ML20135B243
| ML20135B243 | |
| Person / Time | |
|---|---|
| Issue date: | 02/26/1991 |
| From: | Bernero R NRC OFFICE OF NUCLEAR MATERIAL SAFETY & SAFEGUARDS (NMSS) |
| To: | Thompson H NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO) |
| Shared Package | |
| ML20135B134 | List:
|
| References | |
| NUDOCS 9702280141 | |
| Download: ML20135B243 (6) | |
Text
s, -
'o UNITED STATES g
[4y , .t. NUCLEAR REGULATORY COMMISSION 7
% .** ( '** jg f E W ASHINGTON, D. C. 20555 o *
\s a,#ff
, Febru uy 26, 1991 MEMORANDUM FOR: Hugh L. Therapson, Jr.
Deputy Executive Director for Nuclear Materials Safety, Safeguards and Support FROM: Robert M. Bernero, Director Office of Nuclear Material Safety and Safeguards
SUBJECT:
POTASSIUM 10DIDE DP0 This memorandum is in response to your request that I address the residual part of the Potassium locide (KI) Differing Professional Opinion (DPO) filed by Peter Crane on June 16, 1989. As you know, in 1983 I was the Director, Division of Risk Analysis in the Office of Nuclear Regulatory Research, responsible for SECY-83-362, and was a participant in the Commission briefing ,
on that paper which took pl6ce on November 22, 1983. That paper and the briefing are the basis of Crane's DPO.
I have reviewed the following reference documents:
- 6. SECY-83-362
- b. Transcript of the November 22, 1983 Commission Briefing
- c. Memo from Peter Crane to H. L. Thompson, Jr. , dated June 16, 1989
- d. Memo from Peter Crane to H. L. Thompson, Jr., dated January 4,1990
- e. Memo from H. L. Thompson, Jr. to Eric S. Beckjord, dated March 7,1990 j
- f. Memo from Eric S. Beckjord to H. L. Thompson, Jr. , datec March 15, 1990 i
- g. Memo from James M. Taylor to Eric S. Beckjord, dated April 16, 1990 I
- h. Memo from James M. Taylor to Commissioners, dated April 16, 1990
- 1. Memo from Peter Crane to H. L. Thompson, Jr., dated September 24, 1990
- j. Note from Peter Crane to H. L. Thompson, Jr., dated Hovember 9,1990 The last document noted poses the question that appears to be the residual issue of the DP0: "What did the Agency do to investigate the charge, made in j the June 1989 DPO, that the NRC staff misled the Commission and the public on l potassium iodide in 1985?" Crane's memo of September 24, 1990 (reference 1),
gives a more specific description of concerns which I interpret as his basis for the charge that the staff misled the Commission, namely, "that the problen, j with the NUREG was not just one of inaccurate numbers, but also, more fundamentally, of an erroneous gp' roach to applying cost-benefit techniques to an issue of public health policy. I will try to aodress these issues in the I reverse order, first the approach and then the numbers. "
On pages 2 and 3 of Enclosure A to SECY-83-362, the staff listed six questions i relevant to the issue and said: "The decision for requiring the administering of KI to the general public must weigh all of the answers to these questions.
The methods for pulling together all of the answers into 0 readily m _
g 9702280141 970226 PDR ORO NERCTN PDR
l Hugh L. Thompson, Jr. February 26, 1991 {
I understandable form is [ sic] cost / benefit and uncertainty andlysis j
' techniques." The paper then went on to present the analyses. Simply put, the j paper asserted this to be 4 valid approach, without presenting arguments about j why it is valid. Later in the briefing (Tr. 54-56) Chairnwn Palladino and i Commissioner Asselstina both expressed misgivings about the method of cost-benefit analysis. In retrospect, I wish we had taken the time in the peper and the briefing to establish the validity of the approach first, and then proceeded with the applicdtion.
The use of a thyroid blocking agent such as K1 in the event of a nuclear reactor accident is a consequence mitigation measure which will realize its benefits only if 1) a serious accident occurs, and 2) someone in e position to benefit has and takes the egent in time. The costs are essentially fixed and continuous, whether for stockpiling or predistribution; the benefits cannot be treated in the same way since they are not fixed and continuous, but ere rare events. The benefits can only be belanced to costs by taking into account their probability of realization. Consider a parachute as a lifesaving device. The cost of a parachute is indeed small compared to a human life, but it is not a worthwhile medsure in a commerci41 airliner. The likelihood of an dCcident occurring with circumstancqs where the benefits of the parachute can be realized is su small that the cost-benefit ratio is clearly not favorable.
