Description of the Fernald Medical Monitoring Program University of Cincinnati College of Medicine

Transcription

1 Description of the Fernald Medical Monitoring Program University of Cincinnati College of Medicine Medical Director: Epidemiologist: Program Coordinator: Robert Wones, MD Department of Internal Medicine Susan Pinney, PhD Department of Environmental Health Jeanette Buckholz, RNC, MSN, CEN Department of Internal Medicine The Fernald Medical Monitoring Program (FMMP) is a voluntary ongoing medical surveillance program for community residents living within five miles from the perimeter of a former US Department of Energy uranium-processing site. Currently there are 9764 persons enrolled in the Fernald Medical Monitoring Program (FMMP); of these 8770 were adults (18 years or older) at the time of their first examination. Most participants enrolled during a three-year period from September 21, 1990 to November 30, Samples of whole blood, serum, plasma and urine were obtained from all participants at the time of the initial examination, and over 100,000 1 ml aliquots of these biospecimens have been stored at minus 80 degrees C since then. Funding for medical examinations will expire after one more semi-annual examination cycle, at the end of Extensive uranium dose reconstruction using methods developed by the CDC demonstrate that over 60% of the cohort had such minimal exposure to uranium and radon that their cumulative ionizing radiation exposure was less than 3.2% over lifetime background levels. For the US population, background ionizing radiation exposure is estimated to be 3.6 msv per year or 288 msv lifetime (80 years), but the range of individual exposure estimates is wide, depending primarily on geographic location (1-10 msv per year). The median estimate of cumulative lifetime exposure beyond background for this 60% subcohort is only 5.5 msv, very small compared to the yearly individual dose variation. (See Appendix A for details.) Fernald Medical Monitoring Program (FMMP) Upon entry to the FMMP all participants received a thorough medical examination and diagnostic tests including chest x-ray, electrocardiogram, pulmonary spirometry, and mammograms and cervical pap smears for female participants. Laboratory tests included the usual series for hematology, chemistry and lipid profile of blood serum, and urinalysis. At conclusion of the examination physicians recorded any new medical diagnoses apparent at the time of the examination, as well as diagnostic uncertainties requiring further diagnostic testing. Nurses employed by the program conduct participant follow up by phone for at least six months on all those needing further testing or treatment by their usual primary care practitioner. (We believe that this follow-up has been a key to our excellent record of continuing program participation.) New diagnoses recognized through this follow up also are recorded on the medical record. Outside medical records (usually a pathology report) are obtained to validate any new diagnosis of cancer. P:\Fernald\FMMP description.doc 1

2 Fernald Medical Monitoring Program - Number of questionnaires and medical exams (adults) by Program year Number of questionnaires/exams # of questionnaires # of exams 11* 12* 03** 04** Year of the program (* Years 11 & 12 - lower numbers, since this was during the convertion to a calendar year cycle, and some skipped exam/ques by those names; ** the 2003 # of exams includes all those performed during the exam cycle) This same procedure is used at the time of the periodic medical examinations, offered every three years until In December 1998, financial analysis of monies left in the Settlement Fund indicated that re-examinations could be offered every 2 years (1999 to present). To date, the FMMP has conducted 33,854 medical examinations through the end of Female participants age 40 and older are offered mammograms each year (including years when a complete medical surveillance examination is not scheduled), and men over 50 years are offered a PSA every other year. Following the initial questionnaire, participants also are asked to complete yearly questionnaires requesting information about new medical problems and recent hospitalizations and surgeries, and 77,638 completed questionnaires from adults have been received by the program. Yearly questionnaire return rates vary by Program year, but in 2004 were 59% of the original enrollees. Our yearly questionnaires are computer-generated, so that participants who missed a past questionnaire are re-issued critical data items from those missed questionnaires, greatly improving total ascertainment on any data item. Continued participation in the program at the time of re-examinations fell 23% at the first re-exam for adults and has remained at about 50% for each examination since then. (Some participants skip one exam but then return for the next one.) Only 216 participants have withdrawn from the Program and are no longer contacted or followed (except to obtain information when we learn of their death); l84 participants are lost to follow-up. Thus, over 50% of initial participants still get semi-annual examinations and we know the location and vital status of 97% of the initial participants. P:\Fernald\FMMP description.doc 2

