Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
9 Summary of Responses to CDC's Questions CDC asked the National Research Council subcommittee the following six questions regarding the HTDS, the Draft Final Report and related materials and events: . Has the analysis been earned 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 chapter presents the subcomittee's answers to those specific questions. 137
138 Review of the HTDS Draft Final Report HAS THE ANALYSIS BEEN CARRIED OUT APPROPRIATELY AND COMPLETELY? The quality of the data used in the analysis is as important as the analysis itself. Our overall assessment is that the epidemiologic design was of excellent quality. The sample was based on an almost complete census of eligible subjects born in the selected years and geographic regions. Efforts to solicit study participation were thorough and appeared comparable for those with high and low doses; as a result, they achieved similar participation rates across the dose range. The HTDS was successful in obtaining subjects willing to participate, and it essentially met its goals with regard to sample size and thyroid- dose distribution. Care was taken to maintain blinding wherever possible In the study to minimize the potential for selection bias, interviewer-induced response bias, and clinical-examination bias. The clinical examinations and laboratory studies were performed with modern methods of detecting and defining thyroid disease, although more quality-assurance procedures with regard to the cytopathology data would have been desirable. In short, the epidemiology and clinical parts of the HTDS were designed and conducted with great care. The study appears to compare extremely well with other epidemiologic studies in those respects. The subcommittee believes that the methodology used by the HEDR project to estimate thyroid doses and their uncertainties is structurally sound. The HEDR models have been subjected to numerous reviews, and the various codes have been tested by the HEDR project staff independently of the developers to ensure correct implementation. The dose assessment has been found, on the whole, to be reasonably sound for the estimation of thyroid doses for representative, hypothetical individuals. However, errors have been found, and doubts have been raised about the validity of the results for particular environmental conditions and for the estimation of thyroid doses for specific individuals.
Summary of Responses to CDC's Questions 139 Although there are some fairly minor errors in the mode! calculations (use of sagebrush measurements to estimate pasture grass concentrations, use of wet rather than dry grass weights, and possible failure to consider wet deposition on pasture grass), the dose estimates for the period 1944-1947, the period of greatest exposure, appear reasonable. There might have been some underestimation of doses for later years, as pointed out by Hermann and Hermann (1996) and Hoffman and others (1999~. However, those possible errors would have relatively modest impact because the later doses were much smaller than those in 1944-1947 and the study children were older in the later period; other studies have shown that the thyroid is much more sensitive to cancer induction in early childhood than in later childhood and adolescence. There is reason to believe that the dose uncertainties that the HEDR project estimated and that the HTDS study used are underestimates of the total uncertainty. Some sources of uncertainty in the 13~] pathway to humans probably were not included, such as cow-feeding practices and commercial milk- distnbution patterns. But, owing to the scattered nature of the information on uncertainties, as opposed to its being summarized in one source, the subcommittee could not be sure about the uncertainties. In addition, uncertainties in the residential histories, lifestyle, and, especially, milk-drinking habits of the children were not accounted for. One major component of the determination of individual 13~{ exposures was the milk-drinking habits of the study subjects. Although an attempt was made to interview a parent or surrogate to obtain recollections of the milk consumption of each subject in childhood, it was possible to do so for only 62% of the subjects; in the other cases, default assumptions were used in calculating thyroid dose. The defaults that the CIDER mode} used proved to be considerable overestimates of the average doses derived from the reported milk consumption and source in the interviews. The potential impact of the discrepancy between
