Submitted by:

University of Washington

Occupational and Environmental Medicine Program

325 Ninth Avenue, Box 359739

Seattle, WA 98104

October 1, 1997

Authors

Scott Barnhart, MD, MPH Principal Investigator

Tim Takaro, MD, MPH, MS Co-Principal Investigator

Bert Stover, BA

Kate Durand, MHS, CIH

Bill Trejo, BS

Chris Mack, MS

Kathy Ertell, MS, CIH

List of Tables iv

List of Figures v

Executive Summary vi

I. Introduction 1

A. Human Subjects 3

2. Population Identification - Available Databases 3
1. Pending Databases 3
2. Assembly of Master Database 5
3. Estimation of Mortality 6
4. Estimation of Exposure 6
5. Estimate of Need of Medical Surveillance 10
6. Locating Former Workers 12
7. Pilot Mailing 14

III. Identification of the Population of Former Hanford Workers 16

1. Estimated Size of the Entire Population 16
2. Demographics 17
3. Mortality Estimates 17

4. Exposure Estimation 22

1. Hazard Characterization 22
2. Job-Exposure Matrix 24

V. Justification of the Need for Medical Surveillance 31

1. Estimating Need for Medical Surveillance
2. Rationale: Interventions to Alter Course of Disease 31
3. Rationale: Interventions Which Could Lead to 31

4. Asbestos 32
1. Noise 46
2. Beryllium 52
3. Summary of Number of Workers Eligible 56

VI. Feasibility of Contacting Former Hanford Workers 58

A. Locator Resources 58

B. Pilot Mailings 59

VII. Description of Phase II: Approach to Medical Surveillance 62

A. Asbestos Surveillance 64

2. Hearing Surveillance 64
1. Beryllium Surveillance 65
2. Limitations 65
3. Summary, Recommendations and Continuation of Phase I 66

Acknowledgement 67

References 68

Table 1. Estimation of the Size of the Population 18

Table 2. Demographics of Hanford Workers 20

Table 3. Estimated Survival of Workers by Age Group 21

Table 4. List of Hazards 23

Table 5. Number of Workers Exposed by Hazard 26

Table 6. COCS Categories by Decade 27

Table 7. Demographics and Pulmonary Function of Hanford 35

Workers

Table 12. Estimated Need for Medical Surveillance 55

Table 13. Estimated Need for Medical Surveillance 57

Table 14. Response Rates for Pilot Mailings 61

Figure 1. The Hanford Worker Population

Figure 2. Job-Exposure Matrix

Figure 3. Occupational Category by Decade - 1940s

Figure 4. Occupational Category by Decade - 1950s

Figure 5. Occupational Category by Decade - 1960s

Figure 6. Occupational Category by Decade - 1970s

Figure 7. Occupational Category by Decade - 1980s

Figure 8. Occupational Category for Decades 1940 - 1990

Figure 9. Pattern of Hearing Loss

The Defense Reauthorization Act of 1993, Public Law 102-484, Section 3162 mandates, "The Secretary shall establish and carry out a program for the identification and ongoing medical evaluation of current and former Department of Energy employees who are subject to significant health risks as a result of exposure of such employees to hazardous or radioactive substances during such employment." This needs assessment responds to the cooperative agreement from the Department of Energy (DOE) Request for Applications (RFA) soliciting applications for cooperative agreements to Support Medical Surveillance for Former Department of Energy Workers. The RFA calls for a two-phase approach. Phase I is directed at conducting a needs assessment, and Phase II is directed at providing medical surveillance for former DOE workers. The goals of the two phases specified in the RFA are to:

The Department of Energy’s Hanford Site has evolved over the last 53 years from a sparsely populated agricultural area into an enormous and complex industrial facility (1). In 1944, construction began in an effort to build the nation’s first plutonium production facility. Construction continued into the 1950s as the site became more and more complex. A total of nine nuclear reactors and five nuclear materials reprocessing canyons were built and operated at Hanford. As a result of over 40 years of nuclear materials processing, an enormous amount of high level radioactive and chemical waste has been generated and is now stored at the site. The hazards associated with the site have included heavy metals, solvents, asbestos, beryllium, ionizing radiation, noise, and other safety hazards associated with construction and heavy industry (1-21). The extent of exposure to these hazards has not been adequately measured or recorded in a consistent manner, but it is likely that many workers were sufficiently exposed to warrant medical surveillance for the health effects associated with these hazards.

The purpose of this Phase I project was to evaluate the need for medical surveillance of former Hanford workers, to identify those at significant risk for occupational disease, and to demonstrate the ability to contact former workers in order to provide appropriate notification and/or medical surveillance. These results form the basis for a plan (Phase II) to offer medical surveillance to workers at the Hanford site who are at increased risk for occupationally-related diseases and for whom identification of those exposures or illnesses would be of benefit.

