MSK Report

Page 1

Mamser, Eunan Eealth Studies Program, , Division o f Biomedical and De pst sent of Energy, Herewith the re-riew you invited. You gave no ir-dica%ion of the detail you wollld like, so I have adopted a general. approach adh pt the leri&h dwm. The alternative,to have worked thzoq ? each table in tur"",, would hve nade very turgid reading but noreover would hz.;e dist rz c ted at tent i - from whatI consider t o be the major inadequacies of the report. The widespread and rapid pd b li city of the report, even before pb li cation, has already lead toits fb-dings being eq loi ted in s eve zd contexts. involved in evidence f o r two high court hearings in which %he PEd report has been cited to est m te balance of probabilities between sqontaneous and radiationmyelom death was due to ler els of radiation within occqa - > d ly per n it ted In quir j together with m y references t o it by objectcrs %athe proposed I n one, the > E X figures lead toa 2%; probability thata The report also receive i a day of direct public23 at ,the Yindscale I would advise that this report fus very little s-tific merit and that Ad- a little more than i s therefore the findings have very little validity. deducible from the report (ie. a n inpression formed by three authors), I f e e l that l*lancuso left all the technic aspects of the ulalysis t o Stewart and that in t-am s.he left all the details to E&z.d7.le. fie ale is a ~ca,? features in the raw data and does not see contradictions *en the ori gird data and the resslllts of a nethod in * d ? i ~ c hhe -has ime skc 5zplicit faith. with most peosie with whom I bave discussed this report, '2 :find the style of some later the. Such ? ~ 1i nter pet ation is not clear fAmt h e repon, a d hence the table lends its e if to erroneous s33lication. t o each of the

Page 2

and fron the m dys is 02 - , that the only statistic Ll7 sigxi5icaS finding which surri-,-es the v a r i o ~ ~corrections is the Myelom deaths of the three men with doses o f 34 zzds ( 2 3 7 l ) , 29 rads (age 58) and 20 rads (age 71). Even here caution i s necessary since the finding is based on only 3 deaths, two of these at a n age when nyelona incidence i s becoming hi@. drawn attention to these myeiomas as the main finding and concluded a n urgent need to look a to t ker groups for excess myelomas, they would have inad? a better contribution to rau tion epidezio1o.g. One geneml coment about the dzta files used h %e sC2d.y. me that one of the cii2f icd ties in. c5ecking azzy analysis is that.s-i-wequent ugdzting o f the ff led es it h2ossi ble to test exactly t% sace ez lier. We have a similzz data file but 211 transactions with Qe file are recorded on another file so that the file can be regenerated t o the statu3 it had on any particular b te. s o that thy can avoid the problems that Ethel Gilbert has eq erie n ced whea she tried to check the 6ata used by YS. I hope my comegts a-e o f help, I azl sorry tkat t3ey r e ~ dso cri%ic zl ly. analysis of the Ihn2ord ikta. whether the reviex is pu ely for pur own Use or i f it will be published o r =de Ix ou ld be grateful if you would let Ee laow

Page 3

with age iiidicated. for bane m~rowt o 12.2 rads for all cancers are ve qe uck szlallez tf.a any vazues coq at i ble vith othar studies. t o have given verj ctrefirl consideration t o these resydts, There is na evidence that they h v e doce this. o r unccrte.int,g of those nxizers. forth a study, that th ounce r tz k ties s rev eq laz ga. Eeczuse of this one vo..dd -xve eq ec ted the authors The very lest necesszly is toss tk zte the accuraqscientific value without such an assessnect. In s-aixxi~jboth the text ad tke tables are o f such & s czi ty that ury as sock tion of the f i ? di.navi th rz di atl on his been burie", rather tl-iari extracted. .The cl ahs of the authors are of such conseqcence th t the2 - Ether anzlysis referred t o in tk sir disc x sic z sko-2.d hzve 30en carzied oucI before pX i cz5ioz of these results which are in such disagreeztent srith allpe - ri3us -ark appear t o be =ore rel at ad to the statistical methods eq hye d th to iS cx zat iQ n con take d in the data. The sTmey does no Pore tha sazgcs% 312';i tis necessary a greater ra diz tion az soci ation than was hitherto sup?oseZ (based on three deaths The rajcr fixdings