• 

  Collaboration (Plug-in+)

• 

  Collaboration (Plug-in+)

• 

  Collaboration (Plug-in+)

• 

  Collaboration (Plug-in+)

• 

  Collaboration (Plug-in+)

• 

  Collaboration (Plug-in+)

• 

  Collaboration (Plug-in+)

• 

  Collaboration (Plug-in+)

• 

  Collaboration (Plug-in+)

^ We're E.U and French citizens, but we're not native English Speakers. So, we've slowed down the speech rate to articulate each word, and be understood. Advance the speed to x1,5 or x2 to go faster, with settings of the YouTube video.

! It's not the Pitch (Mock), but our first presentation video. We'll release the demo video soon.

JUDGING CRITERIA

40% Impact Potential

WE ARE REVOLUTIONIZING EPIDEMIOLOGICAL STUDIES AND CLINICAL RESEARCH ! We are researchers, clinicians and academicians on the front lines every day in the fight against COVID-19, both in the hospital and in the research labs. Our team is mainly composed of Md-PhD and medical research directors, and WE need this solution urgently! We have been studying since the beginning of the epidemic what would be the most effective digital strategy, in compliance with European RGPDs. We decided less than 2 weeks ago to take up the challenge, after consultation with the CNIL (National Commission for Information Technology and Civil Liberties) and our other academics partners. Please contact us if you have any questions.

20% Technical Complexity & Novelty

The solution we have chosen is clearly innovative: There is no existing service with similar functionalities at the moment. And its complexity is a reflection of its great utility. The service we offer is the most complete so far, and we intend to develop it further.

Project Innovation :

• A new research paradigm
• A rich and powerful search engine
• Full-text or database searches
• And other research and clinical trials …

Automated epidemiological research!!!!!!

20% Prototype Completion

We have developed a functional proof of concept ready for beta testing. Two new full-stack developers will have to be mobilized full time. We want to hire!

20% Business Plan

Our project is non-profit. It will in fine be the property of the EMA (European Medicines Agency) and the WHO. We intend to finance ourselves through public/private partnerships. We urgently need to hire new full-time developers and a communications officer to accelerate the project !

---

A_ Inspiration

We are Md-PhD students with the support of the IHU Mediteranée Infection, the Raoult's hospital and research center in Marseille. As physicians and researchers, we are facing many problems. We decided to launch this project two weeks ago. But we've been working mostly on the scientific, academic, and legal aspects during these 2 weeks.

THE ISSUE

While a definitive way out of the crisis can only be found with a better collaboration of the people facing it, the emergency situation has turned treatment protocols upside down, leading to scientific controversy, medical legibility, and social distrust. Without proper preparation for phase IV clinical trials there is a real risk that we will not be able to emerge quickly from this pandemic.

--> What feedback can doctors get on the risk/benefit ratio of the treatments they prescribe?

B_ What it does

SOLUTION Our platform brings together patients, physicians, and researchers for the benefit of all. It organizes the medical follow-up of patients, facilitates crisis management for physicians, and provides reliable and useful data for researchers in the long-term evaluation of treatments brought to the market by accelerated procedures.

--> Patients Involved in their follow-up thanks to a clear organization and treatment plan. Our system ensures that the dosage is respected and it guarantees perfect compliance.

--> Doctors Contribute to the collection of relevant information during the follow-up of their patients. Our service guides this paradigm shift by facilitating the exchange and generation of data.

--> Researchers Analyze and process the collected data. Thanks to an intelligent solution combining algorithm and AI, relevant information on the risk/benefit balance of a treatment can be deduced.

C_ How we built it

We offer a solution that is pragmatic and, above all, comprehensive, putting patients, doctors, and researchers in synergy for the benefit of all.

Our system refines medical care by specifying the effectiveness of a treatment for a given typology of a patient found in the general population, but also a given context.

To do this, our solution collects and analyzes 3 types of data, classified according to their scientific relevance:

1- Low-Quality Data, coming from the patient himself, thanks to the new medical self-monitoring tools.

2- Medium Quality Data, coming from medical reports, thanks to a raw data extractor that classifies them in a rigid structure.

3- High-Quality Data, coming from existing official clinical trials.

It is the synergy in the linking of these data that makes the strength of our system.