3 FREQUENCY Years of follow-up Questionnaire data collection is designed so that a different version of the form is sent to those who have skipped a year of questionnaire return, covering multiple years rather than a single year of medical history experience. Overall follow-up has been excellent. For a recent analysis we calculated follow-up time, considering the period under observation to be from the first examination date until the last questionnaire or exam contact, or death. For most program participants, we have 12.5 years of follow-up. Death certificates have been obtained, or an NDI search completed, on most of the 662 participants who have died. FMMP medical records are thoroughly reviewed by both quality assurance nurses and coders after every medical examination, performing quality assurance for completeness and clarity of data. Robert Wones, MD, has been the Program Director and Susan Pinney, PhD has been epidemiologist, both serving in those roles since the inception of the Program. Jeanette Buckholz, RN, MSN is the second Program Coordinator during the 15 year tenure. A description of the Fernald Medical Monitoring Program, and some relevant publications, can be found on the website The Program requires that participants sign a consent at the time of each examination, and has IRB approval ( EE, last approved on 2/6/2007), including approval for use of data and biospecimens for research. Archived specimens: At first examination obtained three 1-ml aliquots of whole blood, plasma, serum, urine and urine with buffer, or 15 aliquots per person for future analyses. Since then, these specimens have been stored in minus 80 degree freezers. Although one of four freezers (at Holmes Hospital) failed during the July 4 weekend in 1996, and all samples in that freezer were lost, many of the plasma and serum samples were replaced when participants came for their next examination. Few of the specimens have been used for research. Since 2001 specimens have been stored in six freezers at CCHM Research Lab (old Bethesda Oak Hospital space). The status and temperature of each freezer is continuously monitored by a central alarm system with established protocols for notification and transfer of specimens if needed. The freezers have a CO 2 back-up system and undergo scheduled maintenance through a contract with the SoLow Company. Strengths of this program for research Medical history on this population is extremely well characterized, verified, and all cancer diagnoses have been validated with medical records. Extensive disease risk factor information P:\Fernald\FMMP description.doc 3

4 has been collected, and for most information, prior to disease diagnosis. All members of the cohort have received medical examinations, conducted at the clinical site dedicated to the program, under well-developed protocols. Conduct of examinations is monitored with various quality assurance measures. Blood pressure, pulse, height, weight, and waist circumference are taken during the examination. Physicians record their findings on a 10 page form designed for complete data collection, including a review of systems and a separate section for each component of the complete physical examination. The form includes a section on the last 2 pages where the physician lists medical problems, and orders program follow-up (for limited further diagnostic testing) and records recommendations (for further follow-up by the patient s own physician or for health education/lifestyle changes.) This medical examination form is immediately reviewed by one of the nurse quality assurance specialists. Following receipt of all diagnostic testing results, a summary letter from the physician to the patient reporting all the examination findings is composed and reviewed by the examining physician. At this time an extensive audit of the record is performed, to assure that all test results have been obtained, protocols for follow-up have been implemented, data recorded in the chart, and all pertinent information captured in the letter to the patient. The initial examination also included EKG s, pulmonary spirometry, and hearing testing. Questionnaire data collection has been extensive. The initial questionnaire was a 27 page form, and yearly questionnaires average about 14 pages. Each year participants provide information about new medical problems and the name of the physician who has treated them, and hospitalizations or surgeries, including the health care facility where care was received. Program nurses follow-up on this information through phone calls to participants, and obtain outside medical records where indicated. Participants are also asked to list their medications, copying the name, dose and frequency from the medication container label. Participants bring all medications to their exam, so that accuracy and completeness of the medication information can be verified. All medications have been assigned numerical codes on forms and in the database. Questionnaires are designed with built in redundancy for information verification (so that medications and surgeries can be compared with self-reported and physician-reported medical problems). Systems are in place to follow up with participants on certain health information recorded on questionnaires as well as the follow-up after medical screening. Yearly questions also ask about items such as smoking and alcohol use, diagnostic imaging, and influenza vaccinations. Each year participants also are asked a set of special questions. Past sets of special questions have asked for information about family history of cancer, with specific questions for each type of relative for each cancer, physical activity, sunburns and skin moles, female reproductive history and use of oral contraceptives, and breast feeding. Complete reproductive history information has been obtained from both males and females (6 page questionnaires) and yearly pregnancy updates are obtained on all female participants of childbearing age. During the current exam cycle we ask participants to complete a 2 page check list of malignant and non-malignant health conditions (and provide year of diagnosis for each condition checked). Diet history information was collected during the first and second year of the program, and will be collected again during the next examination cycle. All medical history and risk factor information is coded and stored in a relational database. All information in the chart (medical examination findings, laboratory tests, diagnostic test results, questionnaire information) is housed in a SAS database. The FMMP database consists of over 160 individual files in SAS format. All data has been entered using SAS P:\Fernald\FMMP description.doc 4

