PNM pattern-based characterization of the clinical profiles opens the "missing half" of the information content in standard data, which are used for clinical prognostication: The information supporting clinical decisions can be based both on the PRESENCE as well as on the ABSENCE of characteristic clinical patterns:
The information gain from using PNM stems from preserving the coherence of all the data collected together for an individual patient. This is the largest information resource that PNM can utilize in contrast to conventional methods, which treat every individual clinical variable independently. After capturing the patient's personalized clinical profile pattern into the PNM graph, it is "just" mathematics. Please note that this mathematics has serious foundations (proved theorems, information theory arguments, active component through entropy maximization and more ...), which contribute to the non-empirical, novel, but still directly clinically interpretable results.
