|
Items in cart [4] |
Questions? Call 888-624-8373 |
| Copyright © 2009. National Academy of Sciences. All rights reserved. Terms of Use and Privacy Statement |
Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter.
Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 17
Executive Summary
BACKGROUND
It is well recognized that iodine-13 l, or 13~] (a
radioactive form of iodine), is an important radionuclide because
of the potential for human exposure to it after accidental releases
from nuclear reactors and fuel-reprocessing plants. Deposition of
|3~} onto pasture grass leads to contamination of cows' milk and
ingestion of the radioactivity by humans. Because iodine is
concentrated in the thyroid, radiation doses to the thyroid can
result. The Draft Final Report of the Hanford Thyroid Disease
Study (HTDS) describes a study of the cumulative incidence of
thyroid disease and abnormalities among "downwinder" children
exposed to ]3~{ from the Hanford Atomic Products Operations.
Releases of 13~{ began in December 1944 as a consequence of the
chemical removal of plutonium from the fuel rods irradiated at the
Hanford nuclear site.
The main study objective of the HTDS is described in
the Draft Final Report as a "determination of whether thyroid
morbidity is increased among persons exposed to releases of
radioactive iodine from the Hanford nuclear site." l:n the study,
3,441 subjects who had been born near Hanford in 1940-1946 were
contacted in the 1990s and taken to several locations for medical
examination for thyroid disease. A detailed dose-reconstruction
method developed by the Pacific Northwest National Laboratory
was used to assign likely thyroid doses to study participants.
Estimating dose also entailed querying parents about study
participants' residence and milk-consumption history in 1944-
1957, the period of idol releases.
17
OCR for page 18
18
Review of the HTDS Draft Final Report
The Centers for Disease Control and Prevention (CDC)
asked the National Academy of Sciences-National Research
Council (NAS-NRC) to give an independent appraisal of the study
methodology, results, and interpretation and of the communication
of the study results to the public. Specifically, it asked:
· Has the analysis been carried out appropriately and
completely?
· Are the presentation and the discussion of results
complete?
· Are the conclusions reasonable?
· Was the material accurate and appropriate in providing
guidance to the public in understanding the study findings?
· if these messages about findings need to be amended,
how should the revised messages best be communicated to the
public?
With regard to release of future study reports, how can
CDC improve the public communication process?
This report constitutes the response of the NRC
subcommittee to that request. To respond to the charge, the NRC
subcommittee felt that it needed to go beyond the specific
questions addressed to it by CDC and develop a broad
understanding and critique of the HTDS and the Draft Final
Report. As part of those activities, the subcommittee solicited
comments from outside experts and members of the public
primarily in a public meeting held in Spokane, Washington, in
June 1999, where 14 scientists and members of the public made
formal presentations to the subcommittee about various aspects of
OCR for page 19
Executive Summary
19
the Draft Final Report. Other members of the public also spoke
during four open-comment sessions at the meeting. in addition,
efforts were made to evaluate all information materials prepared
for the public and additional CDC communication plans.
Information was gathered through interviews with journalists,
members of concerned citizen groups in the Hanford region,
members of the CDC scientific and media staff in Atlanta, and the
HTDS investigators.
in this summary, the main points follow the structure
of our report and are presented under several headings:
epidemiologic and clinical methods and data collection, dosimetry,
statistical analyses, statistical power and interpretation of the study,
and communication of the study results to the public. We then
provide a brief synopsis of our response to the questions raised by
CDC.
EPIDEMIOLOGIC AND CLINICAL METHODS AND DATA COLLECTION
The HTDS eligibility criteria called for including all
persons born in the early 1940s in the counties that were predicted
to have the highest exposures to Hanford releases (Benson,
Franklin, and Adams counties) and randomly selected subjects
born in the same period in four counties that were expected to have
intermediate exposure (Walla WalIa County) or low exposures
(Ferry, Stevens, and Okanogan counties). Attempts were made to
determine the vital status of all 5,199 eligible potential participants
and, if they were alive, to trace and enroll them in the study. A
total of 3,441 people received thyroid examinations; thyroid doses
could be estimated for 3, 190 of them.
