8127/frame/fm

 
1.
Levels of Validation
Validation can occur at several levels. The top level is that of validating a new
technique when it is first introduced. This may involve efforts of multiple
laboratories and extremely broad and comprehensive testing. One of the
advantages of interlaboratory validation studies is the opportunity for sample
exchange. This helps establish the limits not only for intralaboratory varia-
tion, but interlaboratory variation, using the same samples. These mammoth
studies are usually only necessary during the initial development or technol-
ogy transfer of a substantively new method. The rest of the field can then
rely, in large part, on the initial body of data that will ideally be peer reviewed
and published. Of course, when substantive changes are introduced into the
original protocol, it must be revalidated in a more-limited fashion to under-
stand the full impact of those changes.
The next level of validation is sometimes called in-house or internal
validation. Although reliability for the general system has been established, it
is still incumbent on those in each individual laboratory to document that
the test works as expected in their hands, and to establish laboratory-specific
limits imposed by the overall system (e.g., protocols, instruments, reagents).
These studies are usually much more limited, but also serve to familiarize
laboratory personnel with the test and what compromised samples might look
like. Regardless of how long a method has been in use and generally accepted,
internal validation must always be performed when setting up a new program.
What one might call “personal” validation translates as training and
proficiency. Even though the field and the laboratory may have great confi-
dence and experience in a particular method, each individual must also build
8127/frame/ch09 Page 224 Friday, July 21, 2000 11:44 AM

---

Good Laboratory Practice — Establishing Validity and Reliability
225
a foundation of personal experience and familiarity. This is accomplished by
a directed program in which the analyst first tests or compares pristine
samples, then practices with various compromised samples. New examiners
might then practice on some adjudicated case samples before taking a formal
proficiency test to demonstrate their competency to work cases.