140 Review of the HTDS Draft Final Report reported and imputed milk consumption on the dose-response analyses needs to be evaluated further. The basic objective of the statistical analysis was to determine whether there was a cause-effect relationship between the occurrence of various thyroid diseases and the magnitude of exposure to 13~{ released from Hanford in the period in question. That was appropriately addressed by modeling the relationship between dose and the probability of occurrence of a thyroid disease. The HTDS also considered a reasonable set of potential confounding variables for thyroid disease. However, there was overreliance on the maximum- likelihood fitting of the linear dose-response model; for several of the important outcome variables (such as thyroid carcinoma), the mode] failed to converge. When the linear model as described failed to converge, an analysis could have been conducted with four or five dose groups and the average value of dose in each category could have been used as the predictor variable. That would probably have resulted in successful convergence of the mode! and retained reasonable power to detect an effect. Dose-response analyses with stratification on geostrata are needed because the HTDS tabulations showed that rates of several types of thyroid disease tended to be higher in areas with low fallout. That means that the geostratum differences would induce a negative association between I'll and thyroid-disease rates and might have masked a positive association between thyroid dose and disease. The HTDS investigators took care to examine the results for study participants who proved never to have been in the dosimetry area during the time of ]3~] exposure (the out-of-area participants). They performed sensitivity analyses to determine the impact of possible dose misestimation for those subjects and found it to be small. However, the HTDS investigators made no attempt to mode! the out-of-area doses for persons who were included in the main analyses. That is, if a person was in the dose-assessment area for only part of the time when there were IT releases, they
Summary of Responses to CDC's Questions 141 calculated his or her dose only for the time that he or she was in the dosimetry area. They implicitly assumed that the dose was zero for any time when the person did not reside in the area. No attempt was made to conduct a sensitivity analysis to evaluate how that assumption could have affected the results. Their approach to this issue could have led to attenuated or biased results in that it potentially estimated total fallout doses for some people but only partial doses for others. The HTDS investigators performed an adjusted comparison of the number of thyroid cancers found in the study versus the number that would be expected in the general US population (without radiation exposure) and found no difference. But there was a great deal of uncertainty in the comparison because the degree to which thyroid screening alters the number of thyroid cancers found is not well known. The subcommittee believes that the HTDS emphasis on analyses of subjects in the study rather than on comparisons with the general population is appropriate. There are sizable uncertainties in the doses reconstructed for individuals because of residential and, especially, dietary histories and because of the imprecision related to the source term, meteorologic conditions, pasture deposition, milk ]3~], source of milk, and iodine metabolism needs to be taken into account. it seems clear that analyses need to address the uncertainties explicitly, and the confidence intervals and the strength of the conclusions have to reflect them. ARE THE PRESENTATION AND THE DISCUSSION OF RESULTS COMPLETE? One of the difficulties that the subcommittee encountered while reviewing the work of the HTDS is that the method that was used to calculate the doses is not clearly descnbed in the documents that it was given. That information presumably is scattered in the large number of documents that were prepared by the HEDR project and the HTDS in the course of their work. it
142 Review of the HTDS Draft Final Report would be helpful to have access to a single document in which the HEDR work is summarized, the results provided to the HTDS are clearly described (with their strengths and weaknesses), and the ways in which the HTDS made use of the HEDR results to estimate individual thyroid doses are described in detail. It is unclear to the subcommittee how the uncertainties in the thyroid doses were estimated by the HTDS and the degree to which they might have been underestimated. In part, the lack of clarity reflects the lack of a single source that describes the dose modeling, including the coefficients and the uncertainty factors. There is a need to address explicitly the magnitude of the uncertainties associated with residential and dietary histories. There is a substantial description of how the dosimetry-related data were collected from people but relatively little information on how they were used. The input into the CIDER program is described as "scenarios", but these are not explicitly described, in particular how the scenarios were constructed from the data. There are a number of references to the use of default values in the CIDER program, but there is no discussion of which parameters used default values or of the degree to which default values changed as life circumstances changed for a given person (for example, if a person moved from a farm to a city). A tabular presentation of the pathways to diagnosis would help the reader to assess how the final diagnoses were assigned. Assignments were made in more than one way for each of the clinical outcomes. A table for each diagnosis with a list of the methods of diagnosis and the number of times each was used should be included in the full report, and these data should be looked at for indications of unsuspected ascertainment bias. Dosimetry-error issues apparently were not fully treated in the analysis of the study power. The same issue is raised by the results, in particular, ignoring dosimetry error could lead to unrealistically narrow estimates of the confidence limits that should be applied to the estimated parameter values. It is unlikely that the estimated dose-response relationships would change in an
