Medical knowledge for clinical problem solving: a structural analysis of clinical questions.

Despite technological advances that support wide-ranging access to and transfer of knowledge, practicing physicians continue to underutilize current biomedical literature. This paper explores the nature of clinically applicable medical knowledge through a structural analysis of clinical questions. The author analyzed a set of sixty questions, based on actual online search requests of practicing physicians, for stated and unstated needs, certainty levels, implicit and explicit assumptions, decision-making processes, and type of answer required. As a result, four states of information valuable in patient care were identified: prediagnostic assessment, diagnosis, treatment choice, and learning. These states are presented in frame-like structures that integrate declarative and procedural components of medical decision making. It is concluded that clinical problem solving requires a blend of declarative and procedural knowledge. The ratio depends, in part, upon the reasoning process underway at the time of the request. Procedural knowledge required for clinical problem solving may be absent from current biomedical journal literature or difficult to identify.