This needs assessment seeks to address four questions posed in a letter dated May 30, 1997 by Dr. Paul Seligman, Deputy Assistant Secretary for Health Studies of DOE. These questions are:

1) Does the report clearly document the need for establishing a medical evaluation and/notification program for the targeted former workers?

To address these questions we have organized our methods section and the report to:

All aspects of this needs assessment involving human subjects were reviewed and approved by the institutional review boards at the University of Washington and the site-specific board at Hanford (Pacific Northwest National Laboratory).

Identification of the population has required stitching together multiple databases. Since award of the contract we have been working closely with the Department of Energy Headquarters, the local Richland DOE Office, Pacific Northwest National Laboratories, Flour-Daniel Hanford Company, Oil Chemical and Atomic Workers (OCAW) and the Hanford Environmental Health Foundation to identify and gain access to key databases. These databases are characterized, when available, for the following information:

A. Name

B. Purpose

C. Location / owner

D. Number of individuals

E. Years covered

F. Types of data included (personal identifiers, job title, duration, exposures, health outcomes, etc.)

G. Comment on data quality (validity, completeness, reliability etc.)

Each of the databases used or anticipated being used pending access is described below.

Access to databases related to the Hanford site is difficult for many reasons including national security concerns, privacy considerations, protection of human subjects and the costs of access. We have received excellent cooperation from the Department of Energy’s Richland Office, Hanford Environmental Health Foundation, Pacific Northwest National Laboratories and Fluor-Daniel Hanford at the site to systematically address these issues. As a result, we have gained access to a sufficient number of databases to provide this initial needs assessment. As discussed elsewhere, the conduct of a needs assessment is an iterative process. We propose to continue these activities during Phase II in order to provide optimal identification of workers who will benefit from surveillance.

Access to three crucial databases has been delayed due to one or more of the following reasons: 1) need for joint University of Washington and local IRB approval; 2) need to secure letters from each of the prime contractors granting access; 3) need to assure compliance with the privacy act; 4) need to negotiate costs of access; and 5) securing approval and execution of a work order to provide the database. As a result access to three key databases for final population enumeration is still pending. These databases are:

The OHH88 database was compiled from the OHH88_OP operators data set and OHH88_CO construction workers data set, received from Jeff Buchanan from Pacific Northwest National Laboratories. This data was originally from the REX Radiological Exposure System, and they were the source files for Ethel Gilbert’s cohort. This database was combined with the REMS database from DOE-HQ. REMS has social security numbers (SSNs) for 41,614 of the 42,874 records. REMS was then matched with OHH88 and there were 10,342 matches on SSN. This resulted in a database with 97,097 records (64,565 + 42,874 = 107,349 total less the 10,342 matches). The Flow Gemini database from the Hanford Environmental Health Foundation contained 47,557 workers who were former workers, current workers, or had too little job data to address employment status. Of the 14,253 Flow Gemini workers without employment information, 7,836 had no match in OHH or REMS. An additional 7,673 records not found in OHH or REMS were added for a total of 104,770. An estimate of the number of current workers was made by querying the August 1997 Hanford Employment Directory. This eliminated 13,816 leaving 90,954.

Gilbert’s study of the mortality of Hanford workers (1945-1986) suggests the mortality experience was similar to or even less than that of the general population in the United States (SMR 0.83) (22-26). Based on these results age specific survival rates were calculated for the population used in Gilbert’s study (OHH88 database). These survival rates were then applied to the entire cohort in order to estimate the proportion of workers surviving in 1997.

Retrospective estimation of exposures for individual workers has been difficult. To estimate exposures we have:

• Reviewed documents describing hazards on site
• Created a job exposure matrix
• Used building location as a proxy for possible exposure to beryllium.

Pending resources for exposure estimation:

• Employee Job Task Analysis
• Individual Worker Exposure Questionnaire

Once workers have been contacted and have signed a consent form to participate in our study, they are sent a follow-up questionnaire eliciting the details of their work history at Hanford, specific information about the hazards to which they were exposed, and what personal protective equipment was used for each job held at the facility. The questionnaire is composed of two parts: Part 1 is the Job History and General Health Form; Part 2 is the Job Specific Information Form. Workers will receive five copies of Part 2 and may request additional copies as needed to complete their job history. As of this report, we are currently piloting the questionnaire. A copy of the questionnaire, cover letter, and reminder postcard is included in Appendix B. The results of this questionnaire will be subject to some problems of recall by the study participants. Nonetheless, they will be extremely useful in refining the estimates in the job-exposure matrix and in obtaining building information. The information gained from this questionnaire will be particularly useful in understanding exposure potential in the early years of Hanford operations as none of our industrial hygienists were on the site prior to the 1980s. Information gathered from questionnaire responses will also be used to assign individual workers to specific medical surveillance programs as will be defined in Phase II.

