Page 16 - 70_03

LENNART K. PAALZOW AN. R. ACAD. NAC. FARM.

However, as shown in Fig. 11, a Bayesian estimates using only two
or three samples in the beginning could predict the dosing for the
following three months without blood sampling. All the subsequent
blood samples could have been avoid if Bayesian technique had been
utilized and by that saved hospital money and distress for the patient.
It has to be inferred, however, that it may also be other factors than
drug blood concentration values that have to be taken into considera-
tion at the same time when performing blood sampling.

FIGURE 11. The patient in Fig. 10 with individualized drug dose predictions based
only on a Bayesian estimate using the first three blood concentration measure-
ments.

Neuroleptic drugs used to treat psychosis are used during long
time periods. It is an interesting class of drugs since there is a dis-
tinct concentration range within which you have a therapeutic effect.
The concentration effect-relationship is however complex for several
of these drugs (8). With increasing concentration the effect against
psychosis increases reaching a maximum and than the effect start to
decline with further increase in concentration, a type of curve that
we can call Bell-shaped. In order to get maximum therapeutic effects
you have to keep the concentration within a certain limited range
and therapeutic drug monitoring using Bayesian dose adaptation
can facilitate this.

An example of these drugs, perphenazin is administered as an
intramuscular depot injection with the aim of delivering the drug for
a whole month. The problem with these depot preparations is that

676