Data decay, significance and confidentiality: a time-oriented data-model for comprehensive care.

Long-term computer-held clinical records present new problems for data organization and for the protection of patients' privacy. To maintain clarity and to protect confidentiality, it is important that most information becomes less accessible and less prominent with time. A data-model is presented which allows the representation and manipulation of complex time-dependent relationships among data, but which imposes sufficient limits to allow efficient corrections and deletions. One useful consequence of the model is a natural analogy and potential link to the methods of production systems and 'expert systems' currently being developed by workers in artificial intelligence.