The Hanford Occupational Health Process (HOHP) is developing a systematic hazard-based surveillance program. The identification of hazards is through the employee job task analysis (EJTA). This program will assess hazards for each worker on the site. In a separate project we are validating EJTAs being performed by facility supervisors and industrial hygienists. Although the EJTAs are being done only on current workers, they will provide valuable information regarding exposures by job and building for the more recent decades during which clean-up work has become the primary focus.

Documents cataloguing exposures on the site were reviewed. The documents reviewed include:

• Office of Technology Assessment. Complex Cleanup: The Environmental Legacy of Nuclear Weapons Production. US Congress OTA-O-484. US Govt Printing Office, Washington, DC, 1991.
• Epidemiologic Surveillance Data Center and Office of Epidemiology and Health Surveillance, US Dept of Energy. Epidemiologic Surveillance 1992: Annual Summary for Hanford Site.
• National Research Council. Building Consensus through Risk Assessment and Management of the Department of Energy's Environmental Remediation Program Commission to Review Risk Management in the DOE's Environmental Remediation Program. National Academy Press, Washington, DC, 1994a.
• BEMR. Volume I. Estimating the Cold War Mortgage. DOE, March 1995. DOE/EM-0232.
• BEMR. Volume II. Site Summaries. DOE, March 1995. DOE/EM-0232.
• The Blush Report. Blush SM, Heitman TH. March 1995. Train Wreck along the River of Money: An Evaluation of the Hanford Cleanup. A report for the US Senate Committee on Energy and Natural Resources.
• Building Consensus through Risk Assessment and Management. National Resource Council (NCR), 1994.
• CERE Report. Health and Ecological Risks at the US Department of Energy’s Nuclear Weapons Complex: A Qualitative Evaluation. March 1995.
• CERE (Xavier). Inventory of Public Concerns. Xavier University. Draft, 1995.
• Chemical Safety Vulnerability Working Group Report. DOE, September 1994. DOE/EH-0396P. Volume 1 of 3.
• Closing the Circle on the Splitting of the Atom. DOE, January 1995.
• Committed to Results. DOE, April 1994. DOE/EM-0152P.
• Confederated Tribes Reports. Scoping Report. Nuclear Risks in Tribal Communities. Confederated Tribes, 1995.
• Environmental Management 1995. DOE, February 1995. DOE/EM-0228.
• Plutonium Vulnerability Management Plan. DOE, March 1995. DOE/EM-0199.
• Hanford Environmental Dose Reconstruction (HEDR), Technical Steering Panel. Summary: radiation dose estimates from Hanford radioactive materials releases to the air and Columbia River. Richland, WA: HEDR PNNL, 1994
• Hanford Thyroid Disease Study (HTDS), Pilot Study Final Report. Seattle, WA: Fred Hutchinson Cancer Research Center, 1995.
• Archived Industrial Hygiene Records. Richland DOE Office. Reviewed by Kathy Ertell.
• Archived Exposure Records Collected by Field Industrial Hygienists. Richland DOE Office. Reviewed by Kathy Ertell.
• U.S. Department of Energy, Office of Environmental Management. Linking Legacies: connecting the Cold War nuclear weapons production processes to their environmental consequences.

The 73 existing Common Occupational Classification System (COCS) Codes developed by the DOE were examined by our industrial hygienists and grouped within the more broad COCS categories resulting in the development of 42 distinct occupational exposure categories. Each of the occupational exposure categories represents a group of job categories likely to have been exposed to the same hazards at Hanford. A list of the COCS codes included in each category are listed in Appendix C. A job-exposure matrix was then constructed such that an estimate of exposure could be assigned for each of the 42 hazards to each occupational category for each of five decades (1943-1990) of Hanford operations.

Because the OHH88 database uses census codes rather than COCS Codes for job classifications, the census codes were re-coded by two of our industrial hygienists (KD and KE) so that COCS codes and the occupational exposure groups could be used in all of our analyses. The re-coding scheme is provided in Appendix C.

Due to the lack of quantitative data available, it would be impossible to make quantitative estimates of the intensity of exposure for the matrix at this time. It is possible, however, to make qualitative estimates of the likelihood of exposures in each occupational category for each time period. This was deemed sufficient for the purpose of estimating the number of exposed individuals in an effort to assess the need for medical surveillance. The estimates are based on the training and experience of the industrial hygienists and review of the referenced materials.

