ž[Object object]

Problem: More and more people are diagnosed with heart diseases. In the last year around 1.7 million people died within the EU due to CVDs’, which accounts for 37% of all deaths in that period. At the same time we see a rise of unhealthy behavioral changes, such as lessening physical activity among the population, increasing chances of succumbing to heart disease. This often leads to overwhelming medical care as well, as they find it increasingly difficult to provide proper care to those in need.

Solution: We wanted to offer NeuraHeart as a possible solution to the described problem. The application offers a way to organize and view patient data, both in text form and using graphs. Our application accepts data from databases as well as home device readings, such as Apple Watch. The app is also capable of making a prediction of a patient's state and bringing it to the attention of medical staff. in case of worrying results.

Impact: The solution has the potential to save many lives. From our research, often the most deadly thing about CVDs’ is a late diagnosis. Our app can help avoid such situations, while simultaneously adding a layer of organization to medical data. The app can also help increase the size of medical databases and constantly improve prediction models for different diseases. As a proof of concept, delivered data the app was able to detect a strong correlation between sudden weight gain and heart failure.

Feasibility & financials: Our research shows that users are greatly engaged with the AI creation process and are happy to provide data. Furthermore, a great number of households already have a device capable of collecting medical data. As the number of users is of the essence, charging them directly is mostly out of the question and the app would probably have to be sustained through contracts with government and private companies. The base of the app would require around 12k € and take another 4 months to complete.

What is new about your solution?: As of now there are very few publicly available solutions on the market. While some research on the topic has already been done, the territory still seems to be mostly undiscovered. Through consultation with some physicians, they seem to have no access to such tools despite the work being already started. The novelty of the solution as well as focus on availability to the public are our main innovative points.

What you have built at the hackathon - text explanation + code (e.g. GitHub link): During the hackathon we created a functional demo of an application, capable of proper display of provided data in the form of text or graph, as well as some ways of processing it. The app can accept data provided from outside and process it as well. We have also created a classifier for structural heart diseases connected to the app, which has achieved over 70% accuracy based on the data provided. https://github.com/HakierGrzonzo/healthHack

What you had before the hackathon, please mention open source as well: We had access to various frameworks and tooling. All of them are contained within our project, which you can view on github.

What comes next and what you wish to achieve: We would be happy to further develop the idea. However at this point in time it may be too risky of an endeavor to pursue, especially when all of us work a part time or full time programmer positions. Michal Danikiewicz is also the only person who will work with medical technologies in the near future. As such we can’t fully commit to the idea without proper support.

Video: https://www.youtube.com/embed/piqZaUrG1nY