Test-retest reliability and influence of practice effects on performance in a multi-user computerized psychometric test system for use in clinical pharmacological studies.

Fifty naive drug-free healthy young male or female volunteers performed psychometric tests on 6 sessions with 3- to 4-days intervals, using a new multi-user computerized test system for use in clinical pharmacology. Tests of simple reaction time, complex reaction time, concentration, motor coordination, and short-term memory (word pairs or figures) were performed. Clear practice effects were shown for almost every psychometric variable recorded. The magnitude differed considerably between tests. The magnitude of practice effects was most evident (46.5-55.0%) for the concentration test, the coordination test, and the Vienna reaction test. Intermediate practice effects (20.8-31.0%) were observed with the complex reaction test (percent correct reactions) and both short-term memory tests (test duration). Only small practice effects (5.1-14.3%) were observed with the reaction times of the simple and the complex reaction test, and the percent correct responses in the short-term memory tests. After 3 test sessions, significant further improvements could not be shown for most tests, but for the reaction times in the simple and the complex reaction test this was true from the first or the second test session, respectively. For the concentration test and the coordination test, significant practice effects could be shown even after 5 training sessions. It is recommended to perform at least one training session before the start of clinical pharmacological studies with psychometric testing. Test-retest-reliability, as determined from session 5 to session 6 by Spearman's rank correlation coefficient (R8), was very good (> or = 0.95) for the concentration test (percent correct responses) and the coordination test (mean steering time). Most other variables showed intermediate (0.44-0.68) reliability (reaction times in the simple and complex reaction test, percent correct reactions in the complex reaction test, percent errors in the concentration test, test duration of the short-term memory tests). The percent correct answers, which is the primary variable in both short-term memory tests, had a relatively poor reliability (0.14-0.18).

Effect of educational level, gender and age on the performance in a multi-user computerized psychometric test system for use in clinical pharmacological studies.

197 native, drug-free healthy young volunteers performed psychometric tests on a single occasion, using a new multi-user computerized test system which consists of tests of simple reaction time, complex reaction time, vigilance, concentration, motor coordination, short-term memory (word pairs or figures) and abstract language-free reasoning (2 versions). Normality of distribution of all psychometric variables was checked. For the reasoning tests and the memory tests, internal consistency and parallel test reliability were determined. Cross-correlations between the variables and factor analysis were done to evaluate whether different tests measure different brain functions. Multivariate variance analysis was carried out to test the effect of the independent factors school education, gender and age on the performance in the psychometric tests. Subjects with a lower school education level performed worse in the reasoning tests, the concentration test and the memory tests. Females were slower in the coordination test and made fewer correct solutions in the concentration test. Older subjects performed worse in the reasoning tests and had a longer working time in the memory tests than younger ones. The results show the necessity of psychometric screening of volunteers before recruitment for clinical pharmacological psychometric studies in order to reduce-individual variability.