A group of four certified industrial hygienists was assembled to develop estimates for the completion of the matrix. (see Appendix D for a list of industrial hygiene staff) Two of these industrial hygienists had had extensive experience at the Hanford site (KE and EB). One had some knowledge of Hanford operations, and some experience doing retrospective exposure assessments of this type, but no experience at the site (KD). The other had extensive experience in the area of epidemiologic exposure assessment, but little familiarity with operations specific to Hanford (NS). Each of the four hygienists were given an opportunity to independently assign qualitative exposure estimates for each hazard to each of the occupational categories for each decade of Hanford operations. Exposure categories were: "probably not exposed" (0), "possibly exposed depending on location and specific tasks" (1), and "probably exposed" (2).

All four industrial hygienists then convened to develop one job-exposure matrix with exposure estimates assigned by group consensus. It should be stressed that the numbers in the matrix are qualitative in nature and are not an indication of exposure intensity

Using the job exposure matrix and work history data, the number of workers with possible or probable exposures to each hazard was estimated. The denominator for this estimation was the 78,427 (75%) of the 104,770 with one or more job titles. This permits an estimation of the likely exposures for each worker.

For some of the hazards, job category will be less predictive of exposure than will building assignment. This is why many of the job categories were assigned a "1" for "possibly exposed" in the job matrix. Workers with the same job title who worked in different buildings might have very different exposures. Thus, we must also consider estimating numbers of exposed individuals by location rather than by job.

Unfortunately, the only database that we have obtained to date that contains any information about building assignment is Flow Gemini. We have used Flow Gemini to identify workers in specific buildings in order to construct populations of workers exposed to targeted substances. Due to the limitations of Flow Gemini, it is difficult to reach conclusions regarding total numbers of people exposed as a result of building assignment based on this database alone. We plan to use REX, which we expect to contain a more complete work history information, including building assignment, to more accurately identify those who have worked in buildings of concern. Another source for this information will be an individual exposure questionnaire (Appendix B).

The Federal Register notice put forth the goals as:

Based on these goals the hazards on the site were reviewed to identify which ones met both the criteria of having the potential to cause occupational illness and lead to beneficial medical interventions. For hazards for which we have adequate data, medical literature on occupational hazards and potential surveillance programs was reviewed to provide a justification for medical surveillance for former workers exposed to noise, asbestos, and beryllium. As additional exposures are characterized, it is likely that medical surveillance will be justified for some of those exposures.

The term medical surveillance is used in the context of this report to include identification of workers at an increased risk, provision of medical screening, provision of recommendations to the worker for further testing, treatment, workers compensation when appropriate, preventive measures, and a summary of findings and recommendations to DOE to assist them with future hazard reduction. In many cases (e.g. beryllium and asbestosis), periodic monitoring for latent diseases is anticipated, but not included in the proposed estimates at this time.

The Flow Gemini database contained health outcome data. These data were reviewed to identify outcomes which may point to adverse effects which are related to past occupational exposures. The data reviewed include pulmonary function and audiometry results. Future analyses will also include blood screening tests (lead, mercury, liver function tests, and renal function).

Pulmonary function (spirometry) results were evaluated with descriptive statistics. Using the percent predicted for FVC and FEV1, and absolute ratio of FVC and FEV1, the percent abnormal (< 80%) was calculated as was the distribution by pattern of abnormality (normal, obstructive, restrictive,) (27). Data on the reliability of measures acceptability of tracings as per the American Thoracic Society’s standards were not available. Abnormal spirometry results were identified using the prediction equations of Knudson, and the following definitions: normal was defined as FVC = 80% predicted, FEV₁/FVC = 0.7; restrictive ventilatory defect was defined as FEV₁/FVC ³ 0.7, and FVC < 80% predicted; obstructive ventilatory defect was defined as FEV₁/FVC < 0.7 and FVC > 80% predicted; mixed obstructive/restrictive defect was defined as FEV₁/FVC < 0.7 and FVC < 80% predicted(27).

To assess the potential for the abnormalities to be associated with job titles with asbestos exposure, the proportionate ratio of percentage abnormal by trade to expected percent abnormal for the entire cohort (based on the average percentage of those less than 80% predicted for FVC and FEV₁) was calculated for those jobs in the job–exposure matrix with no, known or suspected asbestos exposure.

Audiometry data was analyzed to calculate the number of workers with a standard threshold shift (STS) and age-adjusted STS and the number with compensable impairment as defined by the Washington State Department of Labor and Industries which manages workers compensation for the Hanford Site. In addition, to assess whether the pattern of loss was similar to that seen in noise-induced hearing loss (high frequency) the mean loss for those with greater than two tests and whole body impairment or STS was calculated for frequencies 500 through 4000 HZ (28).

