Kashyap, Vipul wrote:
Based on conversations on this topic, there appears to be consensus of the need for multi-layered knowledge representation schemes for heatlhcare.  Will be great if we could brainstorm and come to some sort of consensus on these "layers". Would like to propose a strawman as enumerated below.
 
Layer 0 = Entity - Attribute - Value or RDF triple based rerpesentations.
Layer 1 = MetaClasses, e.g., Observation as in HL7/RIM
Layer 2 = Classes in a Patient Model, Document Models, etc, e.g., the class of HbA1c results for a class of Patients.
Layer 3 = Data that are instances of Classes, e.g., a particular HbA1c result for a patient John...
A term from upper ontology work that is handy for discussing ontological layers is "discriminant".  By this I intend to mean a general predicate, able to usefully characterize something along one particular semantic dimension.  Examples of important discriminants would include:

    Physical vrs Informational
    Natural vrs Artificial
    Real vrs Imaginary
    Composite vrs Characteristic
    Individual vrs Collective
    Atomic vrs Mediating
    Specific vrs Indefinite
    Continuant vrs Occurrent

Not many discriminants can be found which are simultaneously orthogonal (independent of one another) and general (can be applied to anything) and useful (easy to clearly define and decide).   The listing above may in fact be nearly complete (although many would debate its specifics or suggest other candidates).

Regardless of the particulars, I suggest that a better semantic model for your "layer 0" would be all and only those discriminants which have all three qualities - independence, generality, and utility - and hence can be employed to help define any class or instance desired.

Perhaps the specific discriminants listed above could help in more precisely discussing the less-general (domain-specific) layers of your strawman, or the benefits of distinguishing them.

Best regards,
Dan Corwin