The utility of computer versus clinician-authored assessments in aiding the prediction of patient symptomatology

摘要

The present study assessed the differences between clinician-authored and computer-generated personality assessments. The participants, 155 psychologists in independent practice, received one of four packets of client information which contained: a) Minnesota multiphasic personality inventory (MMPI)–2 profile; b) MMPI–2 profile and demographic information; c) MMPI–2 profile, demographics, and a computer based test interpretation report; or d) MMPI–2 profile, demographics, and a clinician-authored report. Participants were asked to assess a patient based on the information they received. These judgments were compared to symptom ratings made by staff where the patient was hospitalized. Results demonstrated that participants who received clinician-authored reports were most able to match hospital staff's ratings of patient symptomatology. Those who received computer-based test interpretation (CBTI) reports, however, were typically not able to make accurate assessments of the patients. In fact, participants who received a CBTI report of an ambiguous MMPI–2 profile (one which did not appear to match the patient's diagnosis) were actually less accurate at assessing symptomatology than were participants who received no report.