Whole Person Impairment is calculated using the Washington State Department of Labor and Industries guidelines based on American Medical Association guidelines as follows. Hearing levels for 500, 1000, 2000, 3000 Hz over 100 or below 0 dB were recoded to 100 and 0 respectively. Percent monaural hearing impairment is then computed by summing of 500 through 3000 Hz hearing levels with any sum over 368 recoded to 368 dB, divided by 4, subtract 25 and multiplied by 1.5%. Binaural hearing impairment is 5 times the monaural impairment for the better ear plus the monaural impairment of the worse ear divided by 6. Whole Person Impairment is then determined according to a table which converts from binaural hearing impairment.

STS was computed by subtracting the mean hearing level for 2000, 3000, 4000 Hz for the baseline test from the mean of the last test for each. A mean loss of 10 dB in either ear is considered as a STS. Age adjustment was performed as allowed, but not required, by the Occupational Safety and Health Administration

(28).

Based on the number of workers identified and alive, their job titles, and (for beryllium) their job location, an estimate of the number of workers who should be eligible for medical monitoring was made. Based on the data obtained from the pilot mailings, estimated survival rates, interest in participating and ability to locate, these estimates are adjusted to reflect the likelihood that a worker will request an evaluation.

Determining the location of former workers is a crucial step in delivering medical surveillance. If workers cannot be located, they cannot be contacted for notification of potential exposures and medical surveillance cannot be delivered. In our pilot project, we have learned about the locating process from other researchers who have been successful in this activity, especially the Fred Hutchinson Cancer Research Center in Seattle, Washington, which offers a tracking resource service for researchers attempting to locate "lost" study subjects. We have evaluated and tested some of the Center’s methods to see which work best for locating former DOE workers.

Locating former Hanford workers on the lists generated from the Flow Gemini database has been done using a variety of methods. Special care was taken to ensure that current workers were not included in the former worker rosters; names were screened against a roster of current workers. The roster of current workers was provided by Lockheed Martin Hanford Company and is updated on a monthly basis.

Our Phase I location of workers used readily available resources: regional phone directories on compact disc, the Social Security Death Index, reverse directories, postal change of address, county records, and historical records. Our goal was to determine the feasibility and utility of these inexpensive, easily accessible resources for location of former DOE workers.

Our first step in locating former Hanford workers was to check the names and last known addresses through current phone directories for the Pacific Northwest region. In our pilot projects we have found that approximately 35% of located workers were found through this initial screening step. Many Hanford workers appear to have stayed in the area after retirement or termination from Hanford. In fact, approximately 75% of located former Hanford workers in our Phase I projects were found in the local TriCity area, comprised of the towns of Kennewick, Pasco, Richland, and outlying rural communities within a twenty mile radius such as Benton City, Prosser, and Grandview.

Phone directory searches were used to locate subjects who have moved outside the local area as well. Although this is more difficult because no identifiers are included in phone directories, it can sometimes produce good results. We have found the most success with this method when the person’s birthplace is known, as people often relocate to their home state or town at some point in their lives. We have also had some success in searching states where other DOE sites are located, as many DOE workers leave one site and go to another.

If potential matches were found in the phone directory searches, a confirmation phone call was made to ensure that the person located was indeed the former worker for whom we were looking. Confirmation was made by asking the person to confirm their date of birth and that they were Hanford workers. The confirmation step could also be done by letter. In any case, confirmation of correct identity is absolutely essential since there are often several potential matches for any name and address.

If initial phone directory searches were not successful, the Social Security Death Index (SSDI) was the next resource used. If the former worker’s Social Security number is known, this resource can be used to determine whether workers are deceased. The Fred Hutchinson Cancer Research Center currently uses the SSDI as one method of tracing study subjects. In their recent experience, approximately 3% of "lost" study subjects were found to be deceased. Since a relatively high proportion of former Hanford workers are of retirement age, as compared to the general population from whom the Center’s population was drawn, we considered that our rate of deceased former workers may also be higher. In one of our Phase I pilot projects, we found that out of 262 former workers selected from a database as being potential beryllium workers, 35 workers (13%) were deceased. This pilot project list was composed of people who had worked in processes during the 1950s-1980s, so older age groups were well represented.

We feel it is best to determine vital status fairly early in the locating process, since contacting the families of deceased workers may cause discomfort and suffering for them. However, because people who have died within the last year will not be found in the Social Security Death Index, and names and Social Security numbers derived from databases may be inaccurate, some contact with the families of deceased workers is inevitable, and must be handled with tact and discretion.

