[Previous Months][Date Index][Thread Index][Join - Register][Login]
  [Message Prev][Message Next][Thread Prev][Thread Next]

[IP] Probability lesson

Time, I think, for a little clarification of what is happening when you
compare meters with lab tests.  As I said earlier, there are multiple
approximations going on.  When there are associated events, the
probabilities are multiplied to obtain the correct probabilities.  With the
meter approximated as +/- 20%, we can look at a result as being 80% likely
to be correct.  If we also assume a variability of 10% based on the
blood/serum differences, and also a small (say 5%) variability due to venous
versus capillary blood, we now have precisions of 80%, 90%, and 95%.  

	.80 X .90 X .95 = .68 or +/- 32%

Or 68% accuracy.  How many of us want to rely on a comparison that is likely
to be incorrect almost one third of the time?

Note that the two variabilities for blood/serum differences and for
capillary/venous blood are number I pulled out of the air.  I don't know
what the real variability is.

Jim Handsfield
email @ redacted <mailto:email @ redacted> 
The opinions expressed are mine alone and do not necessarily represent those
of my wife who runs our house and makes more important decisions than I do.
for HELP or to subscribe/unsubscribe, contact: