Re: Ecologic versus Case-Control Studies

[Muckerheide <muckerheide@mediaone.net>]

Bill Field wrote:

> Jim Muckerheide made the statement that the "power of the statistics
> of even a small ecological study is much more powerful than any human
> case control study"

No I didn't. A good case-control study, with small numbers, can have better
statistics than a "small" ecological study (in my ref to Cohen, his "small"
subset studies may be a million people of the 200 million total population, 20
million in a group in the whole study?). But that is totally dependent on the
premise of a good case-control study being able to precisely know the actual
dose, ie, to esssentially eliminate that variable. That works fine when a dose
is an injection or ingestion controlled fixed quantity.

But, radon case-control studies can't know the dose. Radon studies measure the
house. Two "equivalent" people (women) in the same house will _likely_ have a
significant actual difference in lung dose; 2 people in different houses with
the same measurement are _likely_ to have an even larger variation in actual
lung dose. 

Therefore, the statistical basis for having confidence in performing a
case-control study with small numbers is gone. (As you note below, this
doesn't relate to the power of the study. This is just a question of the
meaning of the basic statistics that go into the study; to the validity of the study.)

In these circumstances, the variations in actual doses vs the measured dose
depend on the statistics of the variations around the measurements, which now
are seen to vary rather substantially. (Again, as you note below, this is just
in the statistics - Statistics 101 if you will, which is distinct from the
matter of the "power of the study" itself.) As you think about it, the "dose"
in the case-control study is now not a known quantity, but is a statistical
mean or average, just as with the ecological study. But with such small
numbers, usually not more than a few thousand, the variation is very
substantial vs the now comparable issue of representing a large population by
the measured values as averages, in a large ecological study. In that case,
the 'power' of the statistics alone narrow the error bands (as Bernie's study
shows, and as Norm Frigerio's study of cancer rates showed using 18 years of
data with more than a million deaths in each data point, which itself was a
representative subset of decades of cancer death statistics.)

Obviously, independent of other factors, just from the lack of controlled dose
data, any case-control study will produce a result that can vary around
central values. In Bernie's many independent studies of various, independent
subsets, because the dose and cancer data statistics are so much more
powerful, the variations in results are small. Not only that, the value of any
confounding factor that could substantially effect the basic results become
very large. 

After considering many confounding factors, Bernie is quite justified in
looking at the other side of the issue and concluding, by the direct
arithmetic of Statistics 101, that a confounding factor that could change the
slope would have to be larger than the effect of smoking (and not seen in any
study to date), but even more significantly would also have to be in a direct
negative correlation with radon concentration levels (and then wouldn't reach
the BEIR slope unless it was MUCH larger than smoking :-)

(At the same time, there is biological evidence that the association is being
confirmed by: animal biology data, and cellular and molecular biology, that
confirms the "stimulating" effect of moderate levels of alpha exposure to the
lung, (and the whole organism, with early studies in mice and rats that
uranium dusts and other alpha-contaminants resulted in long lives and improved
physiological factors, see eg, Henry, JAMA, 1961 review article and summary of
an Oak Ridge report.)

> Jim,
> 
> 1. What is your definition of power of a study?
> 
> 2. In your opinion do you see any differences between the "power of
> the statistics" and the "power of the study".
> 
> 2.  Are you really saying ecologic studies have more power to detect
> an association (or lack of one) than case-control studies?
> 
> 3.  You asked, What is Dr. Cohen's data suffering from?  Could his
> results be suffering from cross level bias within his ecologic
> stratifications?  

It doesn't seem likely. What would be the source?

Is this at all possible in your opinion?  If no, why not?

Because there seems to be no source, certainly not at a level to affect the
basic correlations. Can you propose one?
 
> Bill Field
> bill-field@uiowa.edu


Thanks, Bill.
 
Regards, Jim Muckerheide
muckerheide@mediaone.net
Radiation, Science, and Health
==============================