ML20133D391
| ML20133D391 | |
| Person / Time | |
|---|---|
| Site: | Three Mile Island  |
| Issue date: | 06/07/1981 |
| From: | Tokuhata G PENNSYLVANIA, COMMONWEALTH OF |
| To: | |
| Shared Package | |
| ML20132B167 | List: |
| References | |
| FOIA-85-285 NUDOCS 8507220189 | |
| Download: ML20133D391 (7) | |
Text
., i.
@$-75] *~,
y HEALTH STUDIES IN THE THREE MILE ELAND AREA
- George K. Tokuhata, Dr.P.H., Ph.D.**
Director *
- Division of Epidemiological Research Bureau of Epidemiology and Disease Prevention Pennsylvania Department of Health P.O. Box 90 Harrisburg, PA 17108 i
- Invuted paper presented at rise special session on Healt s Effects of R diation at the annual n cenng of the .4mencan Nuclear Soci-ety (Enviornmental Sciences Divissoni, June 712.1981. .tliami Beach Florsds
- ' Professor cf Epidemicicgy .tnd Eiostatistics, University of Pstriburgh Graduate School of Public Health f.4djuncts Associate Professor of Communitv and Family Sledicsne, Temple Universsty College sfSledscsne f Adjunct!.
f Dr. Tokuhata was designated by Governet of Pennsvivania to assutne oveta'l management of all health rescatch studies related to the TMI nuclear accident of Match 28,1979.
1 8507220189 850524 PDR FOIA LODGE 85-285 PDR w_
l HEALTH STUDIES IN THE THREE MILE ISLAND AREA The Three Mile Island nuclear accident of March 28,1979 has caused a significant impact upon many facets of human life and social structure not only within Pennsylvania, but also elsewhere in the U. S. and abroad. As a mandate of the State Health Department, it became immediately clear l
that possible health effects of the accident must be evaluated. During the 10-day period of crisis,it was not possible to ascertain accurate information regarding radioactive emissions from the dam-aged reactor into the environment. However, the presence of diffuse and growing psychological dis-turbance in the area was apparent.
! Within a short period of days following the accident we were able to conceptualize and de-velop a comprehensive plan for a variety of epidemiological and other health studies designed to assess the impact of the TMI accident. Specific studies conceived during this critical period reflect mostly the existing epidemiological knowledge regarding biological effect of low level ionizing redi-ation and of severe emotional stress. As Director of the Bureau of Health Research,I was designated by the Governor of Pennsylvania to coordinate and manage all health-related research activities re-lative to TMI. At the same time, a special Ad risory Panel was commissioned by the Secretary of Health to oversee and guide all TMI-related health studies administered by the State Bureau of Health Research. Some of the studies are still in progress, while others are either completed or in the early stage of development. I am going to present a brief description of each study and some of the findings available at this time.
j A. Three Mile Island Census: W v:
One of the first projects initiated shortly after the accident was a special census of all persons living within five miles of TMI. The primary purpose of TMI census was to develop a popula-tion profile (TMI Population Registry), which would provide a basic framework for future studies of morbidity and mortality. In addition to usual demographic data, such as age, sex,'
' and race, the census questionnaire included such items as name, address, social security munber, marital status, smoking habits, medical history (particularly cancer and thyraid dis-ease), recent pregnancy experience, medical and occupational radiation exposure, and detailed whereabouts during the 10-day crisis, when abnormal radioactive releases were reported.
l The DfI Population Registry is comprised of 35,537 persons who were living within five miles I of the damaged nuclear reactor at the time of the accident. This cohort includes 254 persons who moved out of the five mile area between the time of the accident and the time of the TMI census in June,1979. Because of the long latency period for radiation related health effects to ,
appear, tracing the residents as they moved in subsequent years is essential. We have designed and adopted an automated system of annual tracing and updating the original cohort in the registry.
1
- s During the first year after the TMI census was tak:n a total of 3.526 ptrsons, excluding nurs-ing home patients and dormitory students, were found to have moved. This yielded an annual mobility rate of 11.0 percent. Of those who moved, only 27.3 percent moved beyond the ten mile radius of TMI.
The Pennsylvania Department of Health will continue to monitor the potentially exposed pop-ulation for 20 years or more in an attempt to detect possible health effects of the TMI nuclear
~acciilent. Both cross-sectional and longitudinal comparisons will be made by using comparable control populations, while the effects of various other factors, sociodemographic and environ-mental, are taken into account. In addition to the existing data sources, such as mortality and natality certificates and hospital discharge statistics, special surveys will be conducted to col-lect new data. Details of such follow.up studies have not yet been fully developed at this time.
B. Pregnancy Outcome Study:
Both ionizing radiation and severe emotional stress can affect human reproductive process and pregnancy outcome. It is known that the fetus is highly sensitive to such environmental in-sults. To evaluate possible health impact of the TMI accident, a carefully designed prospective study of pregnancy outcome was initiated in August,1979. This study covers all pregnant wo-men residing within a ten mile radius of TMI, who gave births during the March 28,1979-March 27,1980 period. This study cohort consisting of approximately 4,000 deliveries will be compared with a control cohort of another 4,000 deliveries during a one-year period in the same geographic area immediately following the study cohort. The study cohort will also be compared with similar data collected in the same general area during the immediately preced.
ing four. year period.
