The mosaic approach is intended to be applied just after preclinical data before or in parallel with phase I studies. It is not a screening method for thousands of drugs. In its present state it consists of a complicated procedure and is rather work-intensive for each single drug. It is intended to supplement the risk assessment resulting from preclinical studies. It should provide some help for risk assessment with data obtained from humans. If e.g. three compounds are selected for further development, the mosaic approach may provide helpful data for preferring one over the other.