DUCK system: a physician/computer interface for clinical practice

Medical informatics systems (MIS) have been in development and operation for more than 20 years, and have been applied to medical record keeping as hospital information systems (HIS), clinical information systems (CIS), expert systems (ES) and practice analysis systems (PAS). Their overall goal has been to control the growth of medical data in this century and to define and implement models of the medical decision making process. Current efforts include developing a physician workstation to provide physicians with clinical tools for data management and decision-support. Models include Explorer I, the Cimino/Barnett Model and ONCOCIN. These efforts have generally focused on providing the physician decision-support programs on fast, user-friendly hardware (e.g., improved graphics, 12-16 mHz clock rates, mouse). None of them, however, has emphasized data management as the key to gaining routine use of the computer by clinicians. The DUCK (direct use of computer knowledge) System is proposed as an interactive physician/computer interface designed to overcome resistance by physicians in using the computer for routine clinical charting. Its design is based on a formal analysis of the factors that influence clinical data management, and its performance is based on meeting the operational needs of clinical practice. The DUCK System is meant to function as the centerpiece of any physician workstation in the context of a larger clinical information system. The approach taken by the DUCK System in gaining physician acceptance is based on the failures of earlier MIS projects that have emphasized decision-support needs. Other workstation models seem to be making this same mistake when they assume that the computer cannot compete with the handwritten record as a data management tool, despite evidence to the contrary. This study evaluates how well the DUCK System competes with the standard handwritten system. In particular, it evaluates the performance of five physicians using the DUCK System in an ambulatory care medicine setting. The experiment compares their handwritten charts with the charts generated by the computer system. The results show that physicians without computer experience are able to create medical records for routine clinical problems using a computer system. Analysis of the computer records indicates that they contain the same amount and quality of information, and were completed as rapidly as the handwritten charts. This is demonstrated in terms of the number and quality of items found in the charts, the amount of time it took to complete the charts and the rate of data less than 20 computer charts, and were not dependent on the physician's computer expertise or level of dexterity.