Case-based reasoning in the health sciences: Why it matters for the health sciences and for CBR

Biomedical domains have been an application domain of choice for artificial intelligence (AI) since its pioneering years in expert systems. Some simple explanations to this phenomenon are the intellectual complexity presented by this domain, as well as the dominant industry market share of healthcare. Following in AI's tracks, case-based reasoning (CBR) has been abundantly applied to the health sciences domain and has produced an excellent as well as varied set of publications, which has fostered CBR research innovation to answer some of the research issues associated with this intricate domain. Some notable examples are synergies with other AI methodologies, and in particular with ontologies [8] and with data mining, the study of the temporal dimension in CBR, the processing of multimedia cases, and novel tasks for CBR such as parameter setting. However CBR has a major endeavor to take on in the health sciences: how to position itself with regard to statistics for studying data? Some claim that CBR proposes an alternative viewpoint on the concept of evidence in biomedicine; others that CBR and statistics complement one another in this domain. In any case, an interesting question to study is whether CBR could become one day as fundamental to the health sciences as statistics is today? This question in particular broadens the health sciences challenge to a universal scope.