Self- and observer assessment in anxiolytic drug trials: a comparison of their validity

Self-rating scales are considered to be less useful for comparing different treatments in anxiety patients than observer-rating scales. However, the empirical evidence for this assumption is not adequate. A self-rating inventory of 35 items related to anxiety was perfectly parallel with an observer-rating inventory. Both instruments were used in the Cross National Collaborative Panic Study to compare the efficacy of imipramine, alprazolam and placebo in an 8-week drug trial in a sample of 1168 outpatients. The variance of the self-rating assessments was about two times higher. Both scales were equally sensitive to change; however, the measurement of change by means of the self-rating scale was slightly less consistent. The discriminative power of the observer-rating scale between placebo and active treatment was two to three times higher than that of the self-rating scale; consequently the observer-rating procedure provides a more valid instrument when the efficacies of different anxiolytic treatments are compared between different groups of patients.