Measurement of antipsychotic-induced side effects: support for the validity of a self-report (LUNSERS) versus structured interview (UKU) approach to measurement.