I doubt that anyone has ever done a strictly quantitative cost-benefit
-analysis for the parachute in this' case but the qualitative cost-benefit j dnalysis is evident dnd sufficient. And to equate the reality of the cost of thf: parachute and its lifesaving benefit is to raise a wholly different question, if the parachute has such evident value why should one. fly in such a dangerous airplane. In e similar vein Roger Blond said in the Commission l briefing, ". . . assuming the accident occurs once every year . . . the question !
should never come down to potassium iodide, but allowing those reactors to operate." = (Tr. 66-67)
The cost-benefit method is well established in the NRC's regulatory process, e.g., 10 CFR Part 50.109. The requirements for adequate safety protection (high quality systems, conteinment, remote siting, etc.) must be met, leaving room for worthwhile safety enhancements if their cost is justified by the risk ,
averted. Peter Crane has suggested that there is some difficulty in that cost-benefit analysis is offered for predistribution of KI but not for other emergency protection measures. I have no doubts that one can identify many details of required emergency planning features which, taken elone, would not pass a cost-benefit analysis test. For example, the analysts who prepared ,
NUREG-0396 might have recommended 410-kilometer (61 mile) radius plume ;
exposure pathway Emergency Planning Zone (EPZ) instead of a 10-mile EPZ. I believe that if the analysis results in NUREG-0396 had been plotted on a metric distance scale, those andlysts might have chosen that round number of 10 km as a reasonable planning range. Later, after that 10-km EPZ was set as a requirement, someone might argue that a 10-mile EPZ (en additional 6 km) would be a more prudent choice for the EPZ. In those circumstances there could be a legitimate cost-benefit appraisal of the added 6 km radius.to the EPZ. Would the fixed and continuous costs of the additional preplanning and 4
. .. . .. . . . _ - - - . - . . .. -~ .. .- .
._.-l
)
Hugh L. Thompson, Jr. -3 February 26, 1991 l exercising be outweighed by the improvements in scfety or averted risk. As I i recall, there was at least one-case, I believe Shearon Harris, where l' intervenors argued the need for a 20-mile EPZ. It might be worth reviewir.g ,
l the litigation of that case to see whether it established a precedent. l What the regulations did do in 10 CFR Part 50.47 and Appendix E to Part 50 was
! to require an EPZ of 10 miles radius. In the supporting document NUREG-0654
! (FEMA-REP-1) there is provided ". . . a colmion reference and guidance source for: . . radiological emergency response plans and preparedness . . ." (p 1).
Later in that document (pp 61-64) the evaluation criteria for protective response include:
"10. The organization's plans to implement protective measures for the plume exposure pathway shall include: ... ;
- a. ...
- e. Provisions for the use of radioprotective drugs, particularly for emergency workers and institutionalized persons within the plume exposure EPZ whose immediate evacuation may be infeasible or very difficult, including quantities, storage, and means of distribution.
- f. State and local organizations' plans should include the niethod by which decisions by the State Health Departnent for administering radioprotective drugs to the general population .
are made during an emergkncy and the predetermined conditions '
under which such drugs may be used by offsite emergency workers;y 1/ See DHEW (new DHHS) Federal Register notice of December 15, 1978 (43 FR 58798) entitled " Potassium Iodide as a-Thyroid- ,
Blocking Agent in a Radiation Emergency." Other guidance '
concerning the storage, stockpiling, and conditions for use of this drug by the general public, is now under development by the Bureau of Drugs, DHHS."
Let me turn now to the " inaccurate numbers" and how they may have " misled the Commission." Reviewing the an61yses now, almost eight years later, I do recognize four areas where 1 believe substantial and significant inaccuracies warrant reconsideration, they are:
- 1. Identification of mortality risk but failure to treat it in the parametric analysis;
- 2. Inadequate quantification of morbidity costs; l
I l
Hugh L. 'i wpson, Jr. 4 February 26, 1991
! 3. 'Inadequatt. quantification of KI distribution costs;
- 4. Inadequate quantification of the availability / effectiveness of the predistributed Kl.
In addition, although I think the conclusions would not change, I believe a
~
- n. ore system 4 tic analysis of various stockpiling options would have been useful, especially in light of Consnissioner Bernthal's concerns. Let me address the four areas above in turn.
In risk analysis of carcinogenic pollutants it is typical to speak of sefety goals, limits, and cost-benefit analyses in terms of cancer mortolity risk, given that for many years the majority of cancer victims do not survive it.