Summary of Responses to CDC's Questions 143 important way, but confidence intervals that take dosimetry error into account would provide further information about the uncertainty of risk estimates. Some results were presented in an abstract, rather uninformative manner. For example, there was a scatterplot of individual thyroid doses on a logarithmic scale, but no table of the frequency distribution of doses, which would have meant more to readers. Similarly, one expects to see in radiation-epidemiology reports tables that show observed and expected numbers of disease outcomes in, say, four or five dose groups. Such key tables were notably absent from the HTDS Draft Final Report. A description of the estimated dose distribution (distribution of median doses for the individuals in the study) according to such important categories as geostratum, year of birth, and amount of milk consumption in childhood would be helpful, especially in interpreting the finding that the least exposed geostratum appeared to have the highest rates of many of the thyroid diseases or abnormalities. it was not very clear from the report how confounders of the dose-response relationship were treated, and results adjusted for possible confounders were rarely given. The HTDS investigators conducted analyses of the various thyroid-disease end points to evaluate a number of possible risk factors for confounding effects or effect modification, but they presented no tables to show a summary of the results of these analyses for any of the disease end points. Of particular value would be a presentation of results stratified by sex, age at initial exposure, magnitude of NTS and global fallout, and history of substantial medical radiation exposure. A tabulation of the number of study participants who were out of the dosimetry area for part of the exposure period is also needed. it has been suggested that the investigators should have given more attention to comparing the rates of thyroid disease that they found with the rates in other, unexposed populations. However, comparisons with an external, general population are
~· ~- 144 Review of the HTDS Draft Final Report fraught with problems in that people living in various geographic areas vary in their risk of thyroid diseases because of dietary and other factors and the rates of detected disease depend heavily on the frequency and sensitivity of thyroid examinations in the population. Those circumstances produce screening effects that cannot be readily disentangled so as to permit meaningful comparisons with rates from other geographic regions that did not have comparable screening. Therefore, the subcommittee believes that the HTDS investigators nghtly chose to emphasize the internal comparisons rather than general population comparisons. The discussion of the results was seriously incomplete in that it said little about whether the confidence intervals were wide enough to be compatible with those of 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. In the statistical-power analyses, it was assumed that the dose uncertainties were all of the "Berkson type"-a type of measurement error that does not affect statistical power. However, if a substantial, fraction of the variability of the HEDR individual dose estimates actually is due to non-Berkson error or to multiplicative errors, or if there is a substantial additional component of error due to uncertainties in milk consumption, lifestyle, and residential history, the power or tne study might nave been reduced below a point that would normally be considered acceptable. Furthermore, the apparent heterogeneity among geostrata might also have reduced statistical power. In contrast, if the doses were systematically underestimated, as has been alleged by others, that might tend to increase the statistical power of the study. ARE THE CONCLUSIONS REASONABLE? Our overall assessment is that the design and execution of the epidemiologic-clinical study was appropriate to the task. The sample was based on an almost complete census of eligible
