Decentralized Management of Patient Profiles and Trajectories through Semantic Web Agents
Download as PPTX, PDF0 likes193 views
This document discusses managing patient health trajectories through decentralized semantic web agents. It addresses the challenges of modeling dynamic patient trajectories from heterogeneous data sources while ensuring data privacy and integration. An ontology is proposed to semantically represent medical concepts within patient trajectories. Agents would cooperate to acquire, aggregate, analyze and predict trajectory data in a decentralized manner. Future work involves further developing the trajectory modeling and validating the agent-based approach through implementation.
![16
Example: Agent protocols
Coaching
τAgent
Processing
τAgent
request prediction
inform-result
[trajectory prediction data]
Patient
τAgent
consent
call-for-data
propose
accept
inform-result
self-organizing
protocol
data request
data stream
Patient
τAgents
consent
call-for-retrospective-data
propose
accept
Therapist
τAgent
Sensor
τAgents
[trajectory sensory data]
inform-result
[trajectory data]
notifynotify
[coaching advice] [patient progress]](https://image.slidesharecdn.com/presentationswh2020novideo-201102165219/85/Decentralized-Management-of-Patient-Profiles-and-Trajectories-through-Semantic-Web-Agents-16-320.jpg)
Decentralized Management of Patient Profiles and Trajectories through Semantic Web Agents
- 1. Decentralized Management of Patient Profiles and Trajectories through Semantic Web Agents Jean-Paul Calbimonte • Davide Calvaresi • Michael Schumacher University of Applied Sciences and Arts Western Switzerland (HES-SO Valais-Wallis) 3rd International Workshop on Semantic Web Meets Health Data Management International Semantic Web Conference ISWC 2020 November 2020 @jpcik
- 2. 2 Context: long-term patient support Icons:Freepik.com,Smashicons,Flaticon,Eucalyp sustained support • chronic diseases • post-surgery • cancer survivors diagnosis treatment surgery eHealth patient support Distributed care + heterogeneous clinical history
- 3. 3 Patient Trajectorieswellbeingdistress diagnosis therapy surgery discharge physical psychological social timeline Illness trajectory Patient trajectory patient context emotional/ social indicators self-reported outcomes monitoring observations diagnosis treatment events
- 4. 4 Case scenario: Cancer survivors support Patients-centered Survivorship care plan after Cancer treatments based on Big Data and Artificial Intelligence technologies diagnosis chemo surgery physical rehab survivor trajectory AIsupporttech Icons:Freepik.com,Flaticon,Eucalyp
- 5. 5 Issues in trajectory management Trajectory information heterogeneity Patient data sources diversity Trajectory data integration & aggregation Icons:Freepik.com,Eucalyp demographics diagnosis morbidities treatments . . . sensors AI self-reported EHR . . .
- 6. 6 Issues in trajectory management Life-long dynamic trajectories Data analytics explainability Privacy and confidentiality trend analysis pattern discovery trajectory classification explainable outcomes Icons:Freepik.com,Eucalyp cohort trajectories dataprivacy+trust+ compliance
- 7. 7 Challenges 1 Trajectory modelling Decentralization Cooperation 2 3 Data volume & velocity Data privacy 5 4
- 8. 8 Ontology-based trajectory modelling schema:Patientgender schema:MedicalSymptom schema:MedicalCause schema:MedicalTherapy schema:MedicalIndication schema:DoseSchedule schema:Drug schema:MedicalEntity schema:MedicalCause schema:DDxElement schema:Drug schema:MedicalRiskFactor schema:MedicalConditionStage schema:MedicalTherapy schema:AnatomicStructure schema:MedicalCondition schema:Drug schema:MedicalCondition Calbimonte, J.P., Dubosson, F., Schumacher, M.: Towards semantic models for proceeding and behavior change in ehealth applications. In: Proc. of the PD Track of the 15th Intl. Conference on Semantic Systems SEMANTiCS. pp. 41–45 (2019)
- 10. 10 Decentralized Data Management Icons:Freepik.com ... user profile state machine domain model comp. persuasion privacy sphere coordinator agent mediator agents coach agent behavior agent Behavior change agent useragents users bot interactions inter-agent interactions Calbimonte, J.P., Calvaresi, D., Dubosson, F., Schumacher, M.: Towards profile and domain modelling in agent-based applications for behavior change. In: Intl Conf on Practical Applications of Agents and Multi-Agent Sys. pp. 16–28 (2019)
- 11. 11 Inter-agent cooperation & negotiation trajectories beliefs behaviors τAgent τAgent τAgent τAgent trajectory data acquisition beliefs metadata health ontologies constraints privacy policies behaviors negotiation cooperation reasoning ML gateways role-specific goals scheduler channels MLprocessing τAgent ontologies trajectory negotiation trajectory data analysis trajectory aggregation channels Calvaresi, D., Schumacher, M., Calbimonte, J.P.: Agent-based modeling for ontology-driven analysis of patient trajectories. J. Med. Systems 44(9), 1–11 (2020)
- 12. 12 Inter-agent cooperation & negotiation Patient τAgent Coaching τAgent Aggregator τAgent Processing τAgentPatient τAgents Sensor τAgents trajectory data data acquisition support + recommendation pre-processed trajectories predictions, analytics, classification trajectory data Calvaresi, D., Schumacher, M., Calbimonte, J.P.: Agent-based modeling for ontology-driven analysis of patient trajectories. J. Med. Systems 44(9), 1–11 (2020)
- 13. 13 Real-time stream processing agents Icons:Freepik.com RSP Agent mailbox beliefs behavior RSP Agent RSP Agent RSP Agent RSP Agent RDF stream metadata metadata RDF stream beliefs metadata stream windows background RDF ontology TBox behavior message dispatch RDF stream pull RDF stream push RSP querying reasoning RT constraints goals scheduler
- 14. 14 Privacy-aware agents for data exchange Icons:Freepik.com Subject Processor Recipient consent consent Controller analytics request Calvaresi, D., Calbimonte, J.P.: Real-time compliant stream processing agents for physical rehabilitation. Sensors20(3), 746 (2020) Subject
- 15. 15 Example: Trajectory predictiondistress diagnosis therapy surgery predictive trajectory physical psychological social timeline schema: MedicalSymptom schema: MedicalCondition schema: MedicalTherapy schema: SurgicalProcedure schema: MedicalCondition schema: Patient retrospective trajectory trajectory 2
- 16. 16 Example: Agent protocols Coaching τAgent Processing τAgent request prediction inform-result [trajectory prediction data] Patient τAgent consent call-for-data propose accept inform-result self-organizing protocol data request data stream Patient τAgents consent call-for-retrospective-data propose accept Therapist τAgent Sensor τAgents [trajectory sensory data] inform-result [trajectory data] notifynotify [coaching advice] [patient progress]
- 17. 17 Future work Trajectory modelling Implementation & validation Decentralized explainability Trust & accountability Patient empowerment
- 18. Questions? Jean-Paul Calbimonte University of Applied Sciences and Arts Western Switzerland HES-SO Valais-Wallis @jpcik