The subcommittee considered the study design to be
generally appropriate to address the aims of the study. The
investigators also chose the best population to study, namely, those
in the most highly exposed areas who were young children at the
time ofthe greatest AT releases. The low-exposure group, up wind
and more distant from the Hanford site, was also a reasonable
choice. However, the one significant weakness of the design was
OCR for page 20
20
Review of the HTDS Draft Final Report
the uncertainty in thyroid dose estimates, with resulting potential
for exposure misclassification.
The epidemiologic methods were exceptionally good.
The sample was based on an almost complete census of eligible
subjects, and location and participation rates were high. There was
a high level of quality control in the epidemiologic procedures, and
interviews and clinical examinations were "blinded" (without
knowledge of exposure or disease status). The HTDS collected
data on an appropriate set of potential confounding variables (risk
factors that might distort or mask findings), including sex; age at
first exposure to ill] Tom Hanford; age at examination; history of
diagnostic, therapeutic, and occupational radiation exposures
smoking history; and ethnicity.
The questionnaire to elicit information from parents or
surrogates on the participants' milk-consumption patterns was
carefully designed and field-tested. However, there is substantial
inherent unreliability in recall of dietary habits of 40-50 years ago,
especially when someone other than the mother was interviewed,
as was the case in 26% of the interviews. Compounding that
problem is that for 38°/0 of the participants no parent or surrogate
was available to be interviewed.
Generally, the clinical examinations and laboratory
studies were performed with good-quality methods. Subjects were
given physical examinations, including thyroid palpation by
thyroid specialists, ultrasonography, and appropriate thyroid-
hormone and thyroid-antibody blood tests. The subcommittee has
mostly minor criticisms of the clinical and laboratory procedures
pertaining to inadequate quality-control procedures in assessing
and reporting cytopathology results and the possibility that some
past thyroid diagnoses were missed, inasmuch as it was impossible
to obtain 7°/0 of the death certificates and 37°/0 of the requested
historical medical records.
The HTDS investigators provided the subcommittee
with additional tabulations for examining deaths in the group
studied. Although there was a small increase in mortality, mostly
OCR for page 21
Executive Summary
21
due to perinatal mortality and congenital anomalies, the tabulations
indicate that the increase was not due to AT exposure, in that it
occurred both before and during the time of the 13~{ releases.
However, a more detailed tabulation of perinatal deaths and
congenital anomalies should be provided to help readers to
interpret these data.
Conclusions:
The study design was generally appropriate.
· An optimal study population was chosen: the most
highly exposed young children.
The epidemiologic methods were of excellent quality.
The clinical and laboratory methods were appropriate
and generally had good quality control.
Some past thyroid diagnoses might have been missed
because medical records and pathology slides were
unobtainable.
Recommendations:
needed.
An adequate review of the cytopathology results is
· The HTDS investigators should indicate for how many
potential past thyroid diagnoses they were unable to obtain any
medical confirmation, with a breakdown by reported type of
thyroid disease and dose.
detail.
. The mortality experience should be tabulated in more
OCR for page 22
22
DOSIMETRY
Review of the HTDS Draft Final Report
The HTDS relied heavily on the Hanford
Environmental Dose Reconstruction (HEDR) method, which in
turn relied heavily on the use of environmental-transfer models.
Models were necessary to estimate movement of IT in the
environment because very few measurements of IT from the
1940s were available. The estimates of the thyroid doses for the
3,190 participants ranged from of O.OOOS mGy to 2,842 mGy, with
a mean of T82 mGy. The HEDR models have been subjected to
numerous reviews and to independent testing. The NRC
subcommittee found the dose assessment, on the whole, to be
structurally sound for the estimation of thyroid doses, but minor
errors have been found and doubts have been raised about the
validity of some assumptions and of results for some
environmental conditions. A review of some of the key parameter
values that went into the HTDS dosimetric mode} showed that
most of them were reasonable, and the resulting dose estimates are
generally supported, at least to within a factor of 2 or 3, by the
validation studies performed by the HEDR project.
