Reference: Shahar, Y. & Musen, M. A. Plan Recognition and Revision in Support of Guideline-Based Care. Knowledge Systems Laboratory, Medical Computer Science, Stanford CA, 1994.
Abstract: Clinical guidelines are a common format in medical domains for representing therapy guidelines, by prescribing a set of rules and policies that the attending physician should follow. Clinical guidelines can be viewed as a shared library of highly reusable skeletal reactive plans, whose details need to be refined by the executing planner over significant periods of time. Guidelines involve collecting and interpreting patient-related data, applying prespecified plans, and revising the plans when necessary. Reactive-planning architectures specific for the task of refining skeletal plans over time have been implemented; their importance is increasing as more clinical data are being captured and represented in an electronic format, and as quality control of medical care grows in importance. Conducting a flexible, intelligent dialog with a human user of such systems requires an ability to reason about the user's goals and plans and about possible modifications to these plans. Automated support for guideline-based therapy can be viewed as a collaborative effort between two planners: a human one and an automated assistant, whose knowledge might be limited, or which might not have access to all the data. Achieving even a modicum of collaboration and of the highly-desirable flexibility involves reasoning about mental states: a recognition of each planner's intentions and plans to achieve them, and a consideration of the available plan-revision strategies. Clinical guidelines could be enhanced considerably by a sharable, explicit, formal representation (1) of therapy-planning-operators' effects, (2) of plan-revision strategies, and (3) of the underlying goals and policies of the guideline as temporal-abstraction patterns to be maintained, achieved, or avoided.
Notes: Working Notes.