5 applications designed for double data entry with verification. In addition, various computer generated quality assurance reports are used for follow-up and to detect data outliers, which are then verified with the paper chart and the laboratory, if needed. A separate set of SAS files were created for each year of the program to maintain and manage the data collected from the physical exam forms and ongoing questionnaires. Some data collected during the program are maintained in single, cumulative SAS files rather than divided up by program year. Mammograms, chest x-rays, and laboratory results are examples of data maintained in single SAS files. The entire SAS database totals more than 1 gigabyte of data and spans over 10,000 variables, although many of the variables are the same from one program year to the next. All diagnoses have been assigned ICD-9 codes by a certified medical record coder. All medication information is coded by type of drug using a 4-digit code. The extensive coding of the medical information on this large cohort provides the immediate opportunity to ask specific research questions. For example, all of the 28,518 mammogram reports and the 27,855 chest x-ray reports in the database have been literally coded (coding of phrases) using a list of over digit codes, specific to each type of imaging. Exposure or risk factor metrics have been developed from questionnaire information, and are available in the database. For example, we have developed matrix of cumulative cigarette pack-years for all participants, for each year after their year of program enrollment. We also have matrices for family history of each type of cancer for each program participant, with number of first degree and total number of blood relatives with that type of cancer. We have collected family structure information from all participants on several different occasions and through several types of questions (family history at the time of the medial examination, questions about who else lived with you, and questions about offspring on one of the yearly questionnaires). In each instance, we have recorded names of these relatives and coded each mentioned family member with their FMMP ID, if they are program participants. With each questionnaire we also ask participants of the names of three personal contacts, and record their relationship to the program participants. The FMMP database contains the information needed to create family pedigrees, and note which family members are Program participants. In the near future, we plan to create family pedigrees using software such as Cyrillic or Progeny. We are familiar with the methods of pedigree creation, through Dr. Pinney s work on other genetic epidemiology studies, and have constructed a few pedigrees from this FMMP cohort. For example, through the work of a Medical Student Summer Research Program project, we easily constructed extensive pedigrees of two families with autosomal dominant polycystic kidney disease, using the information in our database and in the paper charts. The FMMP has repeat measures on blood, serum and urine analytes on all participants, and a Medical Heritage program designed to allow standardization of the data for time dependent changes. At each examination, specimens are obtained for clinical laboratory evaluation of hematology (complete red and white blood count with differential), serum chemistry, kidney and liver function tests, FBS (fasting blood sugar), lipid profile (total cholesterol, triglyceride, LDL and HDL), TSH (thyroid stimulating hormone), and urinalysis with microscopic examination and urinary beta-2-microglobulin. Some additional testing is done on those whose test reaches a certain criterion level. For example, hemoglobin AIC is measured on all diabetics or those with a blood glucose 126 mg/dl a serum beta-2-microglobulin is measured on all those with a urine level 250 ug/l. PSA is measured in all males 50 years at the time of each P:\Fernald\FMMP description.doc 5

