Some thoughts on epidemiology

[RuthWeiner@AOL.COM]

Apologies in advance for a long post, but my thoughts may be of interest to RADSAFERs.


I am not an epidemiologist, but I am an observer.  Thus, I have been thinking about what troubles me about the supposed magnetic field (Fritz is right -- we were all sloppy) impacts on leukemia, miscarriages, asthma, Type-II diabetes, etc.


I had my teeth cleaned with an ultrasound probe today.  I don't get exposed to that magnetic field more than a half hour every three months, but the hygeinist is exposed every day for about 6 hours.  Moreover, nowadays ultrasound pictures of fetuses are quite regularly taken.  If a group were studying the putative impact of magnetic fields on miscarriage rates, dental hygeinists and pregnant women getting ultrasounded would seem to be logical study subjects.  Why weren't they selected?


The whole approach in the magnetic field studies contains an insidious bias.  The Field, et al, radon studies may contain a similar bias.  A really simplistic summary of the way such studies are done is:


Step 1: an environmental phenomenon exists (magnetic fields from AC transmission lines, radon in homes, second hand smoke) that some people think might have some effect on health.

Step 2: the epidemiological investigator tries to determine what the health effect might be, and selects effects that are suspected to be correlated with some environmental insult, and/or that have been correlated with sizable exposure to some environmental insult.  Cancer is everyone's favorite because some cancers meet one or the other criterion.  Allergies and chronic disease meet the first criterion.  Since the Alsea, OR case was refuted, the only reason I can see for looking at spontaneous miscarriage is that it is lurid.

Step 3: The investigator designs a case-control study in which his or her cases are the people who have suffered the insult (cancer, miscarriage).

Step 4: The exposure to the environmental insult is measured.

Final step: The investigator "bins" the exposure data, claims to have accounted for all confounders,  shows some kind of association between the exposure bins and the cases, and claims correlation.


The radon and magnetic field studies stop there (yes, there are some "meta-analyses but they follow basically the same pattern).  Note, however, that second-hand smoke studies  include additional considerations.  Second hand smoke has the same constituents as first-hand smoke, so that its epidemiology can be assumed to be closely related.  The epidemiology of first-hand smoke began with the observation that most lung cancers, that are not clearly related to occupational exposure, occur in smokers and conversely, non-smokers ("never smokers"), who are not occupationally exposed, rarely have lung cancer.  Much the same can be said for the incidence of COPD (emphysema): most non-occupational sufferers are smokers, and non-smokers not occupationally exposed rarely develop it.   In other words, the epidemiology began with the observation of  health effects.


Smokers admit to certain adverse health effects directly related to smoking like persistent productive cough and shortness of breath.


Laboratory experiments (yes, with animals) showed that the constituents of cigarette smoke produce malignant tumors. 


A positive correlation between smoking frequency and years of smoking, and both lung cancer and COPD, has been observed repeatedly.


Plausible mechanisms for both carcinogenesis and COPD initation have been proposed.


The epidemiology of second-hand smoke has considerably firmer bases than the epidemiology of either magnetic fields  or radon in the home.  The latter two have only shown weak correlations with health effects.  Radon, at least, produces a known carcinogen (ionizing radiation), so the question of correlation is more about an exposure threshold  and the type of dose-response relationship than about the carcinogenic nature of ionizing radiation or the existence of some  dose-response relationship.

In the case of magnetic fields, however, the investigators can't even figure out what health effect they are looking for (leukemia? other cancers? spontaneous miscarriage? asthma? Type II diabetes?). 


The study cited in the recent post by Bjorn Cedervall ought to put to rest the notion that magnetic field exposure might cause leukemia, or is even correlated with it.  Actually, similar studies, with similar results, were published more than ten years ago (unfortunately I don't have citations handy).


Let's, for a moment, hypothesize what the public response would have been if tests with laboratory animals had shown that tars from cigarettes did NOT produce tumors.  Smokers would have rejoiced, proclaimed that smoking has nothing to do with lung cancer, and the surgeon general's warning might never have got on cigarette packages.   


Overhead transmission lines are unattractive, to say the least, but that doesn't mean that they can be correlated with an adverse health effect.  Residential and commercial development that requires new transmission lines may be unwelcome, and certainly causes irreversible changes to the landscape, the view, the natural environment, etc.  However, the claim that magnetic fields from the associated transmission lines "cause," or are even correlated with, some adverse health effect, on the basis only of the epidemiological studies cited on the powerline webpage, is unsubstantiated.  Just because you don't like something, or it messes up the natural environment, doesn't mean that it can be correlated with health damage, or even that there is enough justification to go looking for health damage.


 

Ruth Weiner, Ph. D.

ruthweiner@aol.com