% were black; 2% were Hispanic; 1.8% were Asian/Pacific Islander; and 1.6% were other or unknown.

Health care worker donors were asked to indicate their most recent occupation as a health care worker (Table 1). Overall, 60.5% of the 8519 health care worker donors had worked for 5 or more years in their most recent occupation. Of those health care worker donors most likely to have contact with blood and other body fluids (that is, nurse or nurse's aide, physician or dentist, emergency medical technician, and laboratory technician), 54.7% to 72.8% had worked for 5 or more years in their respective occupations (Table 2). Of note, 23.4% of health care worker donors reported their occupation as "other," which included volunteers, pharmacists, and food service workers. Occupational contact with patients' blood presumably would have been rare for many of these workers.

Information regarding the nature and frequency of occupational contact with blood and other body fluids was not collected for health care worker donors because of the high probability of recall bias. As a partial surrogate for exposure data, workers were asked to specify one or more primary work locations. As expected, physicians, nurses or nurse's aides, and laboratory technicians worked in those areas of the hospital likely to involve provision of direct patient care or laboratory services (Table 2).

We reviewed data from the interview study of seropositive donors at these six centers. From March 1990 through August 1991, we identified from two of the blood centers three HIV-seropositive persons who had donated blood at three different hospital or medical facilities. All were men; two (from different centers) were interviewed. One was a psychiatric nurse who had rare nonpercutaneous exposures to body fluids other than blood and reported having sex with men. The other worked as a medical supply technician in a stockroom and reported having sexual contact with a female injection drug user but no occupational blood contact. We could not determine the occupation of the third donor because he never responded to the blood center's notification and request that he return for test results. Thus, the prevalence of HIV infection among all health care worker donors in the six centers ranged from 0.02% to 0.04% (2 or 3 of 8519) (upper limit of 95% CI, 0.10% for 3 of 8519 persons), depending on the occupation of the third seropositive donor. In comparison, the prevalence of HIV infection among all first-time donors (that is, persons who had not been screened previously for HIV antibodies and would be expected to have a higher rate of HIV infection compared with repeat donors) ranged from 0.011% to 0.061% at the six centers during the study period.

Most (90.2%) health care worker donors had donated blood previously. Of the 8519 health care worker donors, 5936 (69.7%) provided the month and year of their most recent previous donation since March 1985, for a total of 4874 person-years. The average interval between donations was 9.9 months. One of the three donors with HIV infection, whose occupation was also unknown, was a first-time donor and excluded from analysis. Another infected donor previously donated blood in 1982 and was also excluded from the analysis. Thus, the estimated overall incidence of HIV infection among all health care worker donors was 0.021 per 100 person-years.

HIV Seroprevalence in 20 Blood Centers

By extrapolating from the survey of the 6 blood centers, we estimated that from March 1990 through August 1991 approximately 36 329 (range, 34 058 to 36 897 persons) health care worker donors were tested for HIV infection during hospital blood drives conducted by these 20 blood centers. During this same time, 20 HIV-seropositive persons were identified from the interview study who had donated blood at a hospital and who indicated during their interview that they had worked in a health care setting in the previous 10 years in a nonadministrative, clerical, or secretarial job. The number of HIV-seropositive health care worker donors ranged from 0 to 7 per center. Seven additional seropositive donors who had donated at a hospital could not be located or did not consent to be interviewed. Thus, the estimated overall seroprevalence of HIV infection among health care workers who donated blood at hospitals in the 20 centers ranged from 0.06% to 0.07% (20 to 27 of 36 329 persons) (upper limit of CI, 0.1% for 27 of 36 329 persons), depending on the occupation of the 7 blood donors who were not interviewed.

Of the 20 interviewed seropositive health care worker donors, 15 (75%) were men and 5 (25%) were women. Eleven (55.5%) of the 20 reported a nonoccupational risk for HIV infection: Eight men reported sexual contact with another man, one man reported sexual contact with a female injection drug user, and one woman each reported having sex with a man who was an injection drug user or having sex with a man who was infected with HIV.

