Methods for knowledge engineering in the development of a computerized admission testing system (CATS)

The authors developed an expert system for routine tests ordered when an adult patient is admitted to the general medicine services of a hospital. They describe the methods for knowledge engineering used to develop a computerized admission testing system (CATS). CATS was developed in successive iterations guided by expert input on discordant results. The methods illustrate the innovative use of quantitative measures of discordance to direct the knowledge engineering process.<
>