Appendices

• Appendix A: The Savannah River Site and Surrounding Area (with Figure 29)

• Appendix B: Demographic Description of the SRRHIS Population (with Table 2)

• Appendix C: Methodology

• Appendix D: Cancer Mortality Studies in this Region

• Appendix E: Completeness of Reporting and Quality Control Issues

• Appendix F: Numbers of Cancers Reported, by Primary Site

• Appendix G: Age-adjusted Incidence Rate for Invasive Cancers, by Primary Site

• Appendix H: Summary Stage by Primary Site by Race, SRRHIS, 1991-1993

• Appendix I: Technical Terms

The Savannah River Site (SRS) is a U.S. Department of Energy (DOE) complex occupying approximately 300 square miles along the Savannah River, principally in Aiken and Barnwell counties of South Carolina (Figure 29). SRS was established in the mid 1950's by the U. S. Atomic Energy Commission to produce plutonium, tritium, and other special nuclear materials for national defense, other government uses, and some civilian purposes. The Westinghouse Savannah River Company (WSRC) has managed and operated the site since April 1, 1989.

The site is located approximately 25 miles southeast of Augusta, GA, and 20 miles south of Aiken, SC in a mostly rural area. Farming in the region is diverse and includes crops such as cotton, soybeans, corn and small grains. Area industrial and manufacturing facilities include textile mills, plants that produce polystyrene foam and paper products, chemical processing plants and a commercial nuclear power plant.

The Savannah River flows along the southwestern border of SRS, with five major streams feeding it. These streams, which receive discharges from various SRS operations, are not used as commercial sources of water. The Savannah River is used as a drinking water supply for about 65,000 people located about 100 miles downriver in Port Wentworth, GA, and in Beaufort and Jasper counties of South Carolina. There is no known use of the Savannah River for irrigation by farming operations downstream of SRS.

The 1992 population estimates for the SRRHIS area include 678,679 persons living in the twelve Georgia counties and 445,577 persons living in the ten South Carolina counties for a total population of 1,124,256. Of these, about 61 percent are white, 38 percent black and slightly over one percent other races. In each race group there are more women than men; overall 51.6 percent are women.

About 60 percent of the population resides in counties making up the 1990 Census designated Metropolitan Statistical Areas (MSA) of Savannah, Georgia (249,641 persons) or Augusta, Georgia (424,913). The other 40 percent of the SRRHIS population should be considered to live in rural counties (449,702). The Savannah MSA includes Chatham and Effingham counties. The Augusta MSA includes the Georgia counties Columbia, McDuffie and Richmond, and Aiken County in South Carolina. The rural counties include Bryan, Bulloch, Burke, Evans, Jefferson, Jenkins, and Screven in Georgia; and Allendale, Bamberg, Barnwell, Beaufort, Colleton, Edgefield, Hampton, Jasper and Orangeburg in South Carolina.

Cancer data for this report are generally grouped into major cancer types or site groups based on the primary site of the cancer and the histological type of the tumor. The primary site reflects the organ where the tumor originated and the histological type reflects the cellular characteristics of the tumor. This information is taken from the medical record and coded by the abstractors according to the International Classification of Diseases for Oncology, 2nd ed. (ICD-O2). Cancer cases used in this analysis were recorded into standard site groups consistent with SEER data as recommended for analytic reports by the North American Association of Central Cancer Registries (NAACCR). While most cancers are grouped by the primary site for epidemiologic analysis, certain cancers such as lymphomas, leukemias, and melanomas are grouped by cell type for medical treatment and etiologic reasons.

The principal statistic used to present the rate of cancer occurrence is the age-adjusted incidence rate, describing the number of new cases of cancer per 100,000 persons living during the period. For this report the rate was annualized, so that three years of cancer data (1991-1993) were included to estimate an annual rate. Age-adjustment means that the rates are calculated for specific age groups, then weighted to a standard population to calculate an overall rate comparable to the cancer rate in other populations. The statistic is conventionally used to compare populations that may have different age structures. For this analysis, age-specific rates were first calculated for five-year age groups, then adjusted to the 1970 U.S. standard population (Breslow, 1987).

Population estimates for rate calculations were developed by the U.S. Bureau of the Census and made available through the National Cancer Institute's SEER program. Population numbers used were 1992 estimates by five-year age groups by race and sex for the twelve Georgia counties and the ten South Carolina counties included in the Savannah River Region Health Information System. Only estimates for 1992 were used as this was the middle year for case data reported for 1991 through 1993, and 1993 estimates were not yet available.