We are now going to detail the 3 parts of data acquisition, in parallel with the user path:

Patients

So to make sure that even mentors who are not familiar with the world of science understand: we are going to focus here, in this presentation, on this non-technical part, potentially usable by all. (Please contact us if you are looking for more information on the rest).

We have created a patient path marked out in 9 steps, divisible into 3 sections : At the first connection, the user will be asked to register.

A- General Epidemiology

Step 1. Gathering Consent

The user is first informed about how the application works through attractive iconography and simple sentences. Then he has to accept the TOS to continue.

However, he has the choice to activate the functionalities using geolocation. He can change his choices later, at any time.

*We have contacted the CNIL on this subject: Our project respects the RGPDs and the laws in place, CNIL add that "volunteering" be real and not a hidden obligation.

Step 2. Traceability

He completes a form concerning his general information, to respond to the need for data traceability in a clinical epidemiology study. (This is an extension of general epidemiology in the field of clinical research. Clinical epidemiology uses the methods of statistics and general epidemiology to provide the clinician with answers to the questions that they are facing when dealing with their patient). This information entered will then be validated :

• In France by matching the patient's information after entering his AMELI number.
• Otherwise, by the doctor himself at the first medical examination.

*Users can find all the information collected about them on their account.

Step 3. AI Survey

The user answers an entry evaluation questionnaire. Simple and accessible, we have developed it, in part, from the form developed by the Institut Pasteur and the ap-hp.

All information that the user can pre-fill out to optimize the doctor's work can be added later. This survey will allow our AI to analyze the real needs and risks, to guide the user at the right time, at the right place.

B- Clinical Epidemiology

Step 4. Testing:

If there is a suspicion of COVID-19 infection by the AI, the patient can make a video appointment with a " filter " physician who will assess the probability of the assumptions, and the need for testing. Following this initial clinical examination, the person may be :

• Redirected to a testing center with a digital prescription.
• Referred to an external doctor for the management of another pathology.
• Reassured; and keep organizing his medical follow-up at home thanks to the platform so that our AI can intervene again if necessary.

Step 5. Diagnosis:

The application informs the patient of the test results. In case of a positive test, he receives at the same time the prescription of the treatment recommended by the doctor. He can follow his entire medical file from the beginning thanks to an intelligent interface.

Step 6. Emergency:

A tool, developed in collaboration with the Huddle project, to reduce the number of fatal outcomes and predict patient's need for ICU transfer in the next 24h that allows optimal allocation of critical care beds.

The majority of hospitals in France do not have the same levels of technology infrastructure. Now facing adversity of a devastating pandemic, now more than ever there is a crucial need to optimize critical care resources and ICU beds.

We are planning to manage critical resources. Two main parts of the solution: An ML model, help hospitals predict which patients are more prone to be admitted to an ICU. With the results, the team can better manage the availability and transfer of patients to different hospitals. Today, there is no overview of the availabilities of COVID - ICU beds. We want to centralize the process.

1 - Model-based Machine Learning Classifier: An ML model helps hospitals predict which patients are more prone to be admitted to an ICU. Teams can then identify which patients are deteriorating and prepare patient transfer in advance.

• Using WHO indicators backed by medical journals
• Support vectors Machin learning models were used to predict patient outcome

2 - ICU Bed Finder: A web application, helping HCPs and health authorities to find the ICU beds for patients very seamlessly. This allows us to obtain better critical care resources and time management to save more lives.

3 - Patient Health Dashboards with AI Chatbot: With the data collected, we will have useful and insightful dashboards for healthcare authorities and hospitals, helping them to make better decisions for transferring patients.

C- Clinical epidemiology: Pharmacovigilance

Step 7. Pharmacy:

The application can inform the user about pharmacies around him that have medical devices and prescribed medications.

Step 8. Compliance:

The application sends a reminder every time the medication is to be taken. It also ensures that the dosage indicated on the prescription is respected.

Before each dose, the patient must scan the box or the pill (recognized by AI) to attest that it has been taken.

Step 9. Evaluation

Every day, he is encouraged to answer a quick questionnaire on his general and functional state, developed by our research team (5 questions that we have evaluated for their usefulness, presented in an attractive format, which the patient can complete with additional data if he has time). If he needs it, he can easily order a kit thanks to our partners. Composed of standardized measuring instruments for his follow-up, we make sure of their conformity.

