Re: Plutonium detection levels in Urinalysis

[ASOLUTION8@aol.com]

The incoming letter brought up the concept of the LLD, which is a concept
that is not well understood and often misunderstood.  I will try to clear
away some of the confusion, but if I am in error, I hope that some radsafer
will straighten me out.

LLD (lower limit of detection) is often used as meaning the same as MDA
(minimum detectable activity) or MDC (minimum detectable count).  Let's use
the MDA to represent the concept.

The MDA has a equation something like this:  MDA = (3 + 4.65 sigma)/k, where
sigma is the standard deviation of the measurement and k is a constant to
convert counts to units of activity.  95% confidence is assumed.

But the MDA is meaningless unless it is used in conjunction with Lc, the
critical level, where Lc = (1.65 sigma)/k.

So if, for example, MDA = 50 microcuries, Lc might equal about 17
microcuries.

By definition, the MDA is an activity that can be detected 95% of the time if
ALL measurements exceeding Lc are considered positive readings.  The Lc is
also set so that if there is no activity present whatsoever (zero activity),
the false positive rate will not exceed 5%.

Therefore, if the MDA concept is used, the correct interpretation of ANY
value above Lc is, "An intake exceeding the MDA may have occurred."  That is
the correct use of the concept.  The MDA is an alarm level.  If you want to
know that any intake exceeding a certain value will be detected, you use the
MDA concept.

Conversely, for any value less than Lc, the corrrect interpretation is, "An
intake exceeding MDA is not likely to have occurred.

Thus, for the example given above, any measurement greater than 17
microcuries would have the interpretation, "An intake exceeding 50
microcuries may have occurred."  For any measurement less than 17 microcuries
the correct interpretation is, "An intake exceeding 50 microcuries is not
likely to have occurred."  Reporting measurements of less that 50 microcuries
as "less than MDA" is a boo-boo.  Doing that has destroyed much of the
usefulness of the data. 

But the MDA concept has little use for recording routine repeated bioassay
results.  People who report readings as "below MDA" are misusing the concept
and should not do so.  If they wish to misuse the concept, but misuse
somewhat less, they should enter numerical values for readings above Lc and
"below Lc" for values below Lc.

The correct way to record the results of routine repeated bioassay
measurements is to record the measured value and specify a range, generally +
or - 2 sigma, representing the 95% confidence level.  The correct
interpretation of that result is that the true intake is likely to be within
the specified range.

This approach has the advantage of correctly specifying the range of possible
intakes, allows an unbiased sum of a worker's annual intake, and allows
propagation of errors to correctly specify the confidence range for the
worker's annual intake.  

Note that even if the uncertainty range for a single bioassay measurement is
quite large in relation to the size of the measurement, when the year's total
intake is calculated and the uncertainty range specified, that range may be
quite small relative to the annual intake.

I hope I have this right!

Michael S. McGuire
Simplified Solutions Software, Inc.