Pregnancy outcome measures being investigated are: fetal death (including abortions of 16+
week gestation), hebdomadal (one week) death, neonctal (28 days) death, prematurity (gesta-tion 4 37 weeks), immaturity (birth weight 4 2.500 grams), congenital anomalies, and low Apgur score (4 7). -
While the study design is that of "before-and.after" comparison of cohorts, measures of radia- '
tion exposure and psychological stress will be related, on an individual basis, to each of the seven outcome measures described in the preceding paragraph. Two levels of radiation expos-ure for each pregnant woman being estimated are (a) maximum possible dose and (b) most likely dose. Source materials used for these calculations are taken from various TLD data, meteorological information, and detailed demographic data relative to the individual mobil-ity, including evacuation of the pregnant women during the 10-day period of crisis.
Psychological stress was determined in terms of overt statements of upset / worry as expressed by individual women, as well as stress coping patterns. such as taking tranquilizers and/or sleeping pills.
1
[
2
Since thIre are many fact rs thIt are known or suspected to influence pregnancy outcome, such influ1nces should b2 and are taken into considzration. Thtse factors includa matarnal ag?, ract, smoking, drinking, education, occupation, employment, and marital status; matemal medical and obstetric histories, including X ray exposures; prenatal care, including provider character-istics, medications, instructions and special procedures; prior birth control method; and birth order of the index offspring.
C. Congenital / Neonatal Hypothyroidism:
By Pennsylvania law, all newbom babies in the State must be screened for hypothyroidism since July 1978. This screening consists of two stage testing: first, low T4 (thyroxin) and high TSH (pituitary thyroid stimulating hormone) on filter paper and second, low T4 and high TSH on serum test. Confirmation of diagnosis is done through thyroid scan. During the initial six-month period (start up period) testing procedures were not fully standardized and the results were not considered to be complete.
Since it is known that maloactive iodine can cause hypothyroidism, that certain amount of I'3' was released from the damaged nuclear reactor during the T31I accident and thatI'J' can be taken up by pregnant women in the area which, in turn, absorbed by the fetal thyroid gland through placenta, we decided to examine the incidence of congenital / neonatal hypothyroidism among newborn infants. The fetal thyroid gland is much more sensitive to radioactive iodine than is the mother's thyroid gland, i.e., a relatively small dose to the mother can be a relatively large dose to the fetus.
In a normal population, the incidence of congenital hypothyroidism is approximately one in 4,500 to 5,000 infants. There are several different diagnostic classes: namely, genetic type (mostly autosomal recessive, resulting from dyshormonogenesis - i.e., lack of enzyme to syn-thesize thyroxine); ectopic type (dysgenesis - i.e., incomplete maturation and/or disniace-ment of the thyroid glands); agenesis (without thyroid gland); and other types.
During the Starch 28,1979-31 arch 27,1980 period only one case of congenital hypothyroid. -
ism was identified within a ten mile radius among approximately 4,000 infants. This incidence rate is well within a normal range of expectation.
An apparent concentration of seven cases in Lancaster County (mostly beyond the ten mile radius) during 1979 was subjected to a special in. depth analysis. Of the seven cases, one was re-ported prior to the T31I accident; one with severe multiple CNS anomalies was born three months after the accident, which are unlikely to be associated with T311: one case was of dys-genesis representing one of discordant twins, thus, non supportive of the etiology secondary to radiation exposure; and another case of dysgenesis with the thyroid glands being displaced from the normal position.
1 Havmg completed detailed diagnostic analysis and epidemiologic assessment of the cases re.
ported in Lancaster County during 1979, we concluded that cases of congenital hypothyroid-ism were not related to the T31: accident. This conclusion was also consistent with the small
- dose of radioactive iodine estimated for the T31I area after the accident.
3
D. Radiati::n Dose Assessmint:
One of the most sensitive and technically difficult tasks was to evaluate the extent to which local residents may have been exposed to radiation from the damaged nuclear facility. While no direct radiation measurements were taken on an individual basis (except for some 700 local residents who subjected themselves to whole-body counts sponsored by NRC during the crisis),
available TLD recordings and other source data monitored by various agencies are useful for such a purpose. This study is being conducted through contract by the Department of Radia-tion Health of the University of Pittsburgh.
Radiation dose assessment is directed towards every person resident within the five mile rad-ius durmg the nuclear crisis and every pregnant woman resident within the ten mile radius dur-ing the same period. From the five mile census data and ten mile pregnancy study data, it is possible to reestablish detailed account of whereabouts during the ten-day period of all indi-viduals, including those who evacuated as well as those who left the area for other reasons.
In addition, detailed daily meteorological data, including wind direction and velocity, as well as plume-dispersion patterns are being incorporated into the combined monitored radiation dose materials from all reliable sources. With the application of computer mapping technol-ogy and elaborate sector analysis, it is possible to assign (estimate) a reasonably accurate dos-age to an individual in the study areas during the ten-day period. A set of two doses, maximum possible and most likely, are being computed for each individual; the former, without consider-ing physical presence or absence in the area and the latter, with daily mobility data taken into account.
