Subject: Re Brain tumor studies (Kundi)(Anon).
Date: Sun, 24 Dec 2000 065650 -0600
From: Roy Beavers
To: guru
--------------------------------------------------
.......Response from EMF-L......
One of our expert readers reinforces Professor Kundi's comment....guru....
Roy,
Thanks for asking for the clarification. In my earlier reply I should
have been more explicit in identifying that I was, and am, referring to
the recently published Muscat research.
Expanding on the original message, Muscat's data has no meaning without
considering at least a two-dimensional or, more appropriately, a
three-dimensional probability basis. I'm very surprised that the
reviewers did not notice and point out the built-in confounders that
were not considered in the report. ANONYMOUS
Roy Beavers wrote:
>
> Which study are you addressing below? It looks like you are refuting Kundi???
> I assume that is not what you intended. He is making some of the same points
> you are........guru......
>
> ANON wrote:
>
> > There is a fatal flaw (at least one) in the statistical methods employed
> > for this study. The study should have employed a joint probability in
> > determining association. By restricting the analysis to only those
> > persons who presented with symptoms during the period, the investigators
> > have placed their work on a questionable, perhaps fundamentally invalid,
> > foundation. The joint probability refers to those persons who should
> > have been included as representative sample candidates but were not by
> > reason of
> > DEATH and those persons who have NOT YET PRESENTED with symptoms It is
> > not mathematically appropriate to restrict the analysis in the manner
> > that has been described. Not good epidemiology.
> >
> > Roy Beavers wrote:
> > >
> > > ........From EMF-L.......
> > >
> > > This response certainly deserves some serious consideration......guru...
> > >
> > > -------- Original Message --------
> > > Subject: Brain tumor studies
> > > Date: Fri, 22 Dec 2000 12:18:39 +0100
> > > From: Michael Kundi
> > > To: Roy Beavers
> > >
> > > Dear Roy,
> > > In your comment on the recently published brain tumor studies you stated
> > > correctly that there are a number of hypotheses which should be tested
> > > in epidemiological studies and that the Muscat (and Inskip) study tested
> > > only the hypothesis that 'cellphone use does not normally influence
> > > brain cancer development'. It might be no surprise to you that in these
> > > studies not even that hypothesis could be tested. It is by far too
> > > general. First of all, there is no such thing as 'brain cancer'. What is
> > > called brain cancer is a multitude of completely different entities
> > > (that have their origin in different tissue types, are slowly or fast
> > > growing, and have different etiology and predisposing factors). If
> > > cellphones influence brain tumor development (without affecting
> > > incidence) there are at least two possibilities: There could be an
> > > influence on speed of development and there could be a reduction of
> > > latency from malign transformation to the onset of clinical symptoms.
> > > For the first possibility there are again at least two hypotheses: The
> > > dissipation of heat in superficial regions of the brain (especially at
> > > the temporal lobe) could reduce the speed of development because
> > > hyperthermia is a known tumor growth inhibiting factor, and the exposure
> > > can lead to a faster development due to its possibly promoting
> > > potential. Again these mechanisms will have an impact which depends in
> > > the type and localization of the tumor. Both studies did not have enough
> > > power to detect any of these effects. Not even an attempt was made in
> > > these studies to test the assumption that exposure could increase speed
> > > of tumor development. This would have involved inspection of the
> > > patients records for early clinical signs and to determine the duration
> > > until tumor diagnosis. A further problem of these studies is that they
> > > didn't account for latency. If cellphones influence tumor development at
> > > an early stage only exposures which have occured maybe many years ago
> > > (depending on type of tumor) should be included. To cumulate exposures
> > > up to hospital admission is nonsense given the hypotheses that were
> > > tested (it makes, however, sense if the tumor growth rate would have
> > > been analysed).
> > > Best regards and Merry Chrismas,
> > > Michael Kundi
> > > --
> > > Department for Occupational and Social Hygiene
> > > Head:Univ.Prof.Dr.Michael Kundi
> > > Institute of Environmental Health
> > > University of Vienna, Austria
> > > Kinderspitalgasse 15
> > > A-1095 Wien
> > > Tel: +43-(0)1-40490 64726
> > > Fax: +43-(0)1-4277-9647
>
> --
>
> Roy Beavers (EMFguru)
> roy@emfguru.com
>
> It is better to light a single candle
> than to curse the darkness.....
>
> WEBSITE: http://emfguru.com
>
> People are more important than profit$$
Archive provided courtesy of WaveGuide, http://www.wave-guide.org
Reprinted with permission of Roy Beavers, http://www.emfguru.com