A prospective study of 167 consecutive patients admitted to a general hospital following an episode of deliberate self-harm was carried out; 102 patients were interviewed by computer and then by a psychiatrist who was blind to the results of the computer interview. The computer interview consisted of a self-rating modification of the Hamilton Rating Scale for Depression and a novel questionnaire developed to assess suicidal ideation. This article explores the preliminary findings in these patients and suggests that not only is the computer interview acceptable to the majority of patients but the data suggest, in line with previous studies, that the patients are prepared to confide information to the computer that they may be unwilling to tell the clinician. Further the data also suggest a significant pathoplastic effect of the personality of the patient on the perception of the psychopathology by the clinician. The computer appeared to be a better predictor of suicidality than the interview by the clinician.