Psychiatrists and a computer as interrogators of patients with alcohol-related illnesses: a comparison

Abstract

A computer-administered 'interview' was developed for eliciting evidence relating to alcohol problems. Thirty-six volunteer male patients on their first visits to a specialist alcohol clinic were interviewed three times, by two psychiatrists and by the computer; information was sought about 72 pre-defined indicants concerning alcohol consumption, drinking behaviour, and symptoms. Each patient was asked to complete an attitude questionnaire anonymously. The extent of agreement between the evidence elicited by the computer and by the psychiatrists was quite high, and their estimated error rates were very similar, all between 10 per cent and 12 per cent in total. With respects to amounts of alcohol consumed, patients reported significantly greater amounts to the computer than they reported to the psychiatrists. The median amounts of pure ethanol consumed ranged from 1-19 kg per week calculated from reports made to one of the psychiatrists, up to 1-58 kg per week calculated from reports made to the computer. The results from the attitude questionnaire indicated a high level of acceptability to patients of computer interrogation.