Re: MDL and Dose Effects

[rkathren@tricity.wsu.edu (Ron L. Kathren)]

Kim --

Interesting rediscovery of the wheel!!!  Brings back memories of my days as
a graduate student at Pitt under the tutelage of one of the unsung heroes of
our profession, the great Allen Brodsky.  Perhaps the definitive work on the
statistics of multiple film and single quarterly film was done by Allen
Brodsky and published in Health Physics 9:463 (1963).  As I recall, Brodsky
called for reporting of any dose, even below the so-called lower level of
detection, noting that the errors would tend to cancel out because of the
Law of Large Numbers.  There was a companion article in the same issue that
I coauthored with Allen that considered the effects of environmental
conditions.  This same problem has been considered in many epidemiologic
studies (eg such as those by Ethel Gilbert), and the National Research
Council report on film badge dosimetry in atmospheric testing of a few years
back even devised a way of getting around the problem of a lower dose
cutoff.  But Brodsky's prescient and three+ decade old statistical
evaluation (which can apply to other dosimeters as well) reigns supreme.
Tankersley, who I note is not a member of HPS and hence likely to be
unfamiliar with the literature, might profit by consulting the older
literature, or some of the comptetent HP talent at Oak Ridge.

Ron Kathren

 >This article appeared in today's Knoxville News-Sentinel. It is available on
>the www at:
>
>        http://www.knoxnews.com/features/munlab/fm110496.html
>
>______________________________________________________________________________
>
>Research finds unrecorded radiation data was significant
>
>By Frank Munger News-Sentinel staff writer
>
>
>A researcher at the Oak Ridge Institute for Science and Education has thrown
>another kink into the controversy about low-dose radiation. 
>
>Bill Tankersley's study shows that some nuclear workers at Oak Ridge -- and
>elsewhere -- received radiation doses significantly higher than they thought
>because records did not reflect doses at low levels. 
>
>This stems from the historic practice of recording a zero if the worker's
>dosimeter measured radiation at or below the ''minimum detectable level.'' 
>
>The MDL is the measured level at which authorities believe it is difficult
>to ascertain whether an exposure took place. It also represents a belief
>that even if a worker did receive an exposure at or below the minimum
>detectable level, it was probably insignificant from a health perspective. 
>
>Tankersley's research, however, found that the magnitude of unrecorded doses
>was significant in the 1940s and '50s.
>
>During that period, the Atomic Energy Commission used an MDL of 30 to 50
>millirems. At the same time, film badges used to measure doses were changed
>weekly -- instead of quarterly or even annually as some radiation readings
>are handled today. 
>
>Therefore, an employee whose job regularly exposed him to low-level
>radiation in the range of 30 millirems could have received an unrecorded
>dose of several rems or more over his career. 
>
>This information might not have any direct relevance in determining whether
>an individual worker suffered health effects as a result of radiation doses,
>but Tankersley said the data could provide more insights for epidemiologists
>studying cause-and-effect issues in a population of nuclear workers. 
>
>''Today . . . even very low exposures are suspected of contributing to
>excess disease,'' Tankersley wrote in a paper published earlier this year in
>the Journal of Applied Occupational and Environmental Hygiene. ''Therefore,
>there is increased interest in dose previously considered unimportant.'' 
>
>While reviewing records at Oak Ridge National Laboratory, Tankersley and
>colleagues found a ''substantial and otherwise inexplicable'' increase in
>radiation dose levels among employees after 1956. 
>
>The increase coincided with ORNL's changeover from weekly to quarterly
>monitoring of employee doses. 
>
>''Although the same practice of censoring readings below (30 millirems) was
>continued, the magnitude of the possibly missed or unrecorded dose during
>the period when monitoring was done weekly was many times greater than that
>likely missed for the same reason during quarterly monitoring,'' Tankersley
>wrote. 
>
>He and other researchers also looked at sets of raw dosimetry data from
>workers at the Oak Ridge facilities to evaluate the reliability of low-dose
>results, which traditionally were reported as zero in employee records
>instead of the actual measurement. 
>
>They found that low doses -- ''even those near the zero level'' -- can be
>''rather accurately measured with a relatively large set of data.'' 
>
>In an interview, Tankersley said there is a level of impreciseness with any
>monitoring system, but he said it was particularly important to researchers
>to have as much actual data on radiation exposures as possible when studying
>nuclear workers. 
>
>The low-dose information may be helpful in explaining some reported health
>effects among workers that previously were not understood, he said. 
>
>Donna Cragle, who heads ORISE's Center for Epidemiologic Research, said
>epidemiologists are now trying to incorporate these findings into current
>studies and, in some instances, retrofitting old studies with new
>information if available. 
>
>In the study's conclusion, Tankersley wrote: 
>
>''It is not possible to determine accurately the fraction of total dose
>unrecorded due to procedures where all readings less than the applicable MDL
>are set to zero. However, our investigations indicate that, in some cases,
>the amount is considerable.'' 
>
>The administrative practice of recording zeros for radiation doses at or
>below the MDL continues today at the government's Oak Ridge plants and other
>nuclear facilities around the country. However, the minimum detectable level
>for dosimeters is much lower than it used to be, and monitoring period is
>longer, too, which reduces the chance that a significant dose of radiation
>will be missed. 
>
>Nonetheless, Tankersley urged authorities to change the policy and report
>all radiation doses as measured. 
>
>
>Kim McMahan, CHP
>Office of Radiation Protection				    Sola fide ...
>Oak Ridge National Laboratory				  Sola gratia ...
>P.O. Box 2008   Oak Ridge, TN  37831-6290	       Sola scriptura ...
>Ph:  (423) 576-1566				      Soli Deo gloria .
>e-mail: mcmahankl@ornl.gov
>
>
>
>