Recently, several scientists have asserted that the source
term (the amount of ill{ released by Hanford) was underestimated.
However, even if those scientists' points are all valid, we estimate
that it would increase the total ]3~{ releases during 1944-1957 by
only about 30°/O. Systematic dose underestimation has implications
for the statistical power of the study. if the doses were
underestimated across the board, the study would have greater
statistical power than was projected, in which case negative results
of the study would be more persuasive. However, if there was
variation by subject in the degree of dose underestimation, or in the
degree to which subjects vary in sensitivity (because of age
differences, and so on), these could result in reduced statistical
power.
The subcommittee believes that the doses might have
been overestimated, in that the HTDS used an estimate of the
fraction of ill] eaten by a cow that is transferred to milk that was
OCR for page 23
Executive Summary
23
about twice as high as estimates in other studies. If the doses were
overestimated, the statistical power of the study would be less than
was stated in the HTDS Draft Final Report. Revisions of the mode}
and recalculation of doses would be required to determine the net
effect of those factors.
A critical feature of the dosimetry system is that the
dose estimates have large uncertainties because they are based on
mathematical models, not direct measurements. It is very likely
that the uncertainties were underestimated by the HTDS because
some sources of uncertainty were not taken into account. A notable
deficiency was in accounting for the uncertainties in the sources
and amount of milk consumption reported by parents or their
surrogates. Those reports were 40-50 years after the fact, so one
would expect appreciable
unreliability in recall of milk-
consumption patterns. The uncertainties would lessen the statistical
power of the study and thereby make its results less definitive.
During the 1950s and early 1960s, the Hanford
population was also exposed to 13~{ from Nevada Test Site CATS)
fallout and global fallout from atomic and hydrogen-bomb tests
over the Pacific, in the Soviet Union, and elsewhere. The HTDS
team performed some analysis of the impact of NTS fallout, but
not of global fallout. A rough assessment of the thyroid doses
arising Dom global fallout was performed as part of our review.
The subcommittee concludes that the variability in thyroid doses
from those sources is much less than that in doses from Hanford
fallout in the population being studied, so NTS and global fallout
are probably not important confounders of the Hanford dose-
response associations. Nevertheless, they should be examined
carefully to be sure.
Conclusions:
· The dose-assessment methods and their implementation
are difficult to review because the information is scattered
among numerous documents.
OCR for page 24
24
Review of the HTDS Draft Final Report
· Dose estimates appear accurate to the degree one
normally expects for environmental-dose reconstruction but
minor errors in the parameter values used in the model need to
be corrected.
· Dose uncertainties were underestimated because errors
were not included for all the factors in the dosimetry model.
This conclusion is the same as stated in a previous National
Research Council (1999) letter report that reviewed the
analysis plan for the HTDS.
Recommendations:
· A single document describing clearly the HEDR dose-
assessment methodology, including uncertainties, and its
implementation by HTDS should be prepared.
· Errors in the dose-estimation mode} should be
corrected.
account.
All dose-related uncertainties should be taken into
STATISTICAL ANALYSES
The observed number of cases of thyroid cancer, the
thyroid-disease end point of greatest interest, was very small: 20,
of which only 14 had dose estimates. That makes it difficult to
perform a meaningful statistical evaluation. For most other thyroid
diseases, the numbers were more substantial, such as 250 cases of
benign thyroid nodules.
The HTDS used the HEDR thyroid doses as its only, or
at least primary, tool to describe patterns of likely exposure. The
subcommittee believes that the HEDR dosimetry should not have
been the sole method for evaluating the association between AT
exposure and thyroid disease, and it suggests that supplemental
OCR for page 25
Executive Summary
25
analyses be done that could help to confirm (or refute) the dose-
response analyses. The dose-response results given are difficult to
interpret; all one sees is dose-response regression coefficients with
no tabulations to help in interpreting the factors that influence the
results.
