Every 40 seconds, someone in the United States has a stroke. Every 3.5 minutes, someone dies of stroke. Every year, more than 795,000 people in the US have a stroke. Heart stroke is also the leading cause of death and disability in India. Early detection and prevention of heart stroke is crucial in order to improve patient outcomes and quality of life.
Traditional methods for identifying individuals at risk of heart stroke, such as blood tests and imaging studies, can be expensive, invasive, and often require specialized equipment and trained personnel. As a result, there is a need for more accessible and cost-effective methods for identifying individuals at risk of heart stroke.
Fitness bands, which are wearable devices that track physical activity and other health metrics, have become increasingly popular in recent years. These devices generate a wealth of longitudinal data, including information about physical activity, heart rate, and sleep patterns. This data has the potential to provide valuable insights into an individual's overall health and risk of developing various diseases, including heart stroke.
In this proposed solution, we are detecting and alerting users of the risk of stroke using wearable technology, machine learning, and mobile application development. We can use any fitness tracker to collect longitudinal data of heart rate and other vitals in real-time, which is then used to train a machine learning model to predict the chances of a person having a stroke based on symptoms such as age, gender, body mass index, work type, and continuous heart rate.
We aim to provide early detection and alerts for users who may be at risk of having a stroke. By providing timely alerts, users can seek emergency treatment quickly, which increases the chances of survival. The application will not only predict the likeliness of a heart stroke, but will also share users location to their contacts and nearby hospitals during the time of an actual heart stroke.
The major challenge that we faced while building Healthify was the 24-hours time limit. With the limited time in our hands, we had to prioritize key features and add the remaining features to our future scope. But the time limit was what made the hackathon even exciting.
Keeping the time limit aside, one major challenge we ran into was to ensure the accuracy of the prediction algorithm. We had to collect and analyze a large amount of medical data to build a reliable algorithm. We performed an intensive search for public datasets that fit our requirements but didn't find one. We then had to create our own custom dataset. My Mi band 4 data from the past two years was really helpful in generating custom data.
Another challenge we faced was to ensure the security and privacy of the user's medical data. We had to implement various security measures to protect the user's data from unauthorized access and ensure that the data is encrypted. We used a package called Crypto in Flutter to implement advanced encyption algorithms to securely store user's health vitals.
To overcome these challenges, we conducted intensive research to implement industry-standard security protocols in our application. The mentors from schneider electric gave amazing insights that have also helped us to overcome challenges in our application.