Confusion on data in the NSWS final report.

Dear Colleagues, I have not been following the discussion on the data from the nuclear shipyard worker study (NSWS).

However, since my name keeps coming up I thought I would take a few minutes to list a few URLs and offer to send you copies of the article by Ruth Sponsler and me on the NSWS. The PI, Prof. Genevieve Matanoski has not published anything on it except the abstract in Radiation Research in 1993. I have not read it. She has not answered any of my e-mail messages for several years. She refused to look over our draft article on the NSWS for errors.

My most recent e-mail to her was sent a month or so ago at the request of the Editor of HPS News. I used the final report as a reference in two different letters to the HPS News. The editor had heard indirectly that Prof. Matanoski had told someone that the NSWS final report shouldn't be referenced. I agreed to e-mail the PI about it. If I receive no reply by 1/15/2003, the editor will contact her.

The NSWS research was well done. If it had found an increase in cancer or a decrease in longevity of the nuclear workers, it would now be widely quoted in NCRP reports. Since the results were the opposite, it has been ignored as much as possible. If you do a search on Google or similar search engine for nuclear shipyard workers, I doubt if you will locate it on the web. If you go to the DOE home page and do a search on the title of the final report you have a good chance of locating it. The URL for it is given below.

Last week I accidentally found the DOE news release in Sept. 1991 at http://dewey.tis.eh.doe.gov/health/epi/docs/hb91-3.pdf I found the news release misleading in that it did not give a comparison of the cohort to the controls. It only gave the comparisons to the U.S. population (white male) and explained the difference as due to "healthy worker" effect.

Here is one paragraph: " The results of this study indicate that the risk of death from all causes for radiation exposed workers was much lower than that for U.S. males. These results are consistent with other studies showing that worker populations tend to have lower mortality rates than the general population because workers must be healthy to be hired and must remain healthy to continue their employment."

There is no mention of the fact that the death rate from all causes of the 28,000 cohort was 24% lower than that of 32,500 age-matched and job-matched controls. That is 16 standard deviations. If you convert 16 std. dev. into P values, I have it from Prof. Don Herbert that P<10^-16. for 16 std. dev. I can see no way that "selection bias" (as claimed on p.. 196 of NCRP Report No. 136) can explain that large a difference.

If you go to the URL given at the end of this reference you will find the final report. Look at the tables 3.6 (a, b, c, d) you will see much of the data you are looking for.

Matanoski, G. 1991. Health effects of low-level radiation in shipyard workers. Final report. 471 pages Baltimore, MD, DOE DE-AC02-79 EV10095. National Technical Information Service, Springfield, Virginia. Available on line at http://www.osti.gov/bridge/product.biblio.jsp?osti_id=10103020

In particular John Jacobus should look at Table 3.6 d and explain what he meant by his last paragraph: "It is interesting that all groups who worked in the shipyard, exposed and non-exposed to radiation, had lower risks than the general population. Maybe that should tell you something about the subject population also." Table 3.6d for the NNW or controls will show that they had a cancer mortality 12% higher than the U.S. public (white males). The NNW death from all causes was 2% higher than for the public. (Maybe the inverse of the healthy worker effect?)

I usually don't read messages that come from John Jacobus unless they are addressed only to me instead of everybody. This message is going to everybody to help correct misunderstandings. The NSWS is a sad science story. I don't know who is to blame for the situation. I suspect a number of people must be involved.

I plan to post our NSWS article by Ruth Sponsler and me on my web page at : http://www.medphysics.wisc.edu/~jrc/ It won't be there for a few days. After you "enter" look under the "ARTICLES" button. The page is under construction but slowly, as I am in Florida and the part time student making the changes is in Madison.

If you wish to communicate with me, send me an e-mail but do not cc: radsafe if you expect an answer. I normally don't reply to messages to radsafe, even if my name is listed in the cc: area.

In a few days I plan to put my new "letter to the editor" to Radiology on my web page. It is titled: "Longevity is the most appropriate measure of health effects of radiation." There is no proof for my hypothesis but the present data from NSWS and the 100 years of British radiologists are convincing to me. We will need a double blind study to determine if my hypothesis is correct.

--
John R. Cameron (jrcamero@wisc.edu)
2678 SW 14th Dr. Gainesville, FL 32608
(352) 371-9865 Fax (352) 371-9866
(winters until about May 10)

PO Box 405, Lone Rock,WI 53556
(for UPS, etc. insert: E2571 Porter Rd.)
(608) 583-2160; Fax (608) 583-2269
(summer: until about Oct. 15)