6 examination, and yearly in those at risk of prostate cancer (25% increase over the previous PSA). Many aspects of the program were designed to foster prospective study of the cohort. The Medical Heritage Program (designed by Drs. Amadeo Pesce and Bradley Copeland) for the stewardship of the clinical laboratory data is only one example. Not only has this program preformed quality assurance by using such measures as 30 day moving averages of the healthy Fernald participants, we have established that it is possible to track laboratory data for each analyte in a manner that will allow for adjustment for possible laboratory analytical drift found in any prospective study (Copeland and Pesce, 1992). An extensive collection of archived samples, most collected in , and currently in excellent condition, already exists and is available for studies of proteomics, genetic variation, and urinary markers of both exposure and disease. Samples of whole blood, serum, plasma and urine were obtained from all participants at the time of the initial examination and in 1997, and over 100,000 1 ml aliquots of these biospecimens have been stored at minus 80 degrees C since then. Quality of the specimens has been assessed on two occasions (primarily for degradation of protein and DNA) and found to be excellent. In 2000, the lyophilization (freeze dry) effect was found to be 7% in two freezers, 4% in one freezer, and minimal in one freezer. In serum, albumin increased (from level measured at time of the sample collection) about 2%, and total protein about 7-11%. Since the concentrations in the archived samples were as high or higher than the initial measurements, these changes indicate concentration due to storage rather than degradation. Some enzymes showed slight levels of degradation (billirubin, which is light sensitive; GTT; lactate dehydrogenase) but at acceptable levels. There were no decreases in AST or ALT. Of 50 whole blood samples from which DNA was isolated, only two yielded almost no DNA. For others, the DNA was of both good quantity and quality. A second more limited quality assurance analysis was conducted in Of 10 whole blood samples, the DNA isolated was of sufficient quality and quantity to be used for an array of genetic analyses (Marshall Anderson lab). More recently (2006), Dr. Shuk-mei Ho s lab found that isolated DNA was of high quality, in fragments of 15kb and greater. Dr. Detlef Shuman tested 10 serum samples for proteomic studies, comparing the samples to fresh standard human serum. None of the samples showed significant protein degradation, concentrations were in the expected range, and protein identification was consistent. Biomarkers of exposure can be measured in both blood and urine samples, and we are preparing an application to CDC for measurement of environmental biomarkers in the FMMP cohort. The FMMP has an established procedure for applications for access to data and biospecimens. (Procedure description and application forms available on request.) After recommendation of the Citizens Advisory Group, the Court-appointed Trustee has approved over 25 requests for access to data. Only one request has been disapproved. There have been only two requests for the archived samples thus far, but both have been approved. The Trustee and the Federal Judge are both very much in favor of the samples (and data) being used for research. The extensive information on participants of the FMMP allows for the selection of tailored control groups. Criteria for controls for studies may differ depending on the disease and the lifestyle factor being studied. For example, the FMMP has provided Dr. Marshall Anderson with a set of controls for studies of identified susceptibility genetic variants related to cigarette P:\Fernald\FMMP description.doc 6

7 smoking and lung cancer. For this study, the controls were persons with no first degree relatives with lung cancer, age 60 years, and with a 20 pack year cigarette smoking history. Our large cohort and extensive database enabled us to identify more than the needed 200 controls from the group without environmental uranium exposure. We will be conducting a similar control selection for Dr. Daniel Nebert, searching for highly resistant persons for his study of susceptibility genes for head and neck cancer, but with a 80 pack year smoking history and no history of any cancer. Persons who meet these criteria are very rare in any population, but we have located 33 controls with these criteria. The cohort is not racially diversified, which enhances the opportunity for detecting relationships between genomic variation and disease, and for discovering biomarker predictors of disease. Since 95% of the cohort is White-non-Hispanic, statistical power of genetic studies and precision of data analyses are improved by the lack of racial diversity of this cohort. Although the lack of generalizabilty to populations with other racial composition is certainly a limitation, the statistical associations found in this population can later be tested in other more racially diverse populations. The database is an excellent resource for developing methods for data mining, and the potential discovery of new statistical associations will lead to hypothesis generation. The extensive database for this longitudinal cohort provides rich opportunities for data mining, and for developing and testing new methods of data mining. Copeland BE and Pesce AJ. The medical heritage concept: a model for assuring comparable laboratory results in long-term longitudinal studies. Annals of clinical laboratory science 1992; 22(2): Pinney SM, Freyberg RW, Levine GH, et al. Health effects in community residents near a uranium plant at Fernald, Ohio, USA. International Journal of Occupational Medicine and Environmental Health. 2003;16(2): P:\Fernald\FMMP description.doc 7