The subcommittee recommends that the following
tables, which were absent from the Draft Final Report, be included
in the revised report: frequency distribution of individual doses,
observed and expected numbers of disease or abnormality
outcomes in several dose categories, and average doses according
to such important categories as year of birth and milk consumption
in childhood. Other potential risk factors were evaluated as
possible confounding or effect-modifying variables, but no tables
were presented to show the results of those evaluations.
The HTDS investigators assigned thyroid doses only
for periods when subjects lived as children in the geographic area
for which exposures were estimated. They made no attempt to
estimate out-of-area doses for persons who were out of the area for
part of the exposure period and to perfonn sensitivity analyses to
determine the impact of the missing doses. (See chapter 5.)
Given that thyroid-disease and thyroid-abnormality
rates appeared to differ by geographic area, the subcommittee
recommends alternative analyses to address the issue. A set of
analyses stratifying on geographic area is needed because the
HTDS investigators' tabulations showed that the rates tended to be
higher in areas with low fallout, so the geographic variations due to
factors other than dose would induce a negative association
between ]3~} and thyroid-disease rates. That feature of the data
could explain why most of the dose-response estimates were in the
negative direction, the question is whether removing its influence
would yield a positive association.
A concern expressed by members of the public is that a
study in which everyone is exposed is not valid and that a
completely unexposed "control" group is needed for proper
assessment of the risk associated with Hanford 13lI fallout.
OCR for page 26
26
.
Review of the HTDS Draft Final Report
However, for several reasons, the subcommittee believes that the
HTDS investigators were correct in emphasizing dose-response
comparisons in the study rather than comparisons with other
general populations. The slope of the dose-response curve can
provide a valid index of the risk even without an unexposed
contra group, as long as a sufficient range of doses are estimated
with reasonable accuracy. Comparisons with an external, general
population are problematic on several accounts. Persons living in
various geographic areas often vary in their baseline risk of thyroid
diseases because of differences in dietary iodine intake and other
unknown factors. The rates of detected disease in the HTDS are
based on thyroid examinations and depend on the methods and
criteria of those examinations. That often produces a large
screening effect (detection of cases of disease that otherwise would
not have been detected until some years later, if at all), so
comparisons with rates from other geographic regions without
comparable screening are not valid. Conclusions drawn from
comparisons with general-population prevalence have more
potential for bias than those drawn from dose-response
comparisons in the study population.
The HTDS investigators analyzed the effect of
interview versus default data for milk consumption, although
results were not presented. The subcommittee suggests that they
also examine associations, using only those with interview
information to minimize dose misclassification. Another suggested
alternative is to conduct dose-response analyses that stratify on
interview versus default milk values.
Tabulation of thyroid-disease rates by reported milk-
drinking habits is suggested to elucidate further whether 13~{
exposure estimates tend to coincide with disease. The fact that
higher thyroid morbidity was found in the less-exposed counties
argues that the pattern of thyroid morbidity did not tend to track
the likely geographic pattern of exposure to its. However, birth
location was not as important a determiner of thyroid dose as were
OCR for page 27
Executive Summary
27
behavioral habits (as can be seen by the large range of doses in all
counties), primarily milk-drinking.
Besides the lack of some important tables in the
analysis, the subcommittee has concerns about the manner of
presentation of the study results. The results were reported in black
and white terms: a statistical test did or did not reject the null
hypothesis. No confidence intervals on the estimates of the size of
effects were given. The subcommittee recommends that confidence
levels be provided throughout the report. Furthermore, had the
investigators presented confidence intervals for the dose-response
mode! on the basis of their statistics, the confidence intervals
would almost certainly have been too small because dosimetry
errors were ignored. A sophisticated statistical method described in
the statistical chapter of the report would have at least partly taken
dose uncertainties into account in the confidence intervals, but it
was not implemented.
Conclusions:
· The HTDS report relies too heavily on dose-response
analyses without providing sufficient associated evidence from
tabulations of factors that could illuminate the results obtained.
