Portal de Periódicos da CAPES

Computerized psychiatric diagnosis in the elderly: AGECAT

1986; Elsevier BV; Volume: 9; Issue: 2 Linguagem: Inglês

10.1016/0745-7138(86)90038-2

1096-374X

Michael Dewey, J. R. M. Copeland,

Machine Learning in Healthcare

This paper describes a computerized diagnostic system, AGECAT (Automated Geriatric Examination for Computer Assisted Taxonomy), designed for use with the Geriatric Mental State Schedule (GMS). AGECAT can be divided into three main parts. In the first part symptoms are aggregated to form symptom components. In the second part these components are grouped and compared, using a logical decision tree approach, to reach levels of confidence on eight syndrome clusters. In the third part these levels are compared, cluster by cluster and if the level of confidence is sufficiently high a decision is reached on the diagnosis. In addition the program provides a variety of other information on alternative diagnoses and any unusual features detected. Developing a portable program for this type of work presents an interesting computing task involving various programming tools. One of the benefits of computer methods is the way in which they enforce a precise statement of the theory involved. This advantage is lost if, as is usual, the version which humans read is a hand translation of the computerized version. The implementation of AGECAT has been designed to access a single master text which is used to produce all the target versions automatically, and the paper considers some of the advantages of this approach.