8 Appendix A: Exposure History: The Fernald plant was part of the US Department of Energy (DOE) nuclear weapons complex and produced uranium metal products including nuclear reactor fuel elements during operation from Releases from the site resulted in exposure to ionizing radiation, especially radon, soluble and insoluble forms of uranium, and various other organic and inorganic chemicals. The medical surveillance program has been funded by a settlement of class action litigation between community residents and National Lead of Ohio, a contractor for the DOE. The Fernald Settlement Fund is administered by the US Federal Court, which maintains oversight of the Fernald Medical Monitoring Program (FMMP). Although the program was founded on the basis of a potential exposure and related health effects, later dosimetry construction has shown that a large proportion of those in the screening program received environmental uranium exposure that was negligible compared to background exposure. In the early 1950's, the U. S. Department of Energy built a nuclear fuel processing plant near Fernald, Ohio, known as the Feed Materials Production Center (FMPC). This industrial facility processed uranium ore, which was chemically converted into a series of uranium salts. Uranium metal was then extracted from the salt in a process conducted at extremely high temperatures, and used primarily to produce highly refined uranium metal rods. These rods were shipped to other DOE sites to be used in nuclear reactors. The Centers for Disease Control and Prevention (CDC) has conducted a historical exposure characterization and developed dose estimation models through the Fernald Dose Reconstruction Project, with an endpoint of developing an algorithm to estimate doses to individual persons who lived within the exposure assessment domain (area within a five mile radius from the perimeter of the plant property). Eligibility for the FMMP requires residence within a five-mile radius of the uranium processing facility at Fernald, Ohio, for a minimum continuous two-year period between January 1, 1952 and December 18, Former workers at the plant are excluded from this medical surveillance program, but are eligible to participate in a separate Fernald Workers Medical Monitoring Program. Although FMMP was established to monitor individuals for health effects of ionizing radiation, findings of dose reconstruction conducted by the Center for Disease Control indicated that a large proportion of those in the screening program received environmental uranium exposure that was negligible compared to background exposure (Killough, 1998a; Killough, 1998b). Extensive uranium dose reconstruction using methods developed by the CDC demonstrate that over 60% of the cohort had such minimal exposure to uranium and radon that their cumulative ionizing radiation exposure was less than 3.2% over lifetime background levels. For the US population, background ionizing radiation exposure is estimated to be 3.6 msv per year or 288 msv lifetime (80 years), but the range of individual doses is wide (1-10 msv per year). The median estimate of cumulative lifetime exposure beyond background for this 60% sub-cohort is only 5.5 msv, very small compared to the yearly individual dose variation. Killough GG, Case MJ, Meyer KR, et al. The Fernald dosimetry reconstruction project: Task 4: Environmental pathways - models and validation. RAC Report No. CDC-3. Neeses, South Carolina: 1998a. P:\Fernald\FMMP description.doc 8

9 Killough GG, Case MJ, Meyer KR, et al. The Fernald dosimetry reconstruction project: Task 6: Radiation doses and risk to residents from FMPC operations from RAC Report No. 1-CDC-Fernald-1998-FINAL. Neeses, South Carolina: 1998b. NCR (National Academy of Sciences/National Research Council). Health Risks of Radon and Other Internally Deposited Alpha-Emitters. BEIR IV Report. Washington D.C.: National Academy Press; P:\Fernald\FMMP description.doc 9