RE: Averill's Editorial

[gj_vargo@ccmail.pnl.gov]

          I'm saddened and embarassed to see arguments such as the one 
          appearing in Averill's editorial embraced by anyone associated 
          with the radiological protection community.  Once again, we see a 
          failure to understand the difference between deterministic and 
          stochastic effects.  
          
          The production of thermal burns is essentially a threshold 
          process - injury occurs when the organ's rate of repair or 
          defense (in this case, the ability to reject or transfer heat) is 
          exceeded.  It is a threshold process in which the appearance of 
          acute injury can be fairly well established (look at the warning 
          labels on many new showers).  Above the threshold, the severity 
          of the injury increases with increasing temperature up to a point 
          where complete destruction of the organ (or organism) occurs.  
          Using an example such as this to attack the linear-no threshold 
          theory is irresponsible and, in my mind, could represent a form 
          of malpractice.
          
          The reference to the Finish radon study is also unfortunate.  The 
          study lacks sufficient statistical power to either accept or 
          reject the null hypothesis of no effect.  For a review of this 
          paper, refer to the Continuing Education Lecture given by Dan 
          Strom at the Health Physics Society Meeting last week in Seattle 
          (dj_strom@pnl.gov).  
          
          I also think its equally irresponsible to apply LNT concepts to 
          miniscule environmental doses and calculate horrendous collective 
          doses that have little, if any bearing on reality.  The LNT 
          hypothesis is a tool, and, like any tool can be misused with 
          unfortunate results.  The concept of collective dose - an idea 
          which is predicated on the use of a linear model for health 
          effects - is useful and appropriate in many situations, 
          particularly when dealing with occupationally exposed 
          individuals.  See NCRP Report No. 121 (1995) for a thorough 
          discussion on the uses and misuses of collective dose.
          
          Certainly one would never equate the effect of one person-sievert 
          to one or two individuals to the same dose delivered to one or 
          two hundred.  On the other hand, we frequently consider reducing 
          the number of workers engaged in an activity, even when 
          individuals doses might be somewhat higher in order to reduce 
          collective dose.  
          
          In the IRPA IX Sievert Lecture  (reprinted in Health Physics 
          71(2):122-125), Beninson states "The linear non-threshold 
          relationship is at present the best tool to predict risk 
          probability at low doses.  It fulfills all the requirements to be 
          considered "realistically representative" using modeling 
          terminology.  Practical decisions can be made using this 
          relationship, and the radiation protection system recommended by 
          the ICRP provides a method for such decisions."
          
          While I believe LNT to be a good tool for prospective purposes, 
          it is inappropaiate to use LNT for retrospective risk 
          calculations - especially individual risk calculations - because 
          of the dependence on sex, age at exposure, age at onset and other 
          individual risk factors (e.g., smoking, etc.)
          
          Instead of attacking the LNT hypothesis as a model for 
          radiological protection, we should seek out and attack its 
          inappropriate application.  
          
          
          George J. Vargo, Ph.D., CHP
          Pacific Northwest National Laboratory