Drug monitoring studies as a method of analyzing response criteria

Two multicenter drug monitoring studies are presented. Some methodological problems are dealt with and the validity of such studies is discussed in terms of differential indication. In a first study (Lehmann et al., 1993) the results of a 12-week xanthinol niacinate treatment (500 to 3000 mg daily) in a cohort of 10,134 outpatients suffering from cerebrovascular insufficiencies were recorded systematically by nonreactive evaluation methods. The therapy was found to be most successful in patients with the target symptoms vertigo, tiredness, lack of concentration, affective disorder, and disturbances of vigilance and vitality. The most frequent side-effects were flush or heat sensations in 9.1% of the patients and gastrointestinal complaints in 3.3%. In a second study (Klieser et al., 1994) we systematically collected data from 219 patients with Major Depressive Disorder during five weeks of treatment with fluoxetine (20 mg daily). The results showed that depressively inhibited, anxious patients with a depression of minor severity, who showed a relatively marked improvement within the first week of treatment, profited the most from this therapy. The first study was designed to use nonreactive evaluation methods. Correlation analyses helped to identify the types of patient with a good response to treatment. The second study was organized on the model of conventional controlled pharmacological studies with the application of commonly used scales. The differential indication was to be inferred from the uni- and multivariate comparison of responders and nonresponders. In the light of these two studies, the problems of target definition, sample design, target variables, practicability, statistical analysis, and validity are discussed.