Comparison of Self-Ratings of Psychiatric Patients with Ratings made by a Psychiatrist

¹ Reader in Psychology, Institute of Psychiatry, de Crespigny Park, London, SE5 8AF
² Physician, The Bethlem Royal and Maudsley Hospital, London, SE5 8AZ

Four experiments were carried out upon a psychiatrist's ability to predict the degree of symptom intensity reported by patients. The psychiatrist's predictions were based on interviews with the patient, each of which was conducted either immediately before or immediately after the patient had reported the intensity of his own symptoms in a reproducible and quantifiable manner by means of a personal questionnaire. The psychiatrist gave expression to his predictions by responding to a duplicate questionnaire, the results of which could be compared exactly with the patient's own responses. Each of the four experiments was carried out on a different psychiatric patient, each of whom suffered from a long-term non-psychotic affective disorder. Each experiment involved between 6 and 17 separate occasions of testing, and each patient was rated on between 20 and 29 symptoms on each occasion. The psychiatrist's predictions were exactly correct about 60 per cent of the time, and were out by one scale point at least another 25 per cent of the time. Discrepancies between psychiatrist and patient did not appear to be related to content but to level of severity, accuracy being greater at the extremes of intensity than in the middle. The results are seen to have implications for research into the nature of psychological disorder and into patient-doctor communication.