e-health case scenario: comorbidity

This work addresses the generation of a personalized treatment plan from multiple clinical guidelines, for a patient with multiple diseases (comorbid patient), as a multi-agent cooperative planning process that provides support to collaborative medical decision-making. The proposal is based on a multi-agent planning architecture in which each agent is capable of

(1) planning a personalized treatment from a temporal Hierarchical Task Network (HTN) representation of a single-disease guideline, and

(2) coordinating with other planning agents by both sharing disease specific knowledge, and resolving the eventual conflicts that may arise when conciliating different guidelines by merging single-disease treatment plans.

The architecture follows a life cycle that starting from a common specification of the main high-level steps of a treatment for a given comorbid patient, results in a detailed treatment plan without harmful interactions among the single-disease personalized treatments.

Fig. 1. Architecture diagram

Figure 1 shows the proposed Multi-Agent Planning (MAP) architecture that is composed of an Initiator Agent and several Planning Agents. A planning agent can be seen as the representation of a clinical specialist capable of (1) planning a personalized, single-disease care plan (HTN planner module); (2) coordinating with other specialists, by sharing its single-disease recommendations and experience (Coordinator module) and (3) detecting and resolving conflicts between its recommendations and those of others (Conflict Solver module).

The mission of the Initiator Agent is to send the global problem to the Planning Agents. A global problem may be defined by a starting time point and a list with the high-level steps of the care plan. For instance, the phases of diagnosis, treatment and follow-up have to be designed for a patient starting from a specific date.

Note that the process of designing a treatment plan for a comorbid patient is a cooperative planning process that iterates over the high-level tasks of the global problem and for each task a joint treatment plan is obtained. Such joint plan is generated by a merging process developed by the Conflict Solver module in which single-disease treatments (automatically generated by HTN planners) are combined in order to detect and solve eventual interactions.

Relevant papers:

Gonzalo Milla-Millán, Juan Fdez-Olivares, Inmaculada Sánchez-Garzón
A Common-recipe and Conflict-solving MAP Approach for Care Planning in Comorbid Patients
XV Conferencia de la Asociación Española para la Inteligencia Artificial (CAEPIA'13), LNAI 8109, pp. 178-187, (2013)

Inmaculada Sánchez-Garzón, Juan Fdez-Olivares, Eva Onaindía, Gonzalo Milla-Millán, Jaume Jordán, Pablo Castejón
A multi-agent planning approach for the generation of personalized treatment plans of comorbid patients
14th Conference on Artificial Intelligence in Medicine, Vol. 7885 pp. 23-27, 2013