15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 1 15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 1 CLARIISA, a Context-Aware Framework Based on Geolocation for a Health Care Governance System LEONARDO GARDINI & MAURO OLIVEIRA LAR-A Computer Network Laboratory of Aracati
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 2 15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 2 Leonardo Gardini State University of Ceara (UECE) Fortaleza, Brazil Reinaldo Braga Federal University of Ceara (UFC) Fortaleza, Brazil José Bringel Federal University of Ceara (UFC) Fortaleza, Brazil Carina Oliveira Federal University of Ceara (UFC) Fortaleza, Brazil Rossana Andrade Federal University of Ceara (UFC) Fortaleza, Brazil Hervé Martin Joseph Fourier University (UJF) Grenoble, France Odorico Andrade Federal University of Ceara (UFC) Fortaleza, Brazil Mauro Oliveira Federal Institute of Ceará (IFCE) Aracati, Brazil CLARIISA, a Context-Aware Framework Based on Geolocation for a Health Care Governance System Team working on this project
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 3 Contextualization 2. LARIISA v1.0 – Ontology Approach 3. LARIISA v1.1 – Bayesian Approach 4. Prototyping & Tests Conclusion Summary
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 4 Health System - Information Era Based on Disease PREVENTION Costs ¢ Encouraged Discouraged Fonte: K. Jennings, K. Miller, S. Materna (1997) Hospital (specialits) Health Agent Primary Health Care $ CONTEXTUALIZATION Descentralization of Public Health System
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 5 PROBLEM Primary Health Care Primary Health Care Hospital (specialits) Hospital (specialits) Health Agent Health Agent Management Information Message Data Acquisition CONTEXTUALIZATION Increasing Complexity of Health Management Increasing Complexity of Health Management
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 6 CONTEXT-AWARE FRAMEWORK SOLUTION ONTOLOGY BAYESIAN NETWORK Metadata Geolocation Decision- Making Information Health Knowledge Inference Mechanism PROBLEM CONTEXTUALIZATION LARIISA: an Intelligent System to support decision-making process
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 7 2. LARIISA v1.0: Ontology Approach
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 8 Medical Sensors Humidity Sensor Atmospheric Pressure Sensor Temperature Sensor Digital Camera Global Positioning System (GPS) Accelerometer Internet Connection Light Sensor Proximity Sensor Compass Gyroscope Geographical Information System (GIS) LARIISA: a Context-Aware Framework
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 9 15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 9 LARIISA: a Context-Aware Framework The photo IMG001 was taken at lat=S 3° 45' “, lon=W 38° 36' “ on February 2nd, 2013 at 8:00AM. Address =Av. G, Conj. Ceará, Fortaleza – CE. Local Temperatura=29°C. Comment= Dengue Habitat. 80%-90% likelihood of being with Dengue (Local Context) Information: lat=S 3° 45' “, lon=W 38° 36' “ on March 3rd, 2013 at 5:00PM. Body temperature=40°C, Heart rate=110 bpm, Blood pressure=140/90. Address=Av. F, Conj. Ceará, Fortaleza – CE. Local Temperature=25°C. Symptoms=Headache, Vomiting, Body aches. Patient B Patient A Health Agent Relocating a health agent for the Patient B’s house Information: lat=S 2° 22' “, lon=W 34° 33' “ on March 24th, 2013 at 3:00PM. Body temperature=37°C, Heart rate=90 bpm, Blood pressure=120/80. Address= Av. da Sé, n° 227, Conj. Palmeiras, Fortaleza – CE. Local Temperature=32°C. Symptoms=Chills, Diarrhea. Who is the patient? Are Data Structured? Who is the patient? Are Data Structured?