The results of this particular study along with the analysis of background radiation and prior radiation exposure will be useful for a variety of current and future studies of possible health effects of TMI nuclear accident, particularly from the radiation standpoint. .
E. Cytogenetic (Chromosome) Study: ' ' '
During the early stage of the development of TMI Health Effects Research Program, when the -
level of radiation emission was not clearly defined, we recommended that a cytogenetic study he carried out to investigate if the incidence of chromosome breakage in human blood cells is unusually elevated among local residents, including those who are employed by the TMI facil-ity and residents living near the facility. Ionizing radiation, if the dose is high enough, is known to be mutagenic and/or teratogenic and can cause physical damage to chromosome structure.
Clinical significance of chromosome aberrations of this type is not fully understood, however.
1 After a lengthy discussion, the TMI Advisory Panel for Health Research Studies recommended to cancel this particular study on the grounds that (a) the amount of radiation emitted in the area was very small, and (b) the cost of the study would be too high to detect a minute differ-ence or increment in chromosome anomalies that may or may not be present.
I i
4
i F. Health Behavi:ral (Stress) Study:
Shortly after the nuclear accident we recognized that one of the major concerns was the psychological impact upon health among local residents. A number of related questions were asked: (a) How many and what type of people felt emotional distress? (b) How did they cope with the crisis situation? (c) What kind of psychosomatic symptoms were present? (d) What social and medical / health services were utilized by those who were disturbed by the accident?
(e) Is the psychological impact short lived or long-lasting?
The initial survey was conducted in July,1979, the second in January,1980, and the third in October,1980. These studies indicate that (a) persons who are younger, more educated, mar-ried and female were most distressed during the crisis;(b) those who reside within 15 miles of the T31I had more stress than did those who reside farther out;(c) the use of sleeping pills and/or tranquilizers to deal with anxiety, as well as certain psychosomatic symptoms increased among certain individuals;(d) the increased level of anxiety experienced following the accident persisted for approximately one year, but declined substantially in October,1980; and (e) persons with more social support tended to be less distressed than others during the crisis.
G. Mental Health Study: ,
A study of mental health of TMI workers, mothers with small children and mental health clinic patients was conducted by the Western Psychiatric Research Institute at nine months and one year following the TMI accident. The mental health assessments include both clinical and sub-clinical measures.
T3fl mothers had an excess risk of experiencing clinical episodes of anxiety and depression during the year after the accident. They also reported more symptoms of anxiety and depres-sion at subclinical levels.
T3fl workers were essentially similar to control workers (Beaver Valley Nuclear Plants) with re-spect to mental health indicators under study both at the clinical and subclinicallevels.
3fental health clinic patients expressed similar level of symptoms or anxiety at both TMI area and Beaver Valley area se;ected as control.
H. Infant Mortality Study:
Sut. sequent to the March 1979 nuclear accident, we initiated a comprehensive evaluation of the existing vital statistics data in order to determine if the TMI accident has had any measur-able influence upon infant mortality in the vicinity of the plant.
Within the ten mile radius of the damaged nuclear reactor, we found no evidence that the rise or fall of the infant mortality was caused by the accident.
The infant mortality rate was not significantly different between the ten mile area and the State of Pennsylvania for any of the three years, 1977,1978. and 1979. The rate within the ten mile radius was already considerably high during the first quarter of 1979. The rate 5
remained at th7 samo levIl during the second cuarttr of 1979 immedistely following the acci-dint, but declined substantially during the following two quarters. If the TSII accidtnt hrd a significant influence, the infant mortality rate would have increased steadily throughout the last two quarters because of different sensitivity of the fetus to radiation at different gesta-tion periods when exposed.
I. Other Studies:
Several other studies have been developed or conducted. These are (a) Health Economics Study, which was designed to assess immediate and short. term excess health costs (direct and indirect) and expenditures due to UfI accident; (b) Cardiac Mortality Study, which was de-signed to evaluate possible psychological impact upon those local residents who had cardiac conditions;(c) Family Unity Study, which was designed to examine how, if any, the nuclear accident may have affected family unity or bond; and (d) Mobility Study, through which the extent of mobility (moving) can be determined that may be attributable to the nuclear acci-dent.
Summary:
The DII nuclear accident nas presented social scientists and biomedicalinvestigators a umque opportunity to evaluate its impact upon local population. Probably the most important concern is that of safety and health effects of the accident.
From the currently available epidemiological knowledge, no significant physical health effects are expected from the low level radiation released from the damaged D1I facility. However, some substantial psychological impacts are apparent. In addition, because of high sensitivity of the fetus to ionizing radiation and severe stress, timely evaluation of pregnancy outcome should be pursued.
Based on the already established DiI Population Registry, there should be a continuous and long-term epidemiologic surveillance of the exposed population. Such endeavor will make it possible to determine and document, if there is any measurable health impact in humans from the low level of radiation that has not yet been fully studied in humans.
6
i J~ (-
/ ptN&
r.