Age-adjusted incidence rates were calculated from SRRHIS data complete as of January 1996 for this report. Rates were calculated for the entire registry area, and for the Georgia and South Carolina counties separately. Rates for the Atlanta SEER area and for all SEER areas combined for the years 1990-1992 were generously provided by the National Cancer Institute SEER Program.

Before the advent of this present report of incidence, the only area cancer studies were those related to deaths from cancer. The three published Savannah River Site (SRS)-related studies of cancer mortality are summarized here for the interest of the reader. Of the mortality studies: two analyzed cancer deaths of residents near the SRS, the other evaluated the cancer mortality experience of SRS workers, past and present.

The first study was undertaken in the late 1970's by Dr. Herb Sauer, a professor in the Department of Community and Family Medicine, University of Missouri, under contract with du Pont, which was managing the Savannah River Plant. He wished to assess the mortality experience of residents living close to the plant, and those most likely to receive low doses of radiation from the facility. To this end, he selected for study the populations of all counties within 50 miles of SRS, and, for comparison purposes, those between 50-149 miles, and others over 150 miles away. Data for the U. S. overall were also used for comparison. He collected mortality data for the period, 1949-1978. This choice provided data for the years immediately prior to production of nuclear materials at SRS (in the mid-50's), and for all available subsequent years. He identified no significant deviations in cancer death rates from what would be expected (Sauer, 84).

The second study was a major study of cancer mortality in persons living close to nuclear production plants throughout the United States. It was undertaken around 1990 by the National Cancer Institute under a congressional mandate. Congress was responding to the often-voiced concerns of its constituents about hazardous materials escaping from such plants. The study, directed by Dr. Seymour Jablon and his colleagues, utilized mortality data over the 35 year period, 1950-1984. The three-volume work was published in 1990. Every nuclear production plant in the US, whether public or private, was included for review. The scientists tabulated death rates before and after the plants began production; and they tabulated them for not only the counties in which the plants were located, but also for three "control" counties in the area and similar to the one in which the plant was located. This permitted death rate comparisons for periods before and after plant start-up, and between similar county populations in the region. They analyzed cancer rates for leukemia (for children and for all ages separately), all other cancers combined, digestive cancers, lung cancer, and breast cancer. The analyses were done for all production sites combined, and for individual production sites. They concluded there was no difference between any of the cancer rates before and after start-up of plants, nor between the test counties and the three comparison counties. At the SRS site they found a slightly elevated rate in leukemia for the 20-39 age group, but they found a reduced rate for stomach cancer for all ages. They attributed such findings to the usual statistical variability in the occurrence of rare events, such as leukemia, and did not conclude there was any causal connection to the SRS operation (Jablon, 1990).

The third study was focused on SRS workers. A study of their cancer mortality experience was published in 1988 by Dr. Donna Cragle, an epidemiologist, and her colleagues of Oak Ridge Associated Universities. They studied all workers from the beginning of production at SRS in 1952 to mid-1981. An update of that study, adding new workers employed between 1981 and 1986, was reported by the author in March 1995, when a news conference was held at SRS. Dr. Cragle and her co-workers constructed a registry of 9,860 white men employed more than 90 days between 1952 and 1986. She reported less all-cancer mortality than the US general population; a slightly higher leukemia rate than the US general population; and called attention to a relationship between radiation dose level and leukemia mortality. Having a lesser all-cancer mortality rate than the general population is to be expected because workers are, as a group, healthier than the general population. This is called the "healthy worker" effect. The slightly higher leukemia rate for these workers, and the correlation with radiation dosage, were felt to require follow-up by the scientists. With the support of the National Institute of Occupational Safety and Health, they have begun a case-control study of leukemia among these workers. The report of that study is a few years away (Cragle, 1988, 1995).

The above studies all tabulate cancer deaths by county of residence. None of them present cancer occurrence for small areas (smaller than counties) close to the River or to SRS. The occurrence of cancers by smaller geographic area of residence, such as zop codes or neighborhoods, can be plotted using SRRHIS registry data. Plans are being made to look closely at these smaller area experiences when the number of incident cases becomes sufficient (five years of data).

