Re: A respectable end to Cohen's LNTT radon debate has arrived!

[Bernard Cohen <blc+@PITT.EDU>]

Steve Miller wrote:



>The recent article in the Health Physics Journal puts to rest for me the validity of Dr. Cohen's assertions that he has shown the LNTT is invalid.   I think a respectable end has arrived!  Even a physicist on the committee agrees the inverse finding is smoking related. Dr. Cohen's rebuttal is not convincing. 

>

          ---With regard to the physicist (he had never been involved in 

Health Physics), he is the Committee member involved in the following 

material which was edited out of my response:



> "After the HBHM paper was completed and submitted for publication, 

> NCRP allowed me to hold direct discussions with one of the Committee 

> members. He was quite gracious about it, and we exchanged more than 25 

> e-mail messages in each direction. I will enumerate some of the 

> misunderstandings of that Committee member not as a reflection on him, 

> but rather to illustrate the highly questionable procedures imposed on 

> Committee members by NCRP. All of the misunderstandings that surfaced 

> in these discussions could have been resolved by back-and-forth 

> communication with me as part of the Committee deliberations rather 

> than after the Report was finalized.

>

> My e-mail exchanges with the designated Committee member revealed that 

> he thought my theory treatment uses an absolute risk model with no age 

> dependence, Eq.(2-9) of BEIR-IV; actually C-95 (my 1995 paper) 

> specifies that I use the BEIR-IV age dependent relative risk model, 

> the only formula given in the BEIR-IV Overview, as Eq. (1.1). Appendix 

> A of C-95 goes through this age dependent calculation in detail. He 

> did not understand that I use the theory to calculate the mortality 

> rates for each age, and apply them to a standard age distribution 

> (applicable to all counties) to obtain an age-adjusted mortality rate. 

> This means that differences in the age distributions for individual 

> counties have no effect. He did not understand that the observed data 

> I used were for age-adjusted mortality rates although this is clearly 

> stated in C-95, so the theory and the observed data were for the same 

> thing - age-adjusted mortality rates. He did not seem to know the 

> meaning or purpose of age-adjusted mortality rates, and he certainly 

> did not understand that using them eliminates differences of age 

> distributions in different counties as a complicating factor.

>

>             He thought I had not given uncertainty estimates for radon 

> levels, although  this uncertainty was given as 17% (standard error) 

> and explained in Sec. D of C-95. He objected to my not stating 

> quantitatively the effects of deleting the retirement states FL, CA, 

> and AZ from my data base, although this was clearly stated 

> quantitatively in Sec. C of C-95. He did not understand that I was not 

> directly testing Eq. (1.1) of the BEIR-IV Overview, but was using that 

> equation to calculate the BEIR-IV age-adjusted mortality rates (as is 

> done in BEIR-IV Table 2-4)  which I then compare with observed 

> age-adjusted mortality rates. He thought that my observed data were 

> for deaths caused by radon, rather than for total deaths; he could not 

> understand how I could obtain the observed deaths caused by radon, 

> which, of course, I could not. He had great difficulty in accepting 

> the fact that I gave no estimates of uncertainties in the age-adjusted 

> lung cancer rates for each county obtained from National Center for 

> Health Statistics (NCHS) because NCHS does not provide that 

> information; I offered to calculate effects of any uncertainties he 

> would propose as plausible, but he made no such proposal. He did not 

> seem to accept my use of several independent sources to obtain 

> separate sets of data (on radon levels and on smoking prevalences) as 

> a method for investigating effects of systematic errors. I point out 

> these problems not as a criticism of the Committee member with whom I 

> exchanged messages, but rather as a strong criticism of the rules 

> under which the NCRP Committee operated. They make me wonder what 

> misunderstandings were involved in the thinking of the other Committee 

> members."

>

          ---The problem with the Committee procedures is explained in 

another portion of my response that was edited out:; it follows:



.   "I assumed that a scientific evaluation would include the members of 

the Committee developing a rather complete understanding of my work, and 

that this would require back-and-forth communication between them and 

myself to clarify confusions, to resolve misunderstandings, and to have 

me do further analyses that they might request. But except for two 

requests for some of my data sources, there was no such communication. 

My only direct contact with the Committee was at its first organizing 

meeting where I was asked to present my work. For that purpose, I 

prepared an elaborate review of my published work with extensions to 

include several important new analyses, all organized into a tightly 

reasoned presentation (Cohen 2005 - hereafter C-05). In HBHM, there is 

no reference to this tightly reasoned treatment, and no evidence that it 

was considered"



            ----If you would like to discuss scientific issues, I would 

be happy to respond..