Ontological Evaluation of Health Models: Some Early Findings

Incomplete or incorrect systems analysis can result in expensive rework at a later stage in system development. It is in the interests of all parties involved in a system development endeavour to get the data model right before development begins. The General Practice Computing Group (GPCG) data model and Core Data Set is an emerging industry standard for General Practice systems. It describes and documents clinical activities in general practice, facilitates the exchange of information between general practice and the broader health system as well as providing a foundation for the development of general practice computing applications. Emergence of the Bunge-Wand-Weber (BWW) ontological model introduced a platform to classify and compare the grammar of conceptual modelling languages. It established a systematic and theoretical basis for the evaluation of grammars. This approach has been adapted in an earlier study, which evaluated a specific instance of a conceptual model rather than the model itself. The model evaluated was in the health domain. The work presented here follows the approach of the earlier study but takes it one step further by gathering empirical evidence to test the hypothesis that ontologically correct models more readily support different users conceptions of a domain. Grounding a model according to ontological principles is an important step in establishing how well it represents reality. This is something that should be undertaken alongside the more ambitious aim of realising the vision of a life-long, portable, fully integrated, health record, and implementing it on a global scale. The work uses the Bunge-Wand-Weber (BWW) model of ontological correctness. This approach is being used increasingly for the validation of conceptual models. This study differs from earlier ones in that it uses the BWW model to identify problems in an instance of a data model rather than the modelling grammar itself. This paper sets out the theoretical foundations for evaluating this model and presents the results of a pilot study