Mon, Aug 21, 2017

MobiGuide System Demonstrated in Brussels

the grand palace

The Grand Place, Brussels.

The project team demonstrated the new version of the MobiGuide system at the European Commission in Brussels in January. The demonstration included applications for Atrial Fibrillation (AF) and Gestational Diabetes Mellitus (GDM) and showed the functionality and interfaces for caregivers as well as for patients. The occasion was the third annual project review, held at the European Commission on 21st and 22nd January. Over a day and a half the team presented the results and deliverables of the third project year to the EC Reviewers as well as demonstrating the system as answering the Reviewers’ questions.

Atrial Fibrillation demonstration

Dr. Carlo Napolitano, cardiologist at Fondazione Salvatore Maugeri (FSM) in Pavia, demonstrated the caregiver interface of the MobiGuide system for the AF case, showing how the cardiologist can manage their patients with the help of MobiGuide.

2 brussels

Atrial Fibrillation application demonstrated by cardiologist Dr Carlo Napolitano.

The patient system, which runs on a personal Body Area Network consisting of a smartphone and sensors, allows the patient to manage their condition from home or while on the move, under the guidance of clinical knowledge adapted to their individual treatment plan and situation. AF patients can record incidence and severity of symptoms on their smartphone and receive appropriate guideline-based but personalized advice from a distributed decision support system. The knowledge in the decision support system is based on a published clinical guideline for management of atrial fibrillation. The knowledge in this guideline, which is written in English, has been formalised to produce a computer interpretable version of the guideline (a Computer Interpretable Guideline or CIG). This CIG is customized to include psycho-social context and forms the major part of the knowledge base in the decision support system. This CIG is then personalized for each individual patient by the cardiologist working together with the patient to decide which options suit their individual health condition and situation.
Measurements such as blood pressure, heart rate, and ECG are uploaded automatically by the sensors of the Body Area Network. Other measurements, such as weight, blood pressure and INR values, dietary information and exercise can be entered manually by patients and each patient can view their own data in their personal logbook running on their smartphone.

3 b

AF patient exercise scenario
AF patients can also use the MobiGuide system to monitor their exercise sessions and give warnings and advice, for instance when they exceed their personal recommended limits. The first screenshot below shows an exercise session. The patient is running; his current activity level is showing as High, with an average level over the session of Moderate-to-Low. The second screen shows how the patient can change their settings for receiving vibration alerts when exceeding their exercise goal or receiving recommendations. The third screen shows a summary of a completed exercise session.

dd b

Gestational Diabetes Mellitus demo
The first part of the GDM demo showed how the system is used in the initial phase, when the doctor enrolls a patient into the system and together they set individual preferences such as altering default times for measuring blood glucose levels to a timetable that suits the patient. They can also record family support level (does the patient live alone or do they have support at home?) and patient preferences relating to reminders (yes/no and preferred timings) and adapt the usual routine to special contexts that the patient expects will arise from time to time such as changed daily patterns when on vacation.

e b

Endocrinologist Dr Mercedes Rigla, patient representative Àngels Pallàs and GUI developer Iñaki Martínez Sarriegui demonstrate the Gestational Diabetes Application.

In this part of the demo the endocrinologist explained the MobiGuide system to the patient and enrolled her into the system. Then the demo showed how a technical expert helps the patient activate the BAN and trains her in how to navigate through and use the functions offered by the GUI during daily management of her condition.

f b

The technical expert trains the patient in how to use the MobiGuide system.

The demo continued by showing the daily routine phase when the patient enters her daily measurements and receives guidance from the system, which helps her in her daily management of her blood glucose levels by means of diet, exercise and, in some cases, insulin therapy, according to the GDM Guideline and her current treatment regimen and context. Functions include manual entry of insulin taken and ketone levels and manual or automatic entry of blood glucose and blood pressure readings, depending on whether the patient uses BlueTooth enabled blood pressure and blood glucose sensors or not.

i b

Wherever suitable sensors that can transfer data automatically are available and affordable, they can be integrated into the MobiGuide system, making the system easier for patients to use and less error prone. However not all sensors are yet available with suitable interfaces, so some measurements must still be entered manually.
MobiGuide is quality aware and the demo illustrated how, when the patient enters a wrong value (in this case during manual entry of blood pressure), the system detects the error and displays a message querying the measurement and offering a second try at entering the value. As with AF, the patient can view their data in their personal logbook.

j b

 In the GDM demo the role of the doctor was played by the endocrinologist working on the project, Dr. Mercedes Rigla of Corporacio Sanitaria Parc Tauli de Sabadell (CSPT), and the patient was played by the MobiGuide patient representative, Àngels Pallàs of Associacio de Diabetics de Catalunya (ADC). The technical support role was played by Iñaki Martínez Sarriegui of Universidad Politecnica de Madrid (UPM), one of the technical developers on the project who is responsible for designing and implementing the GUI.

MobiGuide: a complex distributed system

Lying behind the end-user applications is a complex distributed system involving a number of sophisticated components collaborating together. Each of the components and their interworking was presented in the Work Package presentations in Brussels. On the first day of the Review, formalization and knowledge engineering of the clinical knowledge from the clinical practice guidelines for AF and GDM, development of a distributed decision support system, intelligent data analysis, semantic data integration and mechanisms for ensuring security and data privacy were covered. Presentations on the second review day addressed dissemination and exploitation, and training and project management Work Packages and planning for the fourth and final year of the project.

k b

Yuval Shahar of Ben Gurion University presents WP6: Intelligent Data Analysis.

Seven of the 12 Work Packages of MobiGuide were active in Year 3; each work package manager presented progress and results. Plans for the final year including the upcoming pilots with AF patients in Italy and GDM patients in Spain were also presented. The doctors’ perspectives were presented by Dr Mercedes Rigla and Dr Carlo Napolitano. The patients’ perspective was presented by Àngels Pallàs.
Project coordinator Mor Peleg of University of Haifa summed up the year, saying that the main objectives in Year 3 were: engagement of patients and physicians, education of the public, doctors, nurses and patients about MobiGuide, achieving secure semantic integration of patient health data into a Personal Health Record, interoperating the back end components with the patient’s BAN system and realizing a personalized and distributed decision support system based on intelligent data analysis. We are very happy to say that we achieved our targets for the third year.

MobiGuide International Event in June
Plans are progressing for the MobiGuide International event, planned as a special track in the KR4HC-PROHEALTH workshop to be held on Saturday June 20th 2015 in Pavia, Italy. The KR4HC-PROHEALTH workshop will be co-located with the 15th Conference on Artificial Intelligence In Medicine (AIME) in Pavia, which runs from 17-20th June (http://aime15.aimedicine.info/). To kill three birds with one stone, and use the budget efficiently, MobiGuide will also hold its next Consortium Meeting in Pavia during the same week.

Look out in March for the call for papers for the MobiGuide event and join us in Pavia!