Re: False Positives in Personnel Dosimetry?

["David Whitfill" <DWhitfil@kdhe.state.ks.us>]

A more practical problem is that the personnel exposures reported are
relative to the control dosimeter exposure. If different workgroups are
located in different buildings or areas and only one control is used, small
exposures may be reported just due to natural background radiation
variations. I was monitoring two groups of people in different buildings. I
kept the control in Building A and periodically would get small exposures
reported from people in Building B, with one researcher getting about twice
that of the others (exposures ranged from approximately 20 to 50 mR per
quarter-very small for occupationally exposed personnel). A simple
environmental survey revealed:

a) the background in Building B was twice that of A, and
b) a researcher was periodically hanging his labcoat (TLD attached to the
pocket) adjacent to a source-storage locker.

The moral is that if you monitor personnel in different buildings or
geographical areas, you may need additional control dosimeters at those
locations if natural background radiation levels differ and it is
bothersome to see small exposures routinely reported.

DJWhitfill

Opinions expressed are mine and do not reflect official policies or
positions of the Kansas Department of Health and Environment.



                                                                                                      
                    Chris Alston                                                                      
                    <alstonc@odrge.odr.georg        To:     Multiple recipients of list               
                    etown.edu>                      <radsafe@romulus.ehs.uiuc.edu>                    
                    Sent by:                        cc:                                               
                    radsafe@romulus.ehs.uiuc        Subject:     False Positives in Personnel         
                    .edu                            Dosimetry?                                        
                                                                                                      
                                                                                                      
                    07/27/00 06:04 PM                                                                 
                    Please respond to                                                                 
                    radsafe                                                                           
                                                                                                      
                                                                                                      




Folks

One of the leading vendors has introduced a dosimeter for which the default
minimum detectable reporting level is one (1) mrem. Since many (most?) of
us monitor fairly large groups of people, who are not exposed to radiation
in amounts that are important (with respect to the need for protection) in
the occupational arena, would the use of these badges, at the default
sensitivity, result in an increase in false-positives? I'm wondering if
it's analogous to other statistical tests, e.g., in clinical pathology,
where it's my understanding that, if one uses a probe of very high
sensitivity, say, a blood test for a disease or drug, in a large population
in which the incidence of the "trait" is very low, one accepts the
likelihood of more f-p's, relative to the number that one would expect in a
group of the same size, that was known to have a high incidence of the
trait. Or do I simply misunderstand something fundamental about statistics,
and does the purported phenomenon exists only between my ears?

Hoping not to be eggfaced, I thank you in advance
cja
alstoncj@georgetown.edu

P.S. If knowledgeable persons at the vendor in question, who are (I know)
forbidden to post to the list, wish to email me directly, I promise to
offer the response to the list as if it were from Deep Throat, i.e., with
anonymity guaranteed to the responder.
************************************************************************
The RADSAFE Frequently Asked Questions list, archives and subscription
information can be accessed at http://www.ehs.uiuc.edu/~rad/radsafe.html



************************************************************************
The RADSAFE Frequently Asked Questions list, archives and subscription
information can be accessed at http://www.ehs.uiuc.edu/~rad/radsafe.html