THERE IS a pressing need for objective and repeatable measures of the day-to-day behavior of psychiatric patients. Traditional clinical methods based on global impressions, such as nurses' notes, suffer from incompleteness and unreliability. Improved information and reliability can sometimes be obtained by careful rating of observable behavior. We have previously reported a computerized system using nurses' observations that promises to make the rating approach less costly and more effective.1 Concurrently, we have been working to develop longitudinal measures of cognition. We have been exploring the value of an automated procedure for measuring the acquisition and immediate retention of a ten-member response chain.
Considerable progress has been made by others2,3 in the development of performance measures to assess the status of psychiatric patients. The sensitivity of these measures is necessarily limited by the simplicity of the demands made upon the subjects. It is also true that cognitive performance assessments