(See chapter 5.)
A number of key tables were absent from the report, for
example, tables of frequency distribution of doses; of observed
and expected frequencies of each thyroid disease by, say,
quartiles of dose; of thyroid disease rates by milk-drinking
habits and other risk factors in disease; and of average doses by
year of birth, amount of milk consumption in childhood, and
the like.
· Differences in thyroid-disease rates by geographic area
(called "geostrata" in the Draft Final Report) might have been
an important confounder of the dose-response association.
OCR for page 29
Executive Summary
STATISTICAL POWER AND INTERPRETATION OF THE STUDY
29
The HTDS investigators were generally successful in
achieving the sample size and dose distribution that they had
projected as necessary if the study were to have adequate statistical
power. The Draft Final Report indicates that for many of or all the
thyroid end points the estimated power of the study was good (over
90°/O) for detecting plausible nonzero dose-response relationships.
Nevertheless, the HTDS investigators' discussion of statistical
power did not present how small the expected excess of thyroid
cancers was. Based on their assumptions about the risk coefficient,
the dose distribution, the number of persons in the study and the
length of the follow-up, about 34 thyroid cancers were expected in
the study, of which 19 would have been expected without any 13~{
exposure and ~ 5 were due to radiation exposure. Had these
numbers been presented, they might have tempered CDC's
evaluation of the scientific value of the study in relation to its cost.
Our subcommittee reviewed the factors that influence power,
focusing on the impact of uncertainties in the HEDR thyroid doses.
We note that neither the power calculations nor analytic techniques
used made explicit allowance for uncertainties in the dose
estimates.
The investigators ignored five sources of uncertainty.
First, in making their projections they assumed that the dose-
measurement error was all of a type ("Berkson error") that would
not reduce the strength of associations. That is different from
classical measurement error, which does weaken the strength of
associations and therefore requires a larger sample or a wider dose
range to attain adequate statistical power. [ndividual-based dose
error for example, the uncertainty of a subject's milk-drinking
habits represents classical measurement error that needs to be
taken into account in estimating the statistical power of a study.
Second, multiplicative (as opposed to additive) dose uncertainties
(such as errors in the source term or in the coefficient of transfer
from cow intake to milk) that apply to everyone will add error to
risk estimates and thereby decrease statistical power. Third,
OCR for page 30
30
Review of the HTDS Draft Final Report
correlations among multiplicative dose uncertainties can further
reduce statistical power. Fourth, geographic variations in the
baseline rates of disease can decrease statistical power if not
controlled for. Fifth, some sources of uncertainty in the 13~}
environmental pathway apparently were not included in the
dosimetry uncertainty estimates.
Because those types of uncertainty were not taken into
account in the statistical-power calculations, the subcommittee
believes that the HTDS projections of statistical power are
overestimated, perhaps substantially. The negative results of the
study are therefore less definitive than the Draft Final Report and
press releases stated.
The uncertainties listed above also have an impact on
the width of confidence intervals around the estimates of thyroid-
disease effects, so the study is less clearly negative than was
portrayed. The confidence intervals would be more compatible
with (although the best estimate does not support) the larger risks
seen in other studies, such as the Utah NTS fallout study and the
large pooled study of thyroid-cancer risk associated with external
radiation exposure.
Conclusions:
· The HTDS investigators were successful in achieving
the sample size and dose distnbution that they projected as
necessary if the study were to have adequate statistical power.
· However, the HTDS assumptions regarding statistical
power did not include the possibility that dose uncertainty
would weaken the associations. They ignored several sources
of uncertainty that probably decrease the statistical power of
the study.
· The subcommittee believes that the HTDS projections
of statistical power are overestimated, perhaps substantially.
OCR for page 31
Executive Summary
31
The negative results of the study are therefore less definitive
than the Draft Final Report and press releases stated.
· The HTDS results are probably compatible with the risk
estimates from the Utah NTS fallout study, because the various
uncertainties would yield wide confidence intervals.