0 05 hPRyg p [g ,,e w
.(-
q/q. ,_ :
4 3. e .o Department of Health
~YL.
> W *..
- , ., o_s H AR RISBU RG 717-787-5264 July 5,1984 Mr. Phillip Grant TMI Office U.S. Nuclear Regulatory Commission P.O. Box 311 Middletown, PA 17057
Dear Mr. Grant:
Subsequent to our telephone conversation, I have enclosed a copy of our program description on Epidemiologic Surveillance in Pennsylvania: A Case of Nuclear Power Plants. As you can see, this has come out of the TMI episode and, thus, has been reviewed and supported by our TMI Advisory Panel and the Secretary of Health.
I hope this brief description will give you enough information to have general understanding of this program. Since it is related (perhaps indirectly) to the function of NRC, I thought you might be interested. In the near future, I am planning a joint briefing with you and representatives of other concerned agencies including PEMA and DER.
Please feel free to call me if you have any questions about this.
Thank you for your attention.
Sincerely, P $4 :
Georg okuhata, Dr.P.H., Ph.D.
Director Division of Epidemiology Research GKT:u Enclosure 3 ,
N 3=
c-l OYhADQ
} W JL/ V o f = , gg c'P 3 NE.
=
23
EPIDEMIOLOGIC SURVEILLANCE IN PENNSYLVANIA:
A CASE OF NUCLEAR POWER PLANTS George K. Tokuhata, Dr.P.H., Ph.D.
Director Division of Epidemiology Research Pennsylvania Department of Health June, 1984 e
t N
o . 's EPIDEli10 LOGIC SURVEILLANCE IN PENNSYLVANIA: A CASE OF NUCLEAR POWER PLANTS
- 1. INTRODUCTION The need for monitoring and evaluating the health status of populations was clearly indicated by the accident at Three Mile Island on March 28, 1979.
Had the accident been more serious in a health threatening sense, the infor-mational needs of involved government agencies and a concerned public could not have been adequately served in a timely manner. In A,pril, 1980, the Secretary of Health issued a memorandum to the Three Mile Island (TMI)
Advisory Panel for Health Research Studies regarding the "Need for Baseline Data Around Nuclear Power Facilities." The Panel endorsed the concept of
. improved epidemiologic surveillance in general and approved the project for
. development by the Division of Epidemiology Research.
Several important characteristics of the Epidemiologic Surveillance System should be emphasized. First. the primary intent is to establish an automated method of compiling baseline health data for specific geographic areas which can be updated and monitored reutinely. Second, the System is designed as a
- screening" rather than " diagnostic" device; that is, signif-icant changes or differences in health status based on the data analyzed will be regarded only as " potential warning signals." If and when such signals are detected in a population, further " diagnostic" study would be necessary to substantiate the magnitude of the problem and to identify specific cause(s). Third, the Surveillance System, of itself, is not intended to test any specific hypotheses or to determine health effects of radiation as such or any other specific health hazards, but rather to detect significant chances within or differences from norms in any health indicators under consideration that can be determined with the available data. Fourth, the System would provide comparative data which would be
, . 4. . -2' useful in the continuing effort to assess the potential health effects of the TMI accident. The public demand for health information in the TMI area is beginning to rise again as the length of time since the acci-dent approaches the lower limits of the estimated latency periods of cer-tain cancers and many local residents believe that they had been exposed to large doses of radiation. Furthermore, the System would provide health information which would be invaluable in the event of another nuclear acci-dent or any other health threatening event.
II. PROJECT OBJECTIVES In view of practical considerations and imediate need, the objectives of this System are to:
(1) Develop and implement a baseline health information system which will insure the continuous and systematic compilation, analysis and interpretation of available demographic, natality, morbidity, and mortality data for minor civil divisions (MCDs) within 20 mile radii of the six nuclear plants (5 geographic areas) and selected control areas in Pennsylvania.
(2) Establish baseline health indices fer each selected area. These baseline indices established during years of normal operation of nuclear plants can serve as " controls" against which updated indices can be compared in subsequent years. Health indices developed during pre-operational years (new plant sites) can also serve as
" controls" against which post-operational indices can be compared.
(3) Monitor and evaluate the health indices for study populations in the vicinity of nuclear facilities on a recular basis to detect any significant changes over time or differences from norms that may have occurred or been observed.
. C. '
(4) Determine if special studies should be initiated to establish whether significant changes or differences in health indicators, if and when observed, are related to living in proximity to nuclear power plants or to other factors.
III. GEOGRAPHIC AREAS The initial focus of the Epidemiologic Surveillance System will be on populations residing in minor civil divisions (MCDs) wholly or partially within 20 miles of the six nuclear power plants (commercially operational and under construction in Pennsylvania) and populations in control areas.