Summary of Responses to CDC's Questions 145 subjects born in selected years in what are believed to be the high- dose regions and inclusion of subjects born in the same years and living in regions where the doses from Hanford releases were lower. Efforts to elicit study participation were thorough and appeared comparable for those with high and low doses. Care was taken to avoid methodologic study biases. The clinical examinations and laboratory studies used the best modern methods of detecting and defining thyroid disease. Overall, the authors deserve high marks for the carefully conducted and documented epidemiologic and clinical work they performed. To the degree that the dose is underestimated, the imputed risk estimates will be too large; but systematic, across-the- board dose underestimation does not alter the statistical significance of dose-response trends. Hence, in this study in which, as it turned out, the primary issue became whether there is any association between IT exposure and thyroid diseases-the impact of possible close underestimation might not change the study conclusions appreciably, except for two caveats. First, if there was an across-the-board underestimation of doses, the true statistical power of the study would have been greater than one would estimate it to be, given the reported dose distribution; the negative results would be more persuasive than they are. Second, if the doses were underestimated more for some study subjects than for others, this would, in effect, act as another source of measurement error that would tend to cancel out the gain in statistical power achieved by having generally higher doses. Therefore, a simple generalization about the effects of dose underestimation cannot be given. If the doses were systematically overestimated, the statistical power of the study would be less than claimed. in both versions of the analysis of mortality data whether data are arrayed by year of birth or year of death-SMRs for perinatal conditions and congenital anomalies in the period before the Hanford releases (that is, before 1945) are increased about as much as the SMRs in the period when exposure
146 Review of the HTDS Draft Final Report occurred. That constitutes evidence that exposure to ]3~{ iS probably not responsible for the increased mortality from thyroid disease in the Hanford study group, although some further analysis is recommended. The confidence intervals for the risk of thyroid cancer in this study and the Utah NTS study overlap to some degree. Moreover, on the basis of the considerations above, it is evident that the confidence intervals for the HTDS in fact depend on dosimetry-error assumptions; measurement errors in the dosimetry of the type that would attenuate the dose-response association, which we believe to be present, would cause the confidence intervals for the HTDS to be appreciably wicler than the ones based on the information in the Draft Final Report. Uncertainties in the magnitude of individual thyroid dose estimates and the relatively small sample in this study limit the generality of the conclusions that one can draw from the study regarding the magnitude of thyroid-cancer risk to other populations exposed to similar doses from }3~. Data on thyroid-cancer risk are awaited from the large number of thyroid cancers observed in children exposed to high doses from the Chernobyl accident. The statistical power of the HTDS to detect dose- response relationships might have been overestimated because effects of the types of dosimetry error that would attenuate the associations were ignored. The evidence of heterogeneity of many of the thyroid diseases or abnormalities among geoskata also suggests that the power of the study was weakened by geographic variations in unmeasured or unknown confounders that affected the outcomes. Such reductions of statistical power lower our appraisal of the study in relation to other studies and of the use of this study's results for making predictions about disease risk in other populations exposed to id at low doses and at low dose rates. The results of the HTDS are fundamentally important to the population living around Hanford. The study examined a substantial fraction of the highly exposed population and failed to find direct evidence of an effect of Hanford exposure on thyroid
Summary of Responses to CDC's Questions 147 disease risk, nor was there evidence of abnormally high rates of thyroid disease in the Hanford "downwinder" sample as a whole. It is very likely that no other reasonable study of this population could have found such effects-not only for technical reasons (such as difficulties in reconstructing doses), but because factors other than the Hanford releases appear to be more important in determining the amount of thyroid disease in the population. The fact that the two counties believed to have the least exposure tended to show higher rates of thyroid disease than the most exposed counties bears witness to that. Considerations of basic factors, such as the prevailing wind direction and distance from the Hanford site, indicate that those two counties should have had less~3~T exposure than the other counties in the study, but their disease rates were higher. That implies that there was not a strong association between ]3~} exposure and thyroid disease. Nevertheless, neither this evidence nor the negative dose-response results rule out the possibility of a weak association. We have already noted that the HTDS design and conduct were as good as they could have been, given the size and distribution of the population at risk and the long