Recommendations:
· The HTDS investigators should describe the sources of
uncertainty in as quantitative terms as possible and interpret
their results in the light of these uncertainties.
· The HTDS investigators should recalculate the
statistical power of the study, taking into account the dose
uncertainties if this proves feasible.
· The compatibility of the HTDS study with other studies
of radiation and thyroid disease should be re-examined, taking
into account the impact of dose uncertainties.
COMMUNICATION OF THE STUDY RESULTS TO THE PUBLIC
Compared with the history of a less-than-open public-
information policy of the Department of Energy and its
predecessor agencies, the early plans by CDC and the HTDS
investigators for open communication about the study were
enlightened and promising. So was the decision to establish a
citizen advisory group for the study and the apparent level of
cooperation offered to various other citizen groups in the region
over the years of the study. All those early efforts seemed to build
trust and credibility for the study.
However, when the Draft Final Report was released, a
number of communication errors were made that caused a public
outcry. The draft report outlined a good communication plan,
which included an admirable concern for translating the technical
information in the report into an understandable booklet for the
OCR for page 32
32
Review of the HTDS Draft Final Report
public, and included a Web site to share information with the
public. But several events forced an early release of the Draft Final
Report, which pre-empted the original communication plan.
The main message of the report was problematic. The
written materials and oral presentations made by HTDS
investigators overstated the certainty (the statistical power) of the
study and the conclusiveness of the negative findings. Although
the public materials factually represented what appeared in the
Draft Final Report, the strong statements made publicly were not
tempered by expressing the uncertainties.
In trying to decide how to present the study, CDC was
on the horns of a dilemma: some members of citizen groups had
urged agency personnel in advance not to alter the report before its
release, and CDC had to respect issues of academic freedom
regarding the investigators' views, but after the report's release,
they were blamed for not intervening to counter the strong message
delivered to the public by the HTDS investigators.
On the basis of comments received by the
subcommittee from members of the public, it is clear that some
people with an interest in the findings of the study were
disappointed with the reported (negative) results and upset by how
the results were disseminated and described. A number of factors
contributed to and complicated the problems surrounding the
report's release: an information blackout that included the citizen
groups, a complex briefing schedule by telephone in the Hanford
area to venous state health agencies and citizen organizations only
several hours before the media anal public briefings on the
findings, a leak to the New York Times that related the findings to
the public before most of the briefings in the Hanford area, and a
message that contradicted what most of the public thought would
be the outcome of the study.
Because serious problems were encountered in the
scheduling and conduct of the prerelease briefings, a different
briefing strategy should be used in the future, and telephone
OCR for page 33
Executive Summary
33
briefings should probably be abandoned because they were
disliked by all involved.
Delivering an unpopular message requires sensitivity to
the audience's health concerns and fears. But the media and public
briefings, and all written materials emphasized the overall
statistical results of the study and did not seriously discuss the
outcome for individuals. The implications for individuals and
families that have suffered because of thyroid disease could have
been explained in the written materials and public briefings.
The subcommittee applauds CDC's open-
communication policy and strongly recommends that this policy
continue with the HTDS and similar studies. it recommends that a
new communication plan be devised for the release of the final
HTDS report, taking into account the problems that have already
been encountered. The final report should outline and explain any
significant changes made in the Draft Final Report.
The subcommittee suggests that CDC convene a
workshop of risk-communication experts, scientists, journalists,
and citizens to discuss how to publicly release and discuss
controversial unreviewed draft reports more effectively and to
discuss other issues that could affect the future release of important
CDC reports.
Conclusions:
· The early enlightened plans by CDC and the HTDS
investigators for open communication about the study and for a
citizen advisory group for the study should have helped to
build trust and credibility.
· Early release of the Draft Final Report and public
concerns about CDC changes in that draft led to many of the
communication problems that resulted from the draft report's
release.