The nuclear power plant sites include (1) Beaver Valley 1 and 2 (Shippingport,BeaverCounty);(2)Shippingport(Shippingport, Beaver County; (3) Three Mile Island 1 and 2 (Goldsboro, York County); (4)
Susquehanna 1 and 2 (Berwick, Luzerne County); (5) Peach Bottom 2 and 3
'(Peach Bottom Township, York County); and (6) Limerick 1 and 2 (Pottstown, Montgomery County). The surveillance in these areas will cover a total of 565 minor civil divisions and about 2,970,000 people. One hundred and fifty-four MCDs with a population of 760,000 reside within 10 miles of the six power plants and an additional 411 MCDs with a pcpulatien of 2,210,000 reside within a 10-20 mile ring of the plants. The communities (MCDs) selected for this surveillance program are included in the evacua-tion plans developed by the Pennsylvania Emergency Management Agency (PEMA) and the " Plume Exposure Pathway EPZ" plans available for each nuclear plant site.
The control areas selected for comparative purposes, collectively provide a social-industrial-geographical representation of Pennsylvania; except that there are no nuclear power plants within 20 miles of the perimeters .of any of the areas.
' y, *
.IV. HEALTH STATUS MEASURES (INDICES)
The need for' a minimum data set is recognized. It should be noted that certain health indicators or indices (e.g., cancer) have a long latency period (interval between exposure and diagnosis), wh'ile others (e.g., fetal mortality) may have a short latency and effects, if any, can
, be shown within a year or less.
First, the short-term health indicators included in the minimum data set are largely derived from birth, fetal, and infant mortality data. These include:
(a) Fetal mortality (with or without abortions of 16+ week gestation).
(b) Hebdomadal mortality (during first 7 days of life).
(c) Neonatal mortality (during first 28 days of life).
(d) Perinatal mortality (combination of fetal and neonatal mortalities).
(e) Infant mortality (during first 365 days of life).
(f) Congenital malformations, including neonatal hypothyroidism.
(g) Prematurity (gestation less than 37 weeks).
(h) Low birth weight (birth weight less than 2500 grams).
(i) Low Apgar Score (less than 7 at one minute). _
Second, with respect to lono-term health indicators, data on cancer incidence, prevalence, and mortality of the following organ sites are
, important.
(a) Leukemia (g) Lymphatic system (b) Thyroid (h) Bladder (urinary) i (c) Breast (i) Kidney (d) Skin (j) Pancreas I
(e) Lung and bronchus (k) Stomach (f) Bone (1) Esophaous ,
Third, thyroid disease (other than congenital and cancer) will be
, studied as internediate-term indicators, particularly in relation to'the
. release of radioactive iodine into the environment which can be ingested or inhaled by local residents.
Fourth, in view of the possible effects on ceneral life exoectancy or loncevity of a total population, abridaed life tables will be computed periodically using annual mortality data and base populations by age ano sex.
i i
V. METHOD OF ANALYSIS In general, the methods of analysis to be utilized will be determined by general surveillance needs and the requirements of particular studies.
When the baseline morbidity and mortality data are established, a variety of routine epidemiologic analyses will be carried out. Demographic differ-ences in the populations under consideration which could bias results will be adjusted for wherever possible in all analyses.
First, cross-sectional analyses: Various health status measures will be compared between the nuclear power plant (NPP) and control areas for seier.ted years. Such measures will be based on area-specific mortality and/or morbidity data. ,
Second, temporal analyses: Changes over time in health status indi-cators of populations residing within each of the specified NPP and control areas will be assessed. Specifically,.when the baseline health experience of the population around a given NPP is established, the health experience of that particular population can be compared from one time period to another. Such historical analyses have an advantace over cross-sectional comparisons in that much of the basic environmental characteristics (physical, socioeconomic, cultural, etc.) tends to remain fairly stable. On the other hand, historical analyses will permit the relating of significant changes in health status,. if they occurred, to recognizable socioeconomic changes, 1
such as the establishment of a new industry with subsequent influx of workers in certain age groups into the area communities.
Third, temporal-soatial analyses: This is a combination of geographic comparisons with historical comparisons. For example, time trends in mor-tality from specific cancers in the NPP and control areas will be evaluated in reference to each other. This will permit assessment of sicnificant changesincancer,dhathrateswhichmaybelimitedtoaparticulargeographic area rather thanjas a reflection of the more generalized chances also ob-served elsewhere'.
. .,o -
6-If justified by the results of the routine " screening", special I 1
l in-depth epidemiologic studies will be developed.
, Data available in the system will be useful in planning and carrying out such special studies.
~
~
Health 's'tatus indices, such as the ratio of observed and expected morbidity and mortality measures will be developed for minor civil divi-
- ions. individually or as a pool. Control charts and confidence limits will be ' erived d in order to deterritin'e significant excesses of morbidity /
mortality in any particular area population. Due to the spurious nature of adjusted morbidity / mortality rates sometimes obtained for small popu-lation groups by the direct method, the indirect method of standardiza-tion will be used then indicated, to adjust the measures with respect to factors such as age and sex and to derjve expected numbers and rates.
The various:statisticals enalyses will involve the use of statistical tools ranging from simple univariate measures of central tendency to graphi s, including tests of significance (differences in health indices between copulations or changes in he41th indices over time) as well as multivariate regression and correlation analyses.