time between exposure and the study. The thorough examination of the highest- risk group-the youngest children living in the highest-dose regions failed to find an increased incidence of thyroid cancer, the hallmark effect of high thyroid doses from 13~. Even if some of the factors that affected the distribution of dose (for example, milk consumption) were poorly measured, they were as well measured in this study as is possible 40-50 years after the fact. In evaluating the HTDS, it is useful to distinguish what the subcommittee regards as two aims of the study. The first was the determination of whether patterns of thyroid morbidity among those in the study region during the fallout period correspond to likely patterns of exposure in the HTDS study sample irrespective of specific estimated doses, the likely patterns of exposure being based principally on location of residence during early childhood (distance down wind) and milk-drinking habits. This study, by
148 Review of the HTDS Draft Final Report virtue of its size and how well contact and screening were completed, appears to have had the ability to address that aim effectively, although the presentation of results could be improved. The second aim was to use assigned HEDR doses to estimate dose- response relationships. Many of the caveats discussed in this report about the effect of uncertainties in doses on the power of the study, confidence intervals for the estimated dose-response relationships, and the correspondence of the HTDS results with those of other studies are related to the second aim. The HTDS might have substantially overestimated its ability to assess a dose-response relationship, because of unallowed-for uncertainties, both systematic and random, in the HEDR doses. The absence of a thyroid-cancer risk above expectation, is based on a comparison of rates between high- and low-dose regions and a dose-response analysis that used dose-reconstruction methods and found no increased risk but did have wide uncertainties in the reconstructed-dose estimates and hence in the estimated risk coefficients. From those observations, one can only state that at face value the HTDS was negative. If an exposure of a population to 13~{ radiation of magnitude similar to that estimated in the HTDS were to occur elsewhere, one cannot predict with confidence whether an increase in risk would be seen. The small number of thyroid-cancer cases and the wide uncertainties in individual doses afford little confidence in the risk estimates derivable from this study. At the time of the initial release of the Draft Final Report, the HTDS investigators indicated that residents of downwind areas should fee} relief that their proximity to the Hanford nuclear site did not result in increased risk of any thyroid morbidity. Such statements are, according to the arguments above, reasonable in specific instances. For example, a healthy 55-year- old former resident of the area near the Hanford site (say, Benton County) who remembers drinking a large amount of milk as a child can take comfort in learning that there is no evidence that his or her risk of thyroid morbidity is higher than that of other subjects in
Summary of Responses to CDC's Questions 149 the general HTDS study area. The HTDS, in fact, appears to be definitive on that point for most of the outcomes considered. In contrast, for good reasons discussed above, the HTDS was not designed to compare thyroid morbidity rates among the entire population downwind from Hanford with those among populations living elsewhere. 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. They could be right, and their disease might have been the result of unusual fallout or ingestion patterns. However, thyroid disease tends to run in families, and the particular occurrences could be related to genetic factors in the families, chance occurrences, or even mistaken diagnoses. The small number of thyroid cancers detected in the HTDS was in line with the background rates initially estimated for the study, but the presence of screening effects makes it impossible to compare this intensively screened sample with any other similar population. Other thyroid conditions were much more commonly detected in the sample, but again no comparison group has undergone similar degrees of thyroid screening. The lack of evidence of a dose-response relationship for any of these conditions suggests-but does not prove that the overall risks were unaffected by Hanford releases. The evidence does not rule out (although it provides no support for) a weak association that could affect, for example, those already susceptible to thyroid disease because of predisposing genetic factors. Thyroid cancer is not a common disease, and it would be reasonable in future epidemiology surveys to identify, document, and investigate clusters by using molecular-biology probes to characterize genetic polymorphisms that could make people more sensitive to ionizing radiation or to look for oncogene prevalence in affected groups. These methods are developing rapidly, and it is likely that they will play a role in future environmental-epidemiology studies.
