Considerable efforts are being invested in the attempt to build bridges between molecular and clinical information by exploiting the potentialities of software tools for information retrieval and manipulation. Unfortunately, our existing systems for the coding of clinical data are not up to the task of serving as the clinical platform for the processing of new types of patient-centred biomedical information.

This is because such systems result from the application of a concept-centred methodology. Their terms and definitions echo the medical dictionaries of an earlier era in that they represent not the objects and processes in an ever-changing biological reality but rather word-meanings linked together via associative relations which lack coherent logical definitions.

Such systems provide too little support for the use of logic-based reasoning techniques for leveraging the biochemical and patient-centered information annotated in their terms - and it is for this reason that attempts to bridge the threshold between biochemical and clinical phenomena have thus far relied predominantly on associative, pattern-recognition methodologies. Description Logics are beginning to play a role in supporting more sophisticated methods for hypothesis formulation and testing; but they too can reap their full potential only to the degree that the problems associated with the concept-centred methodology are overcome.

Time and granularity are two central axes in the organization of the biomedical realm, both of which are recalcitrant to treatment in terms of the concept-centred methodology that has predominated hitherto. Concepts themselves are static entities, where the phenomena we are attempting to represent in clinico-biological research are pervasively dynamic in nature. Moreover, our existing ways of organizing life science data rest upon a separation into the different granularities of the gene, protein, cell, organism, and so forth. In the absence of a coherent logical treatment of time and granularity, however, we cannot do justice to disease pathways and to other clinically relevant phenomena which cross granular thresholds as they evolve through time (consider, for example the evolution of a tumor, or of a human fetus).

I will attempt to show how a clinical ontology constructed on a sound logical basis can bring benefits in the treatment of granularity and time in such a way as to support reasoning over clinico-biological information of a sort hitherto unrealized.