VI.,0PERATIONAL PROCEDURES
,-The Division of Epidemiology Researcli and the State Health Data Center will develop the mechanism to as;ure the orderly consolidation and coordin-ation of routinely collected vital statistics, population and morbidity data. Tht! ;urveillance data will be computerized and the automated data s . . -
base.will be updated by the State Health Data Center annually or at other timeintefvals,asneeded. The Division of Epidemiology Research will be responsible for-data analys~is end epidemiologic interpretation. There will be no routine reporting of the results of " screening" analyses. Such findings will bf used solely for the purpose of internal management and 3
recommendation for further action. .
s , ,
. e. ,
. VII. DATA COMPONENTS
- 1. Natality and Mortality Data Component:
"The vital statistics data to be utilized in the System will be limited to special computer files (" minimum data set") containing selected data elements stripped from the Health Data Center's com-puterized birth and death statistics files. This newly created computer file will cover the period 1970-present and contain only data for minor civil divisions within 20 miles of the five surveil-lance areas and selected control sites in Pennsylvania. Because of confidentiality and other constraining factors, the file will not contain identifying information, and direct access will be limited to specified surveillance system staff.
The " minimum data set" will include information on live births, fetal deaths, neonatal deaths, infant deaths, congenital anomalies, prematurity, low birth weight, Apgar scores, and selected causes of death including malignant neoplasms. It will also include such basic data elements as sex, race, date of death, age at death, place of residence (MCD), place of death, cause of death (underlying to 1979, underlying and contributory,1979-present). The data files of the system will be updated with current data by the Health Data Center annually or at other time intervals as needed.
- 2. Morbidity Data Comoonent:
This component will be based on routinely collected morbidit'y data that are readily available by minor civil division of residence.
One of the two computerized morbidity data systems which hold most promise at present is the Pennsylvania Cancer Reoistry, the other is a hospital discharge system of the H_osDital Utilization Project (HUP).
- ~
. .. , 8-Surveillance System staff in cooperation with the Division of Chronic Diseases and the State Health Data Center will develop the operational procedures needed to insure the availability of the most recent Cancer Registry data for surveillance purposes.
The minimum morbidity data set would include discharge date and status, age, sex, race primary diagnosis, other diagnoses, place of residence (MCD) and hospital of discharge. Due to limita-tions in quality and availability, these HUP data will be collected only for the years from 1978 onwards. This minimum data set will be incorporated after the automated surveillance system has been fully developed with the available data-within the State Health Depa rtment.
- 3. Population Data Component:
Baseline population data are essential for a variety of purposes, including the computing of rates. The population data to be utilized will be obtained from the Health Data Center (HDC) which is a census affiliate data center. Data for the years 1970 and 1980 will be used.
VIII. INTERAGENCY COORDINATION The Epidemiologic Surveillance System Around Nuclear Power Plants, which is being developed in Pennsylvania, is unique and the first of its
[ kind in the United States and abroad. The concept and approach of the System, which' evolved from the experiences of the Three Mile Island nuclear accident, have attracted some interests by nuclear industry, media and by
'other states. The System is designed to serve a vital function of the State
, public health agency.
The importance of the,Epidemiologic Surveillance System being discussed here is ' closely related to any future nuclear accident or other environmental health hazards in the Commonwealth which will be monitored and regulated by both state and federal agencies. For this reason, there should be adequate coordination with at least three major agencies including Pennsylvania Emergency Management Agency, Bureau of Radiation Protection of DER, and the U.S. Nuclear Regul5 tory Commission. Steps are be.ing taken to accomplish this objective.
. . LT.10 EPIDEMIOLOGIC SURVEILLANCE IN PENNSYLVANIA:
A CASE OF NUCLEAR POWER PLANTS George K. Tokuhata, Dr.P.H., Ph.D.
Director Division of Epidemiology Research Pennsylvania Department of Health June, 1984 N
s l
EPIDElil0 LOGIC SURVEILLANCE IN PENNSYLVANIA: A CASE OF NUCLEAR P
!. INTRODUCTION
.The need for monitoring and evaluating the health status of populations was clearly indicated by the accident at Three Mile Island on March 28, 1979.
Had the accident been more serious in a health threatening sense, the infor-mational needs of involved government agencies and a concerned public could not have been adequately served in a timely manner. In April, 1980. the Secretary of Health issued a memorandum to the Three Mile Island (TMI)
Advisory Panel for Health Research Studies regarding the "Need for Baseline Data Around Nuclear Power Facilities." The Panel endorsed the concept of improved epidemiologic surveillance in general and approved the project for development by the Division of Epidemiology Research.
Several important characteristics of the Epidemiologic Surveillance System should be emphasized. First. the primary intent is to establish an automated method of compiling baseline health data for specific geographic areas which can be updated and monitored routinely. Second, the System is designed as a
- screening" rather than " diagnostic" device; that is, signif-icant changes or differences in health status based on the data analyzed will be regarded only as " potential warning signals." If and when such signals are detected in a population, further " diagnostic" study would be necessary to substantiate the magnitude of the problem and to identify specific cause(s). Third, the Surveillance System. of itself, is not intended to test any specific hypotheses or to determine health effects of radiation as such or any other specific health hazards. but rather to detect significant changes within or differences from norms in any health indicators under consideration that can be determined with the available data. Fourth, the System would provide comparative data which would be
useful in the continuing effort to assess the potential health effects of the TMI accident. The public demand for health information in the TMI area is beginning to rise again as the length of time since the acci-
-dent approaches the lower limits of the estimated latency periods of cer-tain cancers and many local residents believe that they had been exposed to large doses of radiation. Furthermore, the System would provide health information which would be invaluable in the event of another nuclear acci-dent or any other health threatening event.