150 Review of the HTDS Draft Final Report WAS THE MATERIAL ACCURATE AND APPROPRIATE IN PROVIDING GUIDANCE TO THE PUBLIC IN UNDERSTANDING THE STUDY FINDINGS? For the most part, the written messages in the Draft Final Report, those about the report in public documents, and the messages given orally by the HTDS investigators at the media and public briefings were accurate, but they were sometimes inappropriate and misleading because they included statements that were too positive, definite, and strongly worded, given the uncertainties that applied to the study. In addition, the uncertainties related to the study were not discussed and should have been. Clearly, those problems contributed to the public upset that resulted from the release of the Draft Final Report. However, the problems related to the report's release should not overshadow the attempts made over the years to inform and involve members of the public about the study. They were valuable and generally done well. Such information channels as newsletters, background fact sheets, and a Web site were good ways of trying to reach members of the public directly with information about the study, and they augmented the information in the mass media. The subcommittee recommends that similar public communication efforts be continued regarding this report and others by CDC. Keeping the study process and activities as "transparent" as possible for the public is valuable and should not be abandoned because of the problems encountered with the release of the Draft Final Report of this study. IF THESE MESSAGES NEED TO BE AMENDED, HOW SHOULD THE REVISED MESSAGES BEST BE COMMUNICATED TO THE PUBLIC? Some of the overly positive messages given out at the January briefing have already been softened through CDC efforts during public meetings in Spokane and Seattle in May 1999. Handouts, a summary of the findings, and other written materials prepared by CDC also have conveyed similar information and made more of an effort to explain that no epidemiologic study can
Summary of Responses to CDC's Questions 151 determine "whether an individual person's thyroid disease is or is not caused by Hanford radiation exposure" (Cary, 1999~. Despite substantial attempts to publicize the May 1999 meetings, including paid advertising, they were not well attended by the public. So the "softer" CDC message was not delivered widely, although at least one newspaper did run a story about the new CDC approach. Several factors can account for the low attendance at the two meetings, including the period that had elapsed between the January release of the report and early May, the perception that no new information would be gained at these meetings, and the greater distance of the two meeting sites from Hanford, compared to Richiand, the site of the January briefing. Another interpretation might be that both the HTDS and CDC have lost some credibility on this issue and that people chose not to listen to what they had to say. All those factors and others might have been operating, but the last one should have serious consideration and study before release of the final report. If credibility has been lost, plans must be developed to try to restore it to some degree, particularly for the release of the final report. Given all the communication problems that resulted from the release of the Draft Final Report, the subcommittee recommends that another detailed communication plan be developed for release of the final report. To the greatest extent possible, those working on the plan should brainstorm about unexpected situations like those which affected the release of the draft report and devise plans to handle them more effectively. it is imperative that messages from the final report take into account the venous audiences being addressed and show concern and sensitivity for the thyroid-health issues that people perceive as affecting them. Any changes made to the Draft Final Report must be clearly outlined and explained, including why they were made, what group suggested that they be made, and what impact they have had on the results of the HTDS. Every reasonable effort must be made to present the filet picture of the study results, including all the uncertainties and other problems.
152 Review of the HTDS Draft Final Report CDC and the RIDS personnel should work together on the wording of the messages for the public about the final findings of study. Academic freedom is of great concern in letting researchers give their interpretations of the findings, but that does not mean that alternative interpretations of the data by other groups, such as CDC, should not be offered. It would be better to reach a consensus; if it cannot be done, showing differing interpretations is preferable to having one side dominate the other to the point that a government agency overrides the predominant point of view of the principal investigators, or vice versa. Such advice goes against some risk-communication dogma that states that the message should be kept simple and that all parties involved should show a united front. Sometimes that works, but often when members of the public are actively engaged in and educated about an issue, such a simple approach backfires, as it did with the release of the Draft Final Report. 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. A much simpler briefing structure should be devised to make the information available quickly to as many people as possible. In addition, an embargoed release of the report to journalists should be arranged so that they have a few days to read through it and develop informed questions before the briefing. Many government agencies and scientific organizations do that routinely and get much better, informed coverage because of it. As for details, the executive summary of the final report should be edited carefully to eliminate unnecessary technical jargon and complexity, which made it difficult for even educated readers to understand portions of the Draft Final Report. All written press releases, handouts, and other materials should be dated and numbered for easy reference; that was not always done with the material for the draft report. And all briefings should be videotaped and transcribed for future reference by CDC, the HTDS investigators, and others.