Remember that mortality data, while it defines the ultimate ill effects of cancer, does not reflect the much broader impact of cancer occurrence on the population. That is so, in some measure, because of improvements in cancer treatment and screening which have resulted in cures for some cancers, and longer survival for others. Further, for a number of complicated reasons, mortality data is less susceptible to effective quality control measures than incidence data (new cancer cases). The current best measure of cancer occurrence in the SRRHIS area is the annual number of incident cases in residents, and the first report of that data is this publication.

At the beginning of this project in 1991, South Carolina had no law requiring cancer reporting by physicians, hospitals, or pathologists. The collection of data from the participating hospitals required voluntary cooperation. A separate agreement was established with each hospital. The Medical University of South Carolina is an agency of the State of South Carolina and is therefore able to indemnify each facility pursuant with the procedures of the SC Tort Claims Act, SC Code Section 15-78-10, et. seq. (1976, as amended) with respect to any liability specifically relating to breaching confidentiality of medical records while conducting the SRRHIS project.

In May of 1996, the South Carolina Legislature passed a cancer reporting law, with support from the South Carolina affiliate of the American Cancer Society, the SC Hospital Association, and the State's physicians in general. Thus, the Department of Health and Environmental Control (DHEC), which began the development of a state cancer registry in 1995, gained the cooperation of all hospitals and clinics in the state. The DHEC and SRRHIS registries have worked cooperatively since the beginning of each registry, to avoid duplication of effort. Indeed, the collaboration of the two programs is mutually beneficial, cost efficient and insures the completeness of case-finding.

Cancer is a reportable disease in Georgia. The Georgia Center for Cancer Statistics, at Emory University, under contract with the Georgia Department of Human Resources, is the organization responsible for administration of the Georgia Cancer Registry, and is authorized pursuant to the Health Code of the State of Georgia 31-12-1,2, to collect and analyze data related to cancer incidence in Georgia. The SRRHIS project is a part of the Georgia Cancer Registry, and the mandatory reporting requirements, therefore, apply.

In order to attain the most complete measure of cancer incidence in residents of the SRRHIS catchment area, the cooperation of the Charleston area hospitals is needed as many SRRHIS residents may go to Charleston hospitals for diagnosis and treatment. Gaining access to hospitals within the SRRHIS catchment area was accomplished with ease as all facilities in the outlying areas were eager to participate due to a shared concern for community health and the education of the general population in the area. However, complete participation by hospitals within the three county (Berkely, Charleston, Dorchester) Charleston area was not obtained. With the passing of the cancer reporting law and the advent of the DHEC cancer registry, it is hoped that these difficulties will no longer be an issue.

One special accomplishment of the SRRHIS should be noted here, namely the establishment of a formal program of data exchange between the health departments of the two states. This exchange is especially critical to understanding the true rate of cancer occurrence in South Carolina residents, because so many South Carolina residents along the River seek their medical care in Georgia. For this reason, the true rate of cancer in the South Carolina SRRHIS residents would be severely underestimated if South Carolina resident cases were not identified through Georgia facilities and pathology records. There is less of an impact in the other direction, because fewer Georgia residents receive health care in South Carolina. Because SRRHIS includes a contract with Emory University, which is responsible for the Georgia registries, this project allows for cancers in residents who cross the borders for health care to be included in our cancer statistics.

Data for cases diagnosed between 1991 and 1993 revealed a total of 21 residents from Georgia who were diagnosed and or treated at facilities in South Carolina. Many of these cases were found at a large teaching cancer care center in the three county Charleston area. However, lack of participation by the Charleston Veterans Administration Hospital meant the loss of some cases, such as Georgia SRRHIS area resident veterans who came to Charleston for care. On the other hand, 1,409 cases of South Carolina SRRHIS area residents were reported by hospitals in the Georgia catchment area. The majority of this patient population traveled from Aiken, Beaufort, Edgefield, and Jasper counties for diagnostic and or cancer care in the Augusta and Savannah areas. Advanced treatment modalities are more readily available in these urban areas.

Major North Carolina cancer treatment centers are located within a driveable distance from the border of South Carolina. Thus, it was anticipated that some SRRHIS area residents would seek treatment in North Carolina. Access to the latest modalities in cancer treatment and nationally approved clinical research trials are available at these centers. A request for data on SRRHIS South Carolina and Georgia residents confirmed the above. There were 123 SRRHIS eligible cases reported by the North Carolina State Registry and added to the SRRHIS database.