He can report a side effect on his patient space at any time and make an emergency appointment via the same platform.

Our AI adapts the approach of the management according to the patient's symptomatology. In a normal pathway, our system will guide the patient to a new circuit of clinical examination and screening at 2 separate intervals after treatment (J-0): J7 and J17. ( Viral load and serology measurement).

Doctors

Let's talk quickly about the doctor interface: Knowing how our colleagues use digital patient records, we know that the interface must be clear, easy to use, and, above all, flexible. This is why we have provided an NLP (Natural Language Processing) algorithm to extract accurate and meaningful data from medical reports and to apply a rigid DMP-compatible structure to them.

This service is programmed by the startup Octopus Medical, specialized in the management and visualization of patients' medical data, with whom we have established a collaboration for a rapid implementation of the platform.

Tools used :

1. NLP service: Python main code, nltk library
2. Web service: html, javascript
3. Database service: Mongo DB
4. REST APIs

Researchers

The rapid evolution of the CoViD-19 epidemic and scientific knowledge obliges public authorities, the medical community, and industrial players to acquire efficient means of collecting information in order to adapt response strategies and guide decision-making in a rapid, well-founded and relevant manner.

In our solution, data from patients and physicians, classified and stored in Healthcare-approved SaaS servers, can be linked to official clinical trial data aggregated by our algorithm and analyzed in AI.

Relevant information on the risk/benefit balance of treatment can then be deduced, leading to a procedure for a re-evaluation of the medication by the competent authorities.

Architecture:

1- Collection of Covid-19 publications Ingestion modules have been developed to ingest the different data sources. These modules developed in Python either use APIs or behave like indexing robots.

2- Data processing with data science The main objective is to facilitate the analytical work of our researchers.

3- A research and dataviz tool

+ Creation of a Task Force: Collection, Analysis, Technical and Scientific Proposals (TF RAPTS) at IHU Méditerranée Infection. It organizes the analysis work already in progress and responds to new needs.

The proposed solution will enable the TF RAPTS Task Force to respond rapidly to the numerous requests within the framework of COVID-19.

➔ It will promote a daily monitoring of publications for the medical profession and researchers at URMIT and more broadly the Ministry of Health.

➔ It will allow us to alert on trends or specificities.

➔ It will promote the synthetic feedback of health information to ministers and the government for a citizen's purpose.

D_ What we have done during the weekend

Under development during the hackathon...

• Detection and extraction of named entities (Pathologies, Molecules, Drugs, Dosages) present in the reports.
• Image extraction and analysis
• Realization of a conversational agent based on knowledge extracted from publications
• Extraction of keywords if they were not provided by the authors.
• Alert systems based on the appearance of new entities or their evolution. -Translation to facilitate access to knowledge.

Challenges we ran into

We are essentially a team of clinicians and academicians: Doctors, Researchers, Research Directors, Biologists, Md-PhD students. This is an advantage in terms of the quality of the proposed MVP, but we lack the developers to implement the POC.

It is a long term work, which requires daily and full-time attention. Unfortunately our developers are already busy with their core task, which slows down the project enormously!

We are also temporarily short of people available for communication, which prevented us, for this hackathon, to post in time the new pitch video in preparation (it will be posted soon).

That's why we want to use the prize money to recruit a FULL-TIME programmer for the project, as well as a communication officer.

Accomplishments that we're proud of

Our project is part of a very real demand, that of health experts faced with the realities on the ground, which are often not in line with mainstream fiction. We, therefore, maintain a close relationship with the IHU Méditerranée Infection, with whom we are developing the launch of this service for organizing generalized phase IV clinical trials. Prof. Raoult's IHU already has a well-established communication system, which will facilitate the deployment of the platform during the first stage (local level). Furthermore, this institute is at the heart of the problem we are currently trying to solve: certain medications, such as chloroquine, are prescribed without any clinical evidence validated by the scientific community. By starting with a POC at the local level, our work will make it possible to clarify the data in this international soap opera. Not so bad for the beginning of a scientific collaboration!

What we learned

We learned that such a big project needs full-time workers. So we'll hire a new developer and a communications officer.

E_ Like EU. : What's next for 4WCT

The solution’s impact to the crisis

Our solution will have a significant impact on the advancement of COVID-19 research.