Thyroid cancer is unique in that the survival rete is very high. Therefore, the cost-benefit andlysis for KI dealt with morbidity costs as the dominant concern; it should have dealt with both. On page 4 of Enclosure A to SECY-83-362, the Commission was given the WASH-1400 analysis cf a 10 percent murtality rate for malignant thyroid modules but on page 15 the costs are only for medical treatment associated with thyroid removal. In an analysis tracing back to NUREG/CR-1433, there was derived a benefit value of $20,000 per ,
thyroid module averted, the key figures being 40 percent of all thyroid I n.odules turn out to be malignant and a 10 percent mortality rate for victims
-of. malignancies. As later dialogue illustrated (Reference c, pp 9-10) this
$20,000 cost was for the survivors, and did not take into account the 4 percent ~(10 percent of 40 percent) of victims who died. Proper treatment of the health benefits would have combined three interdependent benefits:
Averted Mortality = '
(Number of Nodules) x (% Malignent) x (% Mortality) x (Life Value)
Averted Malignent (Number Morbidity of Nodules) x (% Malignant) x (% Survival) x (Malignant Morbid'-
Averted Non-Malignant Morbidity =
(Number of Nodules) x (% Non-Malignant) x (1) x (Non-Malignant Morbidity Cost) j This distinction of three separate categories would recognize that for the largest fraction of thyroid nodules suffered, the non-malignent, the thyroid gland would be expected to function after surgical removal of the nodules but ,
there would be medical costs and some impact on the patient's life with lingering uncertainty regarding possible malignancy. For the second largest i group, those who suffer malignant nodules and survive, the medical costs and impact on the patient's life are much greater. Typically, the thyroid gland would be removed, a synthetic hormone treatment established, and the regime of whole body scans for recurrence of malignancy would be followed. The smallest i group, in the case of thyroid nodules, would be those who die of the malignancy. Thus it happens that, in assessing the averted costs of . thyroid nodules, the size of the group affected drops as the severity of effect :
increases; and the dominant element of the averted costs is not evident. I a !
l i i
Hugh L. Thompson, Jr. February 26, 1991 ha Going back then to the analysis of the three interdependent benefits, I would expect that central estimate values then should be*:
l
% Malignant 40 Percent
% Murtality 6 Percent (a more realistic value, then and now)
Life Value $1H ,
Malignant Morbidity Cost $100,000 (see discussion which follows) i Non-Malignant Morbidity Cost $10,000 With these central estimate values there should have been an attendant uncertainty or sensitivity band--I wculd suggest the following tentative ranges: 1
% Malignant 30 - 60 Percent
% Mortality 10 - 4 Percent Life.Value $1M - $3M .)
Malignant Morbidity Cost $20,000 - $500,000 Non-Malignant Morbidity Cost $2,000 - $50,000 I questioned the adequacy of the quantification of morbidity costs. I believe Peter Crane has a telling point, that the simple medical costs are a serious understatement. As my personal reestimates above indicate. I think costs an 3 order of magnitude higher would be more appropriate, better reflecting the costs of pain, loss of productivity, reduced quality of life, etc.
I also questioned the costs of KI distribution. In retrospect, I believe it l was a substantial error to use costs of only a few cents per person per year, j which were based on the bulk purchase costs alone. Distribution and
- maintenance costs would dominate the costs. Simply mailing the tablets to households would easily double the costs used. A more occurate, but difficult way to assess the costs would be for staff to analyze a substantial continuous program of distribution, education and maintenance for KI in various size areas around nuclear plants. Account must be taken of placing doses not only in homes but in places of work, schools, etc. The rate at which we change residence or place of work is just as significant as the shelf life of the KI suggesting a maintenance and replacement cycle of perhaps every two years.
l
'The central estimate values and sensitivity range values presented here reflect only my own current enderstanding. They should be reviewed with competent authorities.
l 'a B
l Hugh L. Thompson, Jr. February 26, 1991 l l
This is a very difficult cost estimate to make; we knew that then and avoided
! it. I don't know what the costs would be but I expect at least a one or two order of magnitude increase.
_ Last, I questioned the availability or effectiveness of predistributed KI.
After eight years I can't remember when I was told of the Tennessee follow-up experience with predistribution of KI around the Sequoyah plant. I recall )
being told that when they went house to ho m , to check and replace after just ;
two or three years, they found a large fraction of the original residents were '
no longer there and, further, that a large fraction of the original recipients )
who were still there could not recall or locate the KI. I believe a i reevaluation of the KI issue should consider this effect as well. 1 1
I Wdh an active pdrticipont in this matter eight years ago and it is difficult for me tu be objective in this analysis. When I consider the charge that the
" staff misled the Commission and the public" it is not clear whether the charge meahs "to lead or guide wrongly" or "to deceive." As I have said in the earlier parts of this memorandum, I recognize some substantial inaccuracies in the 1983 analysis but the method of cost-benefit analysis with i uncertainty / sensitivity analysis was robust because it displayed in a I structured way both the central estimates of staff and the sensithity of the i overall judgnient to uncertainties or alternate views. It was not a close call.
I expect that a thorough reevaluation now will reach the same conclusion "that potassium iodice offers extremely small benefit in relation to its costs and is not cost effective as a preplanned emergency protective measure for the general
'p'uB11c" (SECY-83-362). Therefore, I believe that staff did not " lead or guide wrongly" the Comission or the public.
Nor do I believe that the staff decieved the Commission or the public. In retrospect one must admit that the total package suffers a large measure of i inscrutability. The presentation to the Comission depended heavily on an understanding of SECY-83-362. And SECV-83-362 depended heavily on an understanding of NUREG/CR-1433. Nevertheless, available information was used, analyzed with a defensible method and presented to the Commission with candor.
57 S$
Robert M. Bernero, Director Office of Nuclear Material Safety and Safeguards cc: P. Crane, 0GC E. Beckjord, RES l