OCR for page 34
34
Review of the HTDS Draft Final Report
· The information blackout and complex briefing
schedule for release of the Draft Final Report worked against
effective communication of the report's findings to the public
and unnecessarily upset Hanford-area citizen groups that had
cooperated with the HTDS over the years.
· A
key weakness of the communication effort
surrounding the release of the Draft Final Report was that the
report and the public communications by HTDS investigators
overstated the certainty (the statistical power) of the study and
the conclusiveness of the negative findings and failed to
discuss the uncertainties. CDC officials should have expressed
their own interpretations in addition to those of the HTDS
investigators about the draft report in the briefings and public
documents.
Recommendations:
· Delivering an unpopular message requires sensitivity to
the audience's health concerns and fears. In communications
about the HTDS final report, implications for individuals and
families that have suffered because of thyroid disease should
be carefully explained. if there are plausible alternative
interpretations of the results, they should be acknowledged.
· The subcommittee supports CDC's open-
communication policy and strongly recommends that it
continue. it recommends that a new communication plan be
devised for the release of the final HTDS report and
accompanying public documents, taking into account the
problems that have already been encountered.
· In the HTDS final report and all public documents, any
significant changes made from the Draft Final Report should
be clearly outlined and explained, and all remaining
uncertaintites should be noted and explained.
OCR for page 35
Executive Summary
35
. Careful consideration should be given to how to release
controversial reports to the public more effectively. The
subcommittee suggests that CDC convene a workshop to
discuss this and other communication issues of concern.
SUMMARY OF RESPONSES TO THE CDC'S QUESTIONS
The subcommittee's responses to CDC's six questions
are summanzed below.
Question 1. Has the analysis been carried out appropriately and
completely?
Our overall assessment is that the epidemiologic and
clinical components of the study were of excellent quality,
including the study design, followup success, subject participation
rate, interviewing, thyroid examination, and laboratory methods.
The design of the dose-assessment mode! has been found, on the
whole, to be reasonably sound for the estimation of thyroid doses,
but several questionable assumptions have been identified that
would have some impact on the estimated individual doses. The
estimated-dose uncertainties that the HEDR project produced and
the HTDS study used are underestimates of the total dose
uncertainty because some significant sources of uncertainty were
overlooked.
The basic objective in the statistical analysis was to
determine whether there was an association between the
occurrence of various thyroid diseases and exposure to 13~}
released from Hanford. That was appropriately addressed by
modeling the relationship between the frequency of a thyroid
disease and dose, with consideration of an appropriate set of
potential confounding variables.
Several other analyses that were not presented could
have aided in the interpretation of the apparently negative results, a
OCR for page 36
36
Review of the HTDS Draft Final Report
critical one being an analysis of thyroid dose and disease with
apparent geographic variations in disease rates controlled for.
The investigators made no attempt to mode} the out-of-
area doses for persons who were included in the main analyses.
Their approach to that issue could have led to attenuated results in
that it potentially estimated the total fallout doses for some people
but only partial doses for others. The impact of global fallout on
variations in thyroid-disease risk should also be analyzed.
The subcommittee believes that the HTDS emphasis on
analyses of subjects in the study rather than on comparisons with
the general population is appropriate, inasmuch as the latter are
potentially subject to serious biases.
Question 2. Are the presentation and the discussion of results
complete?
One serious gap is that the methods used to calculate
doses and uncertainties are not clearly or fully described in the
dosimetry documents provided to the subcommittee.
A number of additional tables are needed. A tabular
presentation of the pathways to diagnosis would help readers to
assess how the final diagnoses were assigned. A table of the
frequency distribution of doses would be informative. Similarly,
tables that show observed and expected numbers of disease
outcomes according to four or five dose groups would normally be
expected. A description of the estimated dose distribution
according to such important categories as geostratum, year of birth,
and amount of milk consumption in childhood would be helpful.
.
The discussion of the results was substantially
incomplete in that little was said about whether the confidence
intervals were wide enough to be compatible with other parallel
studies. Most important, there was no adequate discussion of how
dosimetric uncertainties might have affected the confidence
intervals and the statistical power of the study.