Excluding the 11 health care worker donors who reported other risks, the estimated prevalence of HIV infection in the 20 centers was 0.03% (9 of 36 329 persons) (upper limit of CI, 0.05%). Of the 9 remaining health care worker donors with no identified risk, 1 was a female dietitian; 1 was a female housekeeper; 2 were male maintenance workers, both of whom reported working on AIDS wards but with no occupational exposure to blood or body fluids; 1 was a male clinical laboratory technician who did not report any occupational exposures; 1 was a male substance abuse counselor; and 3 reported occupational exposure(s). The first was a female licensed practical nurse who recalled sustaining several needlestick injuries involving hospitalized geriatric patients whose HIV infection status was unknown. The second, a male nurse, reported a cut from a razor blade that had been used on an HIV-infected patient and many mucous membrane and cutaneous contacts with blood while working in two hospitals in areas with high incidences of HIV infection. The third health care worker donor worked in a hospital morgue in an area with a high incidence of HIV infection and reported many cuts and lacerations from scalpels and other sharp objects, some of which had been used on persons infected with HIV.

Discussion

Top Methods Results Discussion Author & Article Info References

The true extent of occupationally acquired HIV infection among health care workers is difficult to estimate. National systems of AIDS [5] and HIV [6] surveillance are important mechanisms for identifying occupationally acquired HIV infection among health care workers. However, this depends on collection of accurate, self-reported data after diagnosis. We need additional methods to supplement existing surveillance programs, particularly to detect previously unreported or unsuspected cases of HIV infection in health care workers.

One approach has been to conduct HIV seroprevalence surveys of selected groups of health care workers, such as dentists and surgeons [7-9]. Military personnel and blood donors can serve as broader, population-based groups to assess both HIV prevalence and incidence. Although the attempted exclusion of persons with known HIV infection or related risk behaviors in these groups can reduce the prevalence of HIV infection, these same features can also highlight new or less frequent patterns of HIV transmission, such as occupationally acquired infection. Surveillance of U.S. Army active duty and reserve personnel has not shown an increased prevalence of HIV infection among medical personnel that can be attributed convincingly to occupational acquisition [10, 11]. Among physicians and dentists in the U.S. Army Reserves, the prevalence of HIV infection (3 of 3347 or 0.09%) was similar to that for nonmedical personnel (1007 of 659 361 or 0.15%) [11]. Army investigators did not collect information about nonoccupational risk factors for HIV infection.

The findings of our study suggest that, among health care workers who donate blood, HIV infection attributable to occupational exposure is uncommon. Overall prevalence rates of HIV infection (including those persons with nonoccupational risks) ranged from 0.02 per 100 health care worker donors who donated in the 6 blood centers to an estimated 0.07 per 100 in all 20 centers. These rates are similar to the overall rate of HIV seropositivity for first-time donations collected at the 6 centers. Furthermore, we determined through follow-up interviews with the seropositive health care workers in these centers that many had nonoccupational risks for HIV infection or worked in settings or occupations with limited opportunities for exposure to patients' blood and body fluids.

Our approach has some potential limitations. Health care workers who donate blood at a hospital or other medical facility may not be representative of health care workers who donate blood at other sites. Infection with HIV may have been overestimated among health care worker donors in this study because many of the blood centers were located in areas with high incidences of AIDS. Alternatively, HIV infection may have been underestimated if those health care workers who knew that they had occupationally acquired HIV infection or who had sustained a recent occupational HIV exposure chose not to donate blood or were deferred. In the six centers, 81 potential blood donors were deferred because of exposure to HIV. However, we could definitively ascribe only 10 of these 81 deferrals to occupational blood contact. Finally, interpretation of the data is limited because the nature and frequency of occupational blood contact could only be inferred indirectly from occupational title, length of occupation, and work location.

Information about the extent of and risk for occupationally acquired HIV infection among health care workers must necessarily be derived from various data sources, each with its inherent strengths and limitations. Health care workers who donate blood can serve as one sentinel population to help evaluate this risk. The observation of a low rate of previously undetected HIV infection in this group presumably due to occupational exposure is reassuring and also suggests that health care worker donors do not pose an increased risk for transmitting HIV through blood donation compared with other donors. Nonetheless, efforts to prevent occupational blood contact, particularly percutaneous injuries, must continue [12].