If initial phone directory searches and SSDI searches were not successful, local reverse directories, such as the Polk Directory, were used to locate neighbors, employers, or spouses. Contact was then made with these individuals to determine if they knew where the former worker is currently living.

If no success was obtained after these steps, local obituaries were searched. Obituaries are maintained alphabetically at local historical societies. Similar obituary records are also maintained in most communities. Obituaries can confirm deaths which have occurred in the past year. Since many obituaries contain the names and current locations of the relatives of the deceased, in some cases it is sometimes useful to review obituaries for deceased with the same last name as the former worker. Doing this provided us with an additional contact.

In some instances, county assessor’s records were also checked to determine the current owner of the property at the former worker’s last known address. The current property owners are then contacted for possible clues about the location of former property owner.

If no positive leads were determined after these steps, a postal change of address inquiry was filed. This involves asking the Post Office for current forwarding addresses. However, forwarding addresses are only available for the past year. Therefore, this method was used last because we believed it was the least likely to produce positive results for our pilot population.

Of the list of 262 potential beryllium workers, 162 (62%) were located and their identities confirmed. Thirty-five (13%) were found and confirmed to be deceased. The remaining 65 (25%) have not been located by the methods above so other locating methods will be used for these workers.

To assess the feasibility of contacting workers, four pilot mailings of study packets were sent to a total of 3,898 former workers. The first mailing was sent to a list of 128 workers whose names were provided by the OCAW as retired union members receiving union pensions. The second and third mailings included two lists of 126 workers generated from the Flow Gemini database, one addition OCAW worker and 14 additional workers who had requested packets as a result of outreach efforts. The fourth mailing went to 3502 workers on a list generated from the Flow Gemini database and one more worker who requested a study packet.

Former workers included in the second and third pilot were contacted by phone to verify addresses. The fourth pilot mailing was sent without first locating workers to verify their addresses. The different methods were used to determine the value of spending the time and money to accurately locate individuals prior to mailing out the packet.

The study packet (Appendix B) included a cover letter, an instruction sheet, an initial contact form, two copies of the consent form, a brochure about our study, and a postage paid return envelope. Additionally, for each pilot mailing a reminder postcard was sent within two weeks of the former worker receiving a study packet.

An Exposure Questionnaire (Appendix C) will be mailed out to each worker who agrees to participate. At this time, a pilot of 43 have been sent out in order to assess the questionaire.

The mailings were analyzed for response rates indicating a willingness to participate and location of workers.

Previous estimates of the number of former Hanford workers have been based on a number of different sources. One source was the PeopleCORE database, which lists all of Hanford prime contractor employees as of 1988. Another source was the list of 52,522 operators and 2,285 construction workers derived from employment records extracted by Ethel Gilbert and placed in CEDR for her 1989 mortality study of workers who began employment between 1945 and 1983 (22-26). Other sources included various union lists. All of these lists are considered to be an underestimation of the entire former Hanford worker population either because of restrictions on the time period included, or because of the omission of the potentially large number of employees of subcontractors and sub-subcontractors.

By obtaining other databases from the DOE and its contractors, we have been able to construct a more comprehensive database of former Hanford workers than has been available to date. We have obtained the original employment history files (OHH88) from which CEDR was derived. This database includes an additional 9758 workers who were excluded from the mortality study (because they were not known to be exposed to radiation, bringing the total number of operators to 53,105 and construction workers to 13,740. Because 2,280 workers are included in both the operator and the construction worker files leaving 11, 460 workers who were solely construction workers. The total number of individual workers from these files is 64,565. Some of these may be current workers, but the exact number has not yet been determined.

In addition, from the DOE headquarters, we have obtained the REMS database which contains all radiation monitoring data between 1987 and 1996. The REMS database contains 42,874 Hanford workers.

We have also obtained the Flow Gemini database maintained by the Hanford Environmental Health Foundation (HEHF), which includes workers who were seen by the medical contractor between 1985 and the present. This database containing a total of 47,542 workers is, in many ways, the least reliable of all the databases we have obtained to date. It contains 14,253 records that lack work history information. These workers may or may not have worked at Hanford. The total number of definite former workers in Flow Gemini is 19,494.

Table 1 outlines the numbers of workers in each of the three databases. There is significant overlap between the three databases as shown in Figure 1. When this overlap is accounted for, there are a total of 104,770 Hanford workers in the three databases. The site currently has approximately 13,816 workers employed in the DOE Richland Office and by the prime contractors. There is no easy method to determine the number of employees of subcontractors working on the site but this may represent up to another 9,000. These are not included in the calculations of eligible workers. In addition, some workers may have been construction workers only. Gilbert estimated that 11,460 workers had been construction workers only and they were excluded from her cohort (22 - 26). For these reasons the estimated total number of former workers (excludes current and construction only) is reduced by 11,460. An estimated 87.4% are alive in 1997. Excluding workers who were only in construction and adjusting for survival and excluding those deceased the final estimate of eligible, living former workers is 67,736. In the final estimates of those who might request medical surveillance, it is likely that not all workers are included. Therefore, this is an underestimate due to lack of ascertainment of subcontractor employees. However, this represents the best available estimate.