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 10 3° 45' " 38° 36' " 17:00 03/02/2013 Av. F, Conj. Ceará, Fortaleza - CE A, B, C 40°C 110bpm 140/90,,,,,,,,,, metadata file Building a metadata file Geolocation Patient Identification via Web Service Health Information
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 11 Context Providers Data Processing Data Acquisition Publishing User Device Internet Health Agent Device Symptoms + sus_id Global Context Local Context Inference Rules Context Aggregator (CA) Health Managers System Security Protocol Metadata LARIISA’s Architecture: a context-aware framework
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 12 Knowledge Representation (ONTOLOGY) Data Acquisition (CONTEXT) DECISION-MAKING LARIISA’s Scenario: a context-aware framework METADATA Step 1 Step 3 Step 2 LARIISA: A Context-Aware Framework LARIISA: A Context-Aware Framework
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 13 Local health context model Global health context model Prototype: Dengue Fever Case Study Metadata
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide LARIISA v1.1: Bayesian Approach (Probabilistic Data)
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 15 LARIISA: Bayesian Approach Specialist Data (Tables) and Relationship
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 16 Probabilistc Methods (BAYESIAN NETWORK) Data Acquisition (CONTEXT) DECISION-MAKING METADATA LARIISA: Bayesian Approach
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 17 IN LARIISA_BAY Inference Module Patient Health Agent Specialist Decision Module Interface Specialist Decision Module 1 1 Specialist Decision: f(%) 3 3 Pass Through A f(%) A = A’ 2 2 Specialist Validation: A f(%) A ≠ A’ RB % SITUATION ROOM Health Agent OUT A’ B’ C’ A’ B’ C’ C B A A C B OTHER CONTEXT PROVIDERS A’ B’ C’ LARIISA INFERENCE MODULE Sensors Health Center Ambulance METADATA User Interface Inference Rules Global Context Repository Local Context Repository Patient Specialist Epidemic Graph Manager LARIISA: Functional Diagram
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide Prototyping & Tests
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 19 ENTRADA DO SISTEMA Módulo de Inferência do LARIISA_Bay Paciente Agente de Saúde Especialista Interface Módulo de Decisão Módulo de Decisão 1 1 Decisão do Especialista: f(%) 3 3 Pass Through A f(%) A = A’ 2 2 Validação do Especialista: A f(%) A ≠ A’ RB % SALA DE SITUAÇÃO Agente de Saúde SAÍDA DO SISTEMA A’ B’ C’ A’ B’ C’ C B A A C B OUTROS PROVEDORES DE CONTEXTO A’ B’ C’ LARIISA LARIISA_Bay Sensores Posto de Saúde Ambulância METADADO Interface do Usuário Regras de Inferência Repositório de Contexto Global Repositório de Contexto Local Paciente Especialista Gráfico de Epidemias Gestor Screens of the proposed System Patient Health Agent Specialist
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 20 Low Risk High Risk Bayesian Networks
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 21 Health Agent Pacient Smartphone / Tablet / Desktop,,,,,,,,,, Lariisa Database WEB Prototype Health Managers Dashboard Structured Data Internet
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 22 Conclusion
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 23 ONTOLOGY Approach BAYESIAN Approach LARIISA v1.2 Hybrid Approach LARIISA: Next Generation
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 24 ONTOLOGY & BAYESIAN NETWORKS Data Acquisition (CONTEXT) METADATA DECISION-MAKING LARIISA: Next Generation
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 25 Conclusão Diga Saude (FUNCAP) SISA (FIOCRUZ) LARIISA (IFCE) GISSA (FINEP) Next Saude (DATASUS / Min Saúde) Submitted: PPSUS (FUNCAP / Min SAUDE) Sponsors This project is being sponsored since 2004 by the Science and Technology Ministry of Brazil and others Brazilian Research Agencies It will be applied to the brazilian public health system This project is being sponsored since 2004 by the Science and Technology Ministry of Brazil and others Brazilian Research Agencies It will be applied to the brazilian public health system
15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 26 15th Healthcom October 10th, 2013 Lisbon, Portugal Slide 26 Leonardo Gardini Mauro Oliveira