How we were inspired?
The initial inspiration came from the PhD of our colleague and an idea incubation at the Bauhaus University. We noticed that transfer of scientific output still lacks behind the technical possibilities of translating written text into more inclusive visuals, icons, audio formats or other modes of digesting and transmitting information. During the Easterhackathon2020 of the CharitéBerlin and HackingHealth we found out that General Practitioners (GHPs) are confronted with an increased amount of questions from their regular patients especially due to Covid19, while they fight the virus at the front line and train and inform the staff members. Since research studies are updating constantly and the GP lacks time more than anything else, we wondered how could we help the GP stay informed and answer questions from staff and patients (+their relatives) more efficiently.
The Solution we bring to the table?
- a mobile app which puts “10,000 (or more) COVID19 research papers into the General Practitioners’ pockets” providing informed answers in an easily understandable format such as printable graphic summaries (also for health care workers) and short videos (for patients).
- Back end: Intelligent processing of complicated research text processed on cloud infrastructure such as AWS SageMaker using data ingested either from a search engine database of curated literature, from APIs, or mixture of the two with curated content and presented in a user friendly format. Delivered in Python and proposed libraries include PyTorch and OpenNMT project.
What is our mission?
We want to put 10.000 research papers and/or articles directly in the pockets of General Practitioners (GPs), the public and hospital workers, allowing them to get their relevant questions answered in the format of short video snippets and visual graphics – all done with a hybrid of curated information and Machine Learning.
What have we done during the weekend?
- Our proof of concept: having interviewed the people on the front-line and focusing on General Practitioners for a fully functioning MVP.
- We further have a code to summarize relevant abstracts and a database of around 30 papers and the first video snippet as well as goals and vision set plus a front end look and feel for later, check out our preliminary website.
- Collected proposed solution such as a literature search engine solution and sought out more collaborative solutions such as with APIs.
- Conducted a security assessment and implementation plan.
The solution’s impact to the crisis?
We help GPs save time and communicate more efficiently while staying up to date with the current research. Patients can disseminate the videos with friends and relatives helping fight fake news and rumors. Staff members and doctors can put research outcomes into context and have a quick overview of the methodology and limitations to create their own opinion on how to apply the outcome into their specific context.
The necessities in order to continue the project?
- Further Front-end development to (UX and Graphic Designers to be paid) plus time and effort to work directly with our GPs/our first user base
- Funding & Partnerships with research databases
- Developing a code to do text analysis based on previous Kaggle challenges such as the following: https://covidex.ai/
- Further work together with researchers, partners and data engineers.
There is no on premise solution. We need:
- AWS or Google Cloud credits in order to use the cloud infrastructure and any associated services available on these companies to process data accordingly.
- API and Python development for further enhancing the existing Gitlab repository OR database search engine development as proposed by an interested development team.
Development and Infrastructure Pipeline:
- Developing an Elastic/Lucene/etc based search engine
- Ingesting medical literature and COVID-19 research APIs
- Further development AWS SageMaker algorithm to ingest metadata
The value of our solution after the crisis?
There is a huge amount of research out there about many topics that will become relevant for the post-coronavirus crisis management. The restoration and innovation of healthcare systems, the preparation for future pandemics and more. A more direct transfer between researchers and practitioners will be necessary in various fields for the renegotiation of many fields in the society. We envision to build next to TheHealthManager App, TheInnovationManager App as well as ThePublicManager App.
Built With
- amazon-ec2
- amazon-web-services
- balsamiq
- ci/cd
- clideo
- css3
- docker
- ecr
- figma
- firma
- gitlab
- javascript
- jekyll
- love
- npm
- opennmt
- python
- pytorch
- s3
- sagemaker
- scss
- stackbit
- unibit
Log in or sign up for Devpost to join the conversation.