B. Demographics

Table 2 displays the mean age of this workforce in 1997 as 56 years old, 70% male, 24% female, and 6% "gender missing". The ethnic distribution is 94% Caucasian, 2% African-American, .1% Asian or Latino, and 3.5% "race missing". Because these data are frequently not complete, the number of workers with information for each variable is presented.

C. Mortality Estimates

Table 3 shows the estimated number of workers alive in 1997 by age group. Of the 104,770 in the master database, the estimated total number of workers alive in 1997 is 91,525.

OHH88 64,565

Operator file 53,105

Construction worker file 13,740

Workers included in both files 2,280

Solely a construction worker 11,460

REMS 42,874

FLOW GEMINI 47,557

Former workers 19,494

Current workers 13,795

Workers without employment information 14,253

Probable Hanford workers 6,432

Probably not Hanford workers (7,821)

Flow Gemini Hanford Workers 39,721

OHH88 and REMS overlap 10,342

OHH88 and FLOW GEMINI overlap 20,320

REMS and FLOW GEMINI overlap 21,968

Total in OHH88, REMS, and FLOW GEMINI Combined 104,770

Exclude Estimated Current Workers -(13,816)

Estimated Solely a Construction Worker (13,740 - 2,280) (11,460)

Total Estimated Deceased Workers (12.6% X 90,954) (11,460)

Total Estimated Eligible & Living Former Workers 68,034

Note: Percentage Alive in OHH88 Data Set and Percentage Assumed Alive in Master Data Set differ due to the changed gender composition of the Hanford work force from 1988 to 1997.

Exposures are characterized in several ways including review of the literature and documentation of exposures at Hanford. Table 4 shows the major exposures which have been present at some time on the site.

To estimate the number of workers exposed to individual hazards a job-exposure matrix was constructed. The 73 existing Common Occupational Classification System (COCS) Codes developed by the DOE were examined by our industrial hygienists and grouped within the more broad COCS categories resulting in the development of 42 distinct occupational exposure categories. Each of the occupational exposure categories represents a group of job categories likely to have been exposed to the same hazards at Hanford. A list of the COCS codes in each category are listed in Appendix C. A job-exposure matrix was constructed such that an estimate of exposure could be assigned for each of the 42 hazards in Table 4 to each occupational category for each of five time periods of Hanford operations.

Figure 2 shows a summary of the job-exposure matrix. This figure demonstrates the exposures for major categories of jobs. A more complete version is shown in Appendix D.

Table 5 provides the number of workers exposed to each hazard based on the job-exposure matrix. These estimates provide the basis of estimating the number of workers who may have been exposed to a hazard and may benefit from medical surveillance.

The types, intensity and duration of exposures likely changed with the changing work processes at Hanford over the past 5 decades. An analysis of the relative proportions of job categories was conducted to examine these changes. Table 6 and Figures 3 - 8 display the relative proportions of major job categories. This table and these figures show the number and proportion of job titles assigned by decade. Because workers might be assigned to multiple jobs these numbers are only an indirect approach to assessing changes in jobs and exposures.

COCS Catagory by Job Begin Year Decade for 78,427 workers with begin job date and COCS code

The need for medical surveillance was estimated by identifying those workers with specific exposures and identifying those exposures where medical surveillance would lead to medical interventions. For this needs assessment medical interventions was broadly defined. Specifically, surveillance was considered for exposures which would lead to:

• interventions to alter the course of a disease; and/or
• interventions which could identify substantial impairment, and/or health risks and reasonably require worker notification.

The rationale for screening examinations which would identify disease at a point where interventions could affect its course is well documented and needs little justification (29-33,43). For selected exposures we have provided specific justification for surveillance below.

The rationale for screening evaluations to identify disease which could lead to worker notification is based on an ethical duty to notify workers of increased risk (33). Notification may be important in modifying diagnostic or treatment interventions (e.g. notification of asbestos exposure might void the necessity of an open lung biopsy in a case of pulmonary interstitial fibrosis) and, in some cases, these workers may be eligible for workers compensation. For many workers this may be limited to risk communication.