II. PROJECT OBJECTIVES In view of practical considerations and immediate need, the objectives
-of this System are to:
(1) Develop and implement a baseline health information system which will insure the continuous and systematic compilation, analysis and interpretation of available demographic, natality, morbidity, and mortality data for minor civil divisions (MCDs) within 20 mile radii of the six nuclear plants (5 geographic areas) and selected control areas in Pennsylvania.
(2) Establish baseline health indices for each selected area. These baseline indices established during years of normal operation of nuclear plants can serve as " controls" against which updated indices can be compared in subsequent years. Health indices developed during pre-operational years (new plant sites) can also serve as
" controls" against which post-operational indices can be compared.
(3) Monitor and evaluate the health indices for study populations in the vicinity of nuclear facilities on a recular basis to detect any significant changes over time or differences from norms that may have occurred or been observed.
o
(4) ' Determine if special studies should be initiated to establish whether significant changes or differences in health indicators, if and when observed, are related to living in proximity to nuclear power plants or to other factors.
III. GE0 GRAPHIC AREAS The initial focus of the Epidemiologic Surveillance System will be on populations residing in minor civil divisions (MCDs) wholly or partially within 20 miles of the six nuclear power plants (commercially operational and under construction in Pennsylvania) and populations in control areas.
The nuclear power plant sites include (1) Beaver Valley 1 and 2 (Shippingport, Beaver County); (2) Shippingport (Shippingport, Beaver County; (3) Three Mile Island 1 and 2 (Goldsboro, York County); (4)
Susquehanna 1 and 2 (Berwick, Luzerne County); (5) Peach Bottom 2 and 3 (Peach Bottom Township, York County); and (6) Limerick 1 and 2 (Pottstown.
Montgomery County). The surveillance in these areas will cover a total of 565 minor civil divisions and about 2,970,000 people. One hundred and fifty-four MCDs with a population of 760,000 reside within 10 miles of the six power plants and an additional 411 MCDs with a population of 2,210,000 reside within a 10-20 mile ring of the plants. The communities '
(MCDs) selected for this surveillance program are included in the evacua-tion plans developed by the Pennsylvania Emergency Management Agency (PEMA)
~
and the " Plume Exposure Pathway EPZ" plans available for each nuclear plant site.
The control areas selected for comparative purposes, collectively provide a social-industrial-geographical representation of Pennsylvania:
except that there are no nuclear power plants within 20 miles of the perimeters of any of the areas.
IV. HEALTH STATUS MEASURES (INDICES)
The need for a minimum data set is recognized. It should be noted that certain health indicators or indices (e.g., cancer) have a long latency period (interval between exposure and diagnosis), while others (e.g. , fetal mortality) may have a short latency and effects, if any, can be shown within a year or less.
First, the short-term health indicators included in the minimum data set are largely derived from birth, fetal, and infant mortality data. These include:
(a) Fetal mortality (with or without abortions of 16+ week gestation).
(b) Hebdomadal mortality (during first 7 days of life).
-(c) Neonatal mortality (during first 28 days of life).
(d) Perinatal mortality (combination of fetal and neonatal mortalities).
(e) Infant mortality (during first 365 days of life).
(f) Congenital malformations, including neonatal hypothyroidism.
(g)_ Prematurity (gestation less than 37 weeks).
(h) Low birth weight (birth weight less than 2500 grams).
(1) Low Apgar Score (less than 7 at one minute).
Second, with respect to _1_ono-term health indicators, data on cancer incidence, prevalence, and mortality of the following organ sites are-important.
(a) Leukemia (g) Lymphatic system (b) Thyroid (h) Bladder (urinary) '
c) Breast (i) Kidney d) Skin _ (j ) Pancreas e) Lung and bronchus (k) Stomach (f) Bone (1) Esophanus ,
Third, thyroid disease (other than congenital and cancer) will be studied as intermediate-term indicators, particularly in relation to the-release of radioactive iodine into the environment which can be ingested or inhaled by local residents.
Fourth, in view of the possible effects on oeneral life expectancy or longevity of a total population, abridaed life tables will be computed periodically using annual mortality data and base populations by age and sex.
5-V. METHOD OF ANALYSIS In general, the methods of analysis to be utilized will be determined by general surveillance needs and the requirements of particular studies.
When the baseline morbidity and mortality data are established, a variety of routine epidemiologic analyses will be carried out. Demographic differ-ences in the populations under consideration which could bias results will be adjusted for wherever possible in all analyses.
First, cross-sectional analyses: Various health status measures will be compared between the nuclear power plant (NPP) and control areas for selected years. Such measures will be based on area-specific mortality and/or morbidity data.
