The Auguste Project:
Case-Based and Multimodal Reasoning in Planning Care
for Alzheimer's Disease Patients

Cynthia R. Marling, PhD
School of Electrical Engineering and Computer Science, Ohio University
and
Peter J. Whitehouse, MD, PhD
University Alzheimer Center and School of Medicine, Case Western Reserve University


Case-Based Reasoning (CBR) is an Artificial Intelligence (AI) paradigm in which new problems are solved by retrieving and adapting the solutions to old problems. It is becoming increasingly useful in combination with other reasoning strategies in multimodal reasoning (MMR) systems. The medical domain is of interest for CBR research, because physicians naturally employ CBR when they use past diagnoses and treatment plans to guide them in making new ones, and because medical education relies heavily on cases to illustrate medical problems. MMR comes into the picture as physicians use heuristics and probabilistic reasoning, as well as past experience, in solving problems.

Here you see Auguste D., the first documented Alzheimer's Disease (AD) patient. She was treated by Dr. Alois Alzheimer, shown at right, in Frankfurt, Germany, between 1901 and 1906.

AD is a degenerative brain disorder characterized by memory loss, personality change, unusual behavior, and a decline in thinking abilities. It is estimated to affect approximately four million elderly Americans and as many as twelve million people worldwide. While the disease can not be predicted, prevented or cured, interdisciplinary teams of health care practitioners, including physicians, nurses and social workers, plan care to maximize quality of life for AD patients and their families.

Planning the ongoing care of AD patients is a challenging problem, which is not presently supported by computer-based decision support tools. The goals of the Auguste Project are to understand the reasoning employed in this task and to build practical decision support tools to ease the task. Research challenges arise from the need to: incorporate the multiple perspectives of members of different health care disciplines; plan flexible care over a period of several years; and support ethical considerations in decision making.

A prototype of the first CBR decision support tool has been completed. This tool will assist with the decision to prescribe neuroleptic drugs for behavioral problems in AD patients. The prototype was implemented by Amit Jain.

Extensive information is available about Case-Based Reasoning, Alzheimer's Disease and Geriatric Interdisciplinary Teams.

The paper "Knowledge Sharing and Case-Based Reasoning in Geriatric Care", from the AAAI-99 Workshop on Exploring Synergies of Knowledge Management and Case-Based Reasoning, is available in PostScript.