OCR for page 37
Executive Summary
Question 3. Are the conclusions reasonable?
37
In concluding its review of the HTDS Draft Final
Report, the subcommittee considered the notion raised by the
public that the HTDS was inconclusive in its findings. The
subcommittee believes that the certainty of the interpretations from
a complex study, such as the HTDS, is always a matter of degree.
Its members believe that the high certainty with which the HTDS
investigators presented the negative findings of the draft report was
overoptimistic. Still, the main finding of the final HTDS report
might indeed be that no radiation effect could be observed: the lack
of evidence of a dose-response relationship for any type of
morbidity suggests that overall risks were unaffected by Hanford
releases. Given the substantial degree of imprecision in the
exposure estimates and the effect of other statistical issues, the
absence of any observable radiation effect does not rule out the
possibility that a small effect exists, although it does mean that
large effects of the 13~} exposure can be excluded as incompatible
with the data. Until estimates are given with appropriate
confidence limits, we will not know how much risk to the thyroid
is compatible with the data. The evidence does not rule out
(although it provides no particular support for) the possibility of a
weak association that could affect, for example, those already
susceptible to thyroid disease because of predisposing genetic
factors.
This carefully designed study, with sound followup and
medical methods, has examined a substantial fraction of the most
highly exposed population and failed to find any obvious evidence
of a radiation effect; that is, there was no evidence of abnormally
high rates of thyroid disease in the Hanford "do~vnwinders"
examined who had the largest estimated exposures. Thus, at face
value, the study was negative and found no increased risk. The
pattern of individual exposure estimates is in accord with such
basic factors as the prevailing wind direction and distance from the
Hanford site. Finding negative results of both geographic and
OCR for page 38
38
Review of the HTDS Draft Final Report
exposure comparisons implies that AT had no strong impact on
thyroid disease. If a similar exposure occurred elsewhere, one
could not predict the results with confidence. The small numbers
of thyroid-cancer cases and the lack of precision in estimating
individual exposures mean that one can have little confidence in
the risk estimates found in the HTDS. At the various public-
comment meetings, people who lived in down-wind areas stated
that they and their families experienced more thyroid disease than
would have been expected in the population at large. That could be
due to genetic factors in families or even to chance, but the
possibility that their disease was the result of unusual fallout or
ingestion patterns or of unusual susceptibility to a thyroid radiation
effect cannot be excluded.
Question 4. Was the material accurate and appropriate in providing
guidance to the public in understanding the study f n clings?
For the most part, the written and oral messages about
the Draft Final Report were accurate, but they were occasionally
misleading in that they included statements that were too strongly
worded, given the uncertainties that applied to the study. Keeping
the study process and activities as "transparent" that is, open
visible and understandable as possible for the public is a valuable
approach that should not be abandoned because of the problems
encountered with the release of the Draft Final Report.
Question 5. If these messages need, to be am encled, how should
the revised messages best be communicated to the public?
Given all the communication problems that resulted
Dom the release of the Draft Final Report, another detailed
communication plan needs to be drawn up for release of the final
report, including planning for unanticipated situations. Messages
must take into account the various audiences being addressed and
OCR for page 39
Executive Summary
39
show concern and sensitivity for the thyroid-health issues that
people perceive affect them. The full picture of the study results
should be given, including all the uncertainties and other problems.
A plan to brief the active citizen groups should be developed so
that they have enough information to be able to respond to media
inquiries about the report. In addition, an embargoed release of the
report to journalists should be used so that they have a few days to
read through the report and develop informed questions before the
briefing.
Question 6. With regard to release of future study reports, how can
CDC improve the public communication process?
The briefing structure should be simplified to try to
eliminate leaks, and citizens who have participated in the advisory
process all along should be given higher priority in the briefing
structure.
It could prove helpful to CDC to conduct a workshop
on improving the public-communication process that includes
experts in risk communication, journalists, outside scientists, and
members of citizen groups.
Representative terms from entire chapter:
statistical power