In reviewing exposures, three exposures were identified for which medical surveillance can be well justified at this time. These are:

• asbestos;
• noise
• beryllium

In addition, we expect that as we continue our Phase I activities, additional exposures will become sufficiently characterized to justify surveillance. At this time, however, there is too little information to warrant their inclusion in a surveillance program. Exposures of specific concern include: ionizing radiation, welding fumes, other respiratory irritants, metal working fluids, solvents, heavy metals, and other carcinogens. Acquisition of the REX database in conjunction with analyses of medical outcome data and the results of additional exposure assessments based on worker surveys will be completed over the next

4-6 months and an updated Phase I needs assessment will be provided.

Justification of medical surveillance for exposed workers is based on identifying evidence of significant exposures and also identifying interventions which will be of benefit.

The hazards of asbestos exposure are well recognized (34-41). These include pleural effusions and fibrosis, parenchymal fibrosis, bronchogenic carcinoma, mesothelioma, and elevated rates of malignancy in the upper respiratory and gastrointestinal tracts . The interaction of asbestos exposure and cigarette smoking in increasing the risk of lung cancer also deserves special attention. Asbestos exposure and smoking appear to have a multiplicative effect (40). Among those with the highest exposures, asbestos insulators, the risk of lung cancer among smoking asbestos exposed workers appears to be increased up to 50 fold (30,40). Lesser exposures appear to have less risk.

Asbestos use, in the United States, began at the turn of the century and rapidly increased at the time of World War II. Asbestos exposure was greatest among construction and maintenance workers. Because asbestos fibers n the respirable range remain suspended in the air, there is substantial potential for secondary exposure to those who work nearby.

Asbestos was widely used at the Hanford Site. This is based on the nature of the work and the need for extensive use of thermal insulation. There is already an extensive asbestos monitoring program in place. This program is reflected in the Flow Gemini database which includes a total of 13,092 workers from the period 1985 to 1997. While some chest radiographs have been read clinically and, according to the International Labour Organization’s system for classifying pneumoconioses, by certified B readers, the results are not available in the Flow Gemini system. Given these limitations asbestos exposure was estimated through the job exposure matrix. An estimated 27,988 workers who worked in jobs where asbestos exposure was possible or probable. An estimated 21,555 workers are alive who worked in jobs were asbestos exposure was likely, Of these 67 percent are known to have worked at Hanford for one year or more. Demographics of these workers are described in Table 7.

In order to assess the potential effects of respiratory toxins such as asbestos, an analysis of the lung function data in the Flow Gemini database was conducted. While this database began to assimilate data in the 1980s the findings likely represent, for some workers, the cumulative effects of exposure from earlier years. Table 7 displays the demographics and baseline lung function of workers in the Flow Gemini system who have had spirometry. Table 8 displays the proportion of workers in each COCS code with various patterns of abnormality. Normal lung function was defined as FVC = 80% predicted, FEV₁/FVC = 0.7; restrictive ventilatory defect was defined as FEV₁/FVC ³ 0.7, and FVC < 80% predicted; obstructive ventilatory defect was defined as FEV₁/FVC < 0.7 and FVC > 80% predicted; mixed obstructive/restrictive defect was defined as FEV₁/FVC < 0.7 and FVC < 80% predicted(27).

There are several limitations to these analyses. There are likely many reasons why spirometry was obtained on these workers. The potential for respondent bias and limitation in exposure assessment limit the conclusions which can be drawn from these data. The lack of data on smoking status and exposure to other respiratory toxins (e.g. silica, beryllium, welding fumes) raises the potential for misclassifying or attributing abnormalities from smoking to asbestos exposure.

Despite these limitations, there appears to be a substantial cohort of asbestos exposed workers (Table 5). Analysis of available lung function data demonstrate higher than expected rates of abnormality (Tables 6 - 8). For these reasons, it is reasonable to be concerned that sufficient asbestos exposure may have occurred to result in increased risks of lung cancer and pleural and parenchymal fibrosis.

Medical surveillance for asbestos related malignant and non-malignant respiratory disease can be justified on several grounds. First and foremost, for those with asbestos exposure who smoke, identification and patient education concerning the risk and importance of smoking cessation will have the benefit of reducing risk over time (40). While there is limited efficacy in smoking cessation programs, there is evidence that quit rates of 5 to 20% can be achieved. Second, many with asbestosis are at risk for misdiagnosis. Appropriate diagnosis can result in avoiding non-beneficial and potentially invasive and expensive evaluations of dyspnea and respiratory disease. In addition, while there is general concensus that screening for lung cancer with chest radiographs (or other measures such as sputum cytology) is not beneficial. This is not uniformly accepted. There are data that screening chest radoigraphs may identify cancers at an earlier stage permitting resection. For these reasons a screening examination which focuses on smoking status, respiratory symptoms, chest radiograph and spirometry is warranted.