The Necessary Forms of the Knowledge

2.0 Abstract

A patient’s presentation for diagnosis generally calls for ascertainment (and documentation) of a multidimensional diagnostic profile (Sect. 1.2). In the context of a given type of presentation with sickness (chief complaint, demographic category), the set of diagnostic indicators – a dozen in number, say – definitional to the various possible case profiles generally implies an enormous number of possible profiles for diagnosis about any particular one of the illnesses that could be present, for setting the probability of its presence in the context of the profile at hand. It thus generally is wholly impractical to (learn and) codify the probability for a particular illness separately for each of the possible diagnostic profiles that are possible in the context of a given type of patient presentation.

The need thus is to codify the diagnostic probability for each of the possible illnesses as a function – joint function – of the set of diagnostic indicators, defining subdomains of the presentation domain. The idea (in ‘clinical epidemiology’) that diagnostic probability can be based on the diagnostic ‘accuracy’ of each of the diagnostic indicators is a serious misunderstanding.

Analogous probability functions are needed for etiognosis and prognosis just the same.