• Patients will be able to follow a protocol adjusted to their own parameters thanks to our AI.

• Physicians will have access to a huge public anonymized database (reports of research conducted inside).

• Researchers will be able to analyze these masses of relevant data according to a rigorous scientific methodology.

---> This is unprecedented in the history of Science!

It will also allow :

• Rebuild the prestige and confidence of the IHU Mediterranean infection
• Rebuild the prestige and confidence of French governance
• Rebuild prestige and confidence in the European Union
• Rebuild prestige and confidence in WHO

The necessities in order to continue the project

We have established a calendar of our actions based precisely on the "final deliverable".

Step 1- Local Level:

We have to be flexible!

By starting on a small scale, we will avoid big mistakes in beta testing.

At the local level, these data and this research tool will enable Prof. Raoult's team from IHU Méditerranée Infection to evaluate the parameters to be taken into account in the protocol they implement. Starting the deployment on a small scale and on a flexible system also allows us to better target user demands. They'll be able to reassess and ensure the long-term safety of treatments whose approval procedure has been rushed in times of health crisis.

-->Kick-off meeting with IHU Méditeranée Infection --> Installation of the architecture in Saas mode --> Tests and learn: data ingestion, data processing, Datavisualisation --> Development of Datascience modules: ● NLP ● Machine translation ● Alert ● User testing --> Tests and learn: Data management, data processing, Datavisualisation --> Deployment on sovereign Internet hosting DIRISI Ingestion of new data and new use cases in the context of research

Step 2- National Level:

Expansion of the platform to the national level.

Our academic team is in fact already national: Ecole Polytechnique, Paris, ENS, Lyon, IHU, Aix-Marseille, ...

Step 3- European Level:

Our platform will be at the disposal of EMA. (European Medicines Agency)

We are eagerly looking for a mentor specialized in European administration, who could help us through the maze of files to be filled in for calls for projects !

Step 4- World Level:

4WCT will soon become an indispensable tool in pharmacovigilance.

This is why we also plan to fully integrate it into the WHO tools, which will become its legal owner.

The value of our solution(s) after the crisis

As we have already said, 4WCT will be very valuable after the crisis!

We see 4WCT as a novel platform for large-scale epidemiological studies, not only for COVID-19. We will adapt it for all pathologies known until now and in the future.

Collaborations

We have thus established collaborations with the aim of increasing the potential and attractiveness of the service:

• we will add an optional geolocation service by contaminated case neighborhood, based on informed consent
• we plan to provide "kitcovids" to patients requiring regular follow-up at home.

→ In the future, we think it would be interesting to include genetic data in the patient record in order to identify possible factors that would explain the manifestation of severe forms. We are evaluating the legal possibilities for cooperation with the 23andMe company. Fully integrating personalized medicine in the fight against COVID-19 would then take on its full meaning.

+ Plug-in for Research :

COVOX PROJECT (We're already working together)

Detect Covid-19 respiratory symptoms using several sampling records with your smartphone and based on citizen involvement.

In order to decrease the spread of the global Covid-19 pandemic, tests should be available for everyone, wherever they are, which is not the case today. That’s why we need an easily accessible testing solution, that can be used by everyone no matter their location.

Our solution combines both medical exams done in hospitals and medical data sent by other people. It provides a first diagnosis that should be validated by a medical staff afterward. Our project is non-invasive and can be done remotely, with nothing more than a smartphone - available to everyone. All the people contributing to the solution are protected and anonymized in our database.

HOW IT WORKS

1- Get involved

A secure and anonymous way to collaborate with the medical research. Our application is intuitive and accessible. You only need your smartphone

2- Run the tests

Use your smartphone and launch the application to start your test immediately. Our solution is based on the medical process used by hospitals to evaluate your profile.

3- Get a first diagnosis

Get a detailed report of your condition as well as advice on what to do. For instance, consult a doctor, go to the nearest hospital, etc.

4- Share the initiative on social media

Help us spread the word, and increase medical research by sharing this initiative with your friends and family.

---

we urgently need funding to hire a full-time communications officer on the project, as well as new developers.

If you are a specialist mentor in European administration, please contact us!

Built With

• html
• javascript
• json
• linkeddata
• mongodb
• natural-language-processing
• nltk
• python
• react