Summary of Responses to CDC's Questions 153 Little information has been available about the quality of communication to people who participated in the HTDS. A random sample of those people should be surveyed about their responses to the communication efforts carried out in connection with the draft report to seek ways to improve communication with them about the final report. The study participants seem to be an overlooked group; they need to be brought more into the communication loop. WITH REGARD TO RELEASE OF FUTURE STUDY REPORTS, HOW CAN CDC IMPROVE THE PUBLIC-COMMUNICATION PROCESS? Attempts to establish and maintain an information blackout before release of the Draft Final Report ran counter to the previous spirit of information-sharing with citizen groups in the region. The blackout, although somewhat understandable and probably policy-driven, led to ill feelings among citizens who had worked closely with the HTDS and CDC, and it led to a leaked report. Trying to maintain such a blackout dunng multiple briefings particularly in Washington, DC-is something that CDC should reconsider when a controversial report of great public interest is involved. If nothing else, the subcommittee recommends that the briefing structure be simplified and that citizens who have participated in the advisory process all along be given higher priority in the briefing structure. Multiple conference-call briefings, which are relatively inexpensive, appear to be ineffective. None of those who participated in them and provided information for the present report liked them. They have many disadvantages, and the subcommittee suggests that different ways be devised to brief groups at different locations. Briefings by satellite, although expensive, might pay off in the long run in more effective communication. Other possible uses of advanced communication technology, such as the World Wide Web, also should be considered. A lower-technology way to brief groups in different areas is to have simultaneous briefings led by different people .
154 Review of the HTDS Draft Final Report involved in the study. That method has its problems, but they seem less serious than those posed by conference-call briefings. Many of the suggestions made for handling the final HTDS report can also be applied to fixture CDC reports and need not be restated here. However, the subcommittee feels that it would be helpful for the CDC to ask a group of about a dozen people to come together during a 1- or 2-day workshop to discuss how the agency can improve its public-communication process and the release of future reports. Some of the people invited to such a workshop should be experts in risk communication, but there also should be a wider representation, including journalists, outside scientists who have worked on CDC studies, and members of citizen groups who have served on CDC advisory boards. Given careful planning, perhaps with a few case studies distributed to participants in advance, such a workshop might help CDC to avoid some of the pitfalls that occurred with the release of the HTDS Draft Final Report. One point for such a group to consider is the advisability of publicly releasing draft reports before external peer review and, if such a release is required by law or contract, how to do so effectively. Although circumstances in this particular situation might have forced the HTDS investigators and CDC to release a draft report to the public and the media, it is problematic to have preliminary information conveyed to the public if external peer review could suggest important changes in the final report. If substantial changes are made, what will the public think about the study and the investigators? That their initial results were not scientifically valid? That they raced to release information that was politically favorable to the government and not to citizens in the region? This is not a criticism of the actions that occurred, but a call to think about these issues more carefully, not only for this project, but also for others. Release of draft scientific studies, although it potentially serves the information needs of the public, also has the potential to cause confusion and to undermine the credibility of researchers and government agencies. Of course,
Summary of Responses to CDC's Questions 155 keeping information from the public can do the same thing, so this is not an easy issue to deal with. It should, however, be carefully considered not only by CDC, but also by other government agencies, Congress, the National Academy of Sciences and National Research Council, and the scientific community as a whole. Science journalists, too, might need to re-evaluate how to apply their own guidelines that advocate writing primarily from peer-reviewed studies and reports as they are faced with more and more cases in which government agencies and others announce the results of unreviewed draft reports at news conferences.