Second, temporal analyses: Changes over time in health status indi-cators of populations residing within each of the specified NPP and control areas will be assessed. Specifically, when the baseline health experience of the population around a given NPP is established, the health experience of that particular population can be compared from one time period to another. Such historical analyses have an advantage over cross-sectional comparisons in that much of the basic environmental characteristics (physical, socioeconomic, cultural, etc.) tends to remain fairly stable. On the other hand, historical analyses will permit the relating of significant changes in health status, if they occurred, to recognizable socioeconomic changes, such as the establishment of a .1ew industry with subsequent influx of workers in certain age groups into the area communities.
t Third, temporal-spatial analyses: This is a combination of geographic '
comparisons with historical comparisons. For example, time trends in mor-tality from specific cancers in the NPP and control areas will be evaluated in reference to each other. This will permit aseessment of significant chances in cancer death rates which may be limited to a particular geographic area rather than as a reflection of the more generalized chanaes also ob-served elsewhere.
If justified by the results.of the routine " screening", special in-depth epi'emiologic d studies will be developed. Data available in the system will be useful in planning and carrying out such special.
studies.
Health status indices, such as the ratio of observed and expected morbidity and mortality measures will be developed for minor civil divi-sions individually or as a pool. Control charts and confidence limits will be derived in, order to determine significant excesses of morbidity /
mortality in any particular area population. Due to the spurious nature- '
of adjusted morbidity / mortality rates sometimes obtained for small popu-lation groups by the direct method, the indirect method of standardiza-tion will be used when indicated, to adjust the measures with respect to factors such as age and sex and to derive expected numbers and rates.
The various statistical analyses will involve the use of statistical tools ranging from simple univariate measures of central tendency to graphics, including tests of significance (differences-in health indices between populations or changes in health indices over time) as well as multivariate regression and correlation analyses.
VI. OPERATIONAL PROCEDURES The Division of Epidemiology Research and the State Health Data Center will develop the mechanism to assure the orderly consolidation and coordin-ation of routinely collected vital statistics, population and morbidity data. The surveillance data will be computerized and the automated data base will be updated by the State Health Data Center annually or at other time intervals as needed. The Division of Epidemiology Research will be responsible for data analysis and epidemiologic interpretation. There will be no routine reporting of the results of " screening" analyses. Such findings will be used solely for the purpose of internal management and recommendation for further action.
g- ,
. . - ,}
VII. DATA COMP 0NENTS
- 1. Natality and Mortality Data Component:
The vital statistics data to be utilized in the System will be limited to special computer files (" minimum data set") containing selected data elements stripped from the Health Data Center's com-puterized birth and death statistics files. This newly created computer file will cover the period 1970-present and contain only data for minor civil divisions within 20 miles of the five surveil-lance areas and selected control sites in Pennsylvania. Because of confidentiality and other constraining factors, the file will not contain identifying information, and direct access will be limited to specified surveillance system staff.
The " minimum data set" will include information on live births,
. fetal deaths, neonatal deaths, infant deaths, congenital anomalies, prematurity, low birth weight, Apgar scores, and selected causes of death including malignant neoplasms. It will also include such basic data elements as sex, race, date of death, age at death, place of j
residence (MCD), place of death, cause of death (underlying to 1979, ,
underlying and contributory, 1979-present). The data files of the system will be updated with current data by the Health Data Center annually or at other time intervals as needed.
- 2. !1orbidity Data Component:
This component will be based on routinely collected morbidity data that are readily available by minor civil division of residence.
One of. the two computerized morbidity data systems which hold most promise at present is the Pennsylvania Cancer Reqistry, the other is a hospital discharge system of the Hospital Utilization Project (HUP).
u
, , r. ,
Surveillance System staff in cooperation with the Division of Chronic Diseases and the State Health Data Center will develop the operational procedures needed to insure the availability of the most recent Cancer Registry data for surveillance purposes.
The minimum morbidity data set would include discharge date and status, age, sex, race, primary diagnosis, other diagnoses, place of residence (MCD) and hospital of discharge. Due to limita-tions in quality and availability, these HUP data will be collected only for the years from 1978 onwards. This minimum data set will be incorporated after the automated surveillance system has been fully developed with the available data within the State Health Depa rtment.
- 3. Population Data Component:
Baseline population data are essential for a variety of purposes, including the computing of rates. The population data to be utilized will be obtained from the Health Data Center (HDC) which is a census affiliate data center. Data for the years 1970 and 1980 will be used.
VIII. INTE_RAGENCY C0 ORDINATION The Epidemiologic Surveillance System Around Nuclear Power Plants, which is being developed in Pennsylvania, is unique and the first of its kind in the United States and abroad. The concept and approach of the System, which evolved from the experiences of the Three Mile Island nuclear accident, have attracted some interests by nuclear industry, media and by other states. The System is designed to serve a vital function of the State public health agency.
The importance of the Epidemiologic Surveillance System beina discussed here is closely related to any future nuclear accident or other environmental health hazards in the Commonwealth which will be monitored and regulated by both state and federal agencies. For this reason, there should be adequate coordination with at least three major agencies including Pennsylvania Emerqency Management Agency, Bureau of Radiation Protection of DER, and the U.S. Nuclear Regulatory Commission. Steps are being taken to accomplish this objective.