During the Covid-19 pandemic, primary health care services and call centers have been criticized about their decisions on whether a patient should be hospitalized or not. The medical staff working in these services have to remotely estimate the condition of the calling citizens based on vague descriptions of their symptoms. Missing information about the patients, the congestion and the rapidly changing medical protocols as they are combined with the subjective opinion of each medical specialist working in these centers often leads to incorrect decisions. Coronario platform helps at this stage of pre-hospitalization symptoms’ assessment. With equal importance, the rehabilitation phase can also be monitored using the developed platform. A recent incident in Greece, where a patient with kidney failure spread Covid-19 virus to other patients and the personnel of a hemodialysis center, shows the significance of monitoring people that need regular medical attention with a platform like the one proposed here.
What it does
The proposed platform consists of the Coronario mobile application on the user side, connected to the cloud where all the information is stored. eHealth medical sensor infrastructure (temperature, blood pressure, body position, glucose meter, or more advanced sensors like ECG) is connected potentially to a microcontroller or uploading independently the sensor indications to the cloud. The supervisor doctor accesses the information of the cloud through Coronario Supervisor app. An medically certified AI tool can also be plugged on the supervisor side assisting the diagnosis. The end user is authenticated through the Coronario mobile app, selects the language of the UI, and downloads from the cloud the latest medical protocol and sensor strategy that has to be followed. Any messages sent by the supervisor are displayed on his screen. He can fill a symptoms' questionnaire arranged as a flat set of questions or as a decision tree. He can record and analyze the sound of his cough or his breath and upload this information to the cloud. Any messages from the supervisor, instruction on how to do the require examinations using the available sensors are also displayed in the Coronario mobile app while the user can also send text messeges to the Supervisor. The usage of the various platform modules in brief:
- Coronario Supervisor application allows the supervisor to: connect to the cloud account, select user interface language, view sensor data in readable format (information is stored in the cloud in JSON format), send a message to the end user, select a medical protocol (including the definition of the questionnaire, the sensor sampling strategy, and the generation of non-life critical alerts from rules based on sensor values).
- In the Coronario mobile app, the user selects the UI language interface, is authenticated (or registered as a new user) and connects to the cloud. The application displays the messages sent by the supervisor and downloads the latest medical protocol that has to be followed. Some personal info can be given (age, gender, habbits) and the questionnaire defined by the supervisor can be filled. It can be a flat set of questions or questions asked following a decision tree structure. The answers can be yes or no (check-boxes) or analog indications (sliders). The location of the patient is determined and his country or the prefecture is recognized. The coronavirus cases in this region are retrieved by an external webpage while his position can also be displayed on the map. This information can be used to assess the risk of being affected by covid-19. The recorded sound of his cough or breath can be analyzed in the frequency spectrum and displayed in a color canvas. The frequency power features of the spectrum can be used to classify the as a symptom of COVID -19. Finally, the sampling strategy as defined by the supervisor doctor is displayed in readable format with instructions on when or how often to perform an examination.
Implementation details about the developed application can also be found in https://www.youtube.com/watch?v=kp1unR1mQXQ
How we built it
Microsoft Azure cloud services have been used for information storage (values uploaded either from the sensors or the Coronario mobile app, medical protocols, sampling strategies, messages exchanged). The most important Azure modules used are the coronario storage and IoT test hub. JSON format is used for all the information exchanged. Although Azure can provide a wide variety of data mining, analysis and presentation tools (eg. Power BI) we preferred to use an external tool on the Supervisor side the so called Coronario Supervisor App in order to allow portablility of the platform to different cloud services (Google, Amazon AWS, etc).
Both the Coronario mobile app and the Supervisor app have been developed in C# and Microsoft Visual Studio 2019. The Coronario mobile app has been developed on Xamarin platform in order to allow its deployement either as a mobile app (for Android or iOS) or as a tablet/desktop/laptop application.
Challenges we ran into
During the EUvsVirus even, we worked in collaboration with people that we didn't know before. We exploited the valuable comments of the mentors and the coaches of EUvsVirus. Three days of brainstorming on how can the prototype system be improved: a) prioritization of the user inputs in oder to reach valid diagnosis results, b) adaptability of the system to new medical protocols, infections, regional settings without needing to rebuild the applications, c) ergonomic design in order to be accepted by both patients and medical staff that is not necessarily qualified in computer science, d) achievement of financial benefits for all the participants (patients, medical centers and health institutions, supervisor doctors, developers of this application)
Accomplishments that we're proud of
• Mature design. 80% of what is described here has been implemented! • Reduction in the congestion of primary healthcare units • Will be used to monitor post-hospitalization, gathering critical statistical information about the health of people that have recovered from Covid-19 • Will be used to monitor vulnerable population (patients with cancer or heart diseases, diabetics, transplanted) • Unified treatment rules that are applied to all patients • The protection of the patient privacy is guaranteed (GDPR compliant) • The system is open to extensions in order to support other similar infections, different mobile and cloud platforms and different regional settings including language of the UI and authentication methods
What's next for Coronario
- Find investors/sponsors
- It has to be tested in real environments such as primary health care units that will be our research partners
- Official medical protocols have to be tested
- Its usage with patients in the rehabilitation phase or patients that need regular medical attention has to be tested
- The UI interface will be made more user friendly hiding unnecessary details
- Authentication methods common to all EU countries such as eIDAS will be supported
- The application has been developed and tested using Azure cloud services and Android phones. It will be extended to support other mobile and cloud services
- It will also support more diverse regional settings for example the languages of the user interface
- Finally it will also be tested for different medical cases