Case-Based Reasoning in the Health Sciences: Foundations and Research Directions - Computational Intelligence in Healthcare: Advanced Methodologies - page 123

Fig. 3. Data supporting the importance of interdisciplinary applications, from the 2008
Journal Citation Reports in the ISI Web of Knowledge [18]
endeavors. It is notable that the Science Citation Index of the Institute for Scien-
tific Information (ISI) Web of Knowledge lists among computer science categories
a specialty called “Computer science, Interdisciplinary applications” [18]. More-
over, this area of computer science ranks highest in the discipline in terms of the
number of articles published, and second highest in the number of total citations.
These statistics, shown in Figure 3, support the other data pointing toward the
importance of applications in computer science. Health sciences applications in
particular, which are based on complex interdisciplinary work, represent a major
specialization area in computer science.
4 CBR in Health Sciences Domains: Rationale and Roots
CBR has proven to be especially useful for problem solving and decision support
applications in health sciences domains. Motivations for applying CBR to these
domains have been present from the beginning, with the goal of improving public
health. This application area has provided a rich experimental environment for
evaluating and enhancing the CBR paradigm and for expanding CBR system
development beyond the laboratory setting. There are many reasons why CBR
is a good fit for health sciences applications, including:
- Case histories are established educational tools for health care professionals
- Health care professionals publish accounts of the treatments of individual
patients in the medical literature
- Reasoning from examples comes naturally to health care professionals
- A complex biological system like the human body is hard to completely
formalize, but there are many examples, or cases, of diseases, treatments
and outcomes
- Many diseases are not well-enough understood for there to be cures or uni-
versally applicable specific treatments
- Cases can complement general treatment guidelines to support personalized
medical care for individuals
- Extensive data stores are available, such as those in electronic health records,
which may be mined for case examples
In this section, we describe early work in CBR in the Health Sciences. Current
trends and research directions are discussed in the following sections. The earliest