Death clearance is a method for computer-matching the personal identifying information of persons who die with cancer as reported on death certificates to state vital statistics offices, to those identifiers of the incident cases reported to the registry. Cancer deaths which are not matched to an earlier cancer case are then followed back to ascertain whether it was a cancer case which was missed. These are instructive exercises, and they are critical because they constitute the best measure of the completeness of reporting of the area cancers. Procedures are being established to initiate a thorough follow-back mechanism as part of the casefinding process. Cases which are reported by means of Death Clearance Only are additional incidents of cancer, if follow-back reveals that such cases truly meet the reporting criteria of SRRHIS. Due to the lack of cancer reporting legislation, and the shortage of staff time, the death clearance process was done only for one year. A standardized follow-back procedure was developed, but it could not be initiated in time for this report.

The quality of any data base is highly dependent on the mechanisms used to gather, edit, and analyze the data. It was not surprising that some deficiencies in these abilities were evident in the early years, and that, as a consequence, the 1991-1993 data are somewhat under-reported. Two deficiencies were especially frustrating: follow-back to physicians because of incomplete pathology reports, and obtaining SRRHIS case information from hospitals distant from the SRRHIS area. The ability to contact physicians was limited to a degree because the case information was confidential, and permission to get the hospital case information on South Carolina cases was conditioned on avoiding contact with physicians or patients (or family members). Casefinding at distant hospitals was limited by staff time, and by the unwillingness of several Charleston area hospitals to participate in the project. The passage of the cancer reporting law, and the advent of the state registry are expected to solve these deficiencies.

Nonetheless, SRRHIS activities and findings indicate an overall high degree of quality and completeness of reporting. Indeed, initial results are quite encouraging because they appear to be similar to cancer rates that would be "expected" for a region like SRRHIS.

The confidentiality of the patient, physician, and facility is guarded at all times. The same standards of confidentiality that exist in the traditional medical care setting extend to the SRRHIS registry. Patient, physician, and facility identifiers are not shared with staff outside of SRRHIS. Security procedures are in place which restrict access to registry data and ensure the data is used only for its intended purpose.

• Table 12: Summary Stage by Site for Whites, SRRHIS, 1991-1993

• Table 13: Summary Stage by Site for Blacks, SRRHIS, 1991-1993

Source: SRRHIS, March 1996

Source: SRRHIS, March 1996

Age-adjusted rate: An age-adjusted rate is a weighted average of the age-specific cancer incidence (or mortality) rates, where the weights are the proportions of persons in the corresponding age groups of a standard population. Because cancer is partly a disease of aging, age-adjustment allows for more accurate comparisons of rates among populations with different age structures. In the current publication, the 1970 U.S. population was used as the standard.

Cancer: A group of diseases characterized by uncontrolled growth and spread of abnormal cells.

Carcinogen: A substance that causes cancer.

Carcinoma: A malignant tumor originating from epithelial tissue from various sites of the body.

Distant (Remote): Malignant invasive neoplasm that has metastasized to distant organs or lymph nodes.

Incidence rate: The cancer incidence rate is the number of new cases of a specific site/type occurring in a specified population during a year, expressed as the number of cancers per 100,000 people.

In Situ: A neoplasm with all the characteristics of malignancy except invasion; it has not penetrated the basement membrane nor extended beyond epithelial tissue (i.e. non-invasive).

Invasive: The ability of a malignant tumor to spread into surrounding normal tissues. In those sites that have a basement membrane, invasive tumors penetrate the basement membrane.

Localized: Malignant invasive neoplasm confined to the organ of origin.

Malignant: "Highly injurious disease" as in malignant hypertension, malignant hyperthermia, malignant hypercalcemia, malignant cardiac arrythmia, malignant cattarhal fever virus, etc. In reference to tumors it means tumors that invade, infiltrate, and destroy adjacent normal tissues.

Mortality rate: The cancer mortality rate is the number of deaths with cancer given as the underlying cause of death ocurring in a specified population during a year, expressed as the number of deaths due to cancer per 100,000 people.

Regional: Malignant invasive neoplasm in an organ that has invaded other organs by direct extension, or regional lymph nodes.

SRRHIS: Savannah River Region Health Information System (See Summary)

SRS: Savannah River Site (See Appendix A)

Stage of Disease: General summary scheme developed by the Cancer Surveillance, Epidemiology, and End Results Reporting Program (SEER) that indicates the progression of a cancer. Extent of disease is determined at the time of diagnosis.

Surveillance, Epidemiology and End Results (SEER) Program: A cancer incidence and survival reporting system operating under the National Cancer Institute which reports incidence on an estimated 10% of the U.S. population.