Our team decided to tackle a project related to healthcare technology, as we all had interests within the industry. Looking over available prompts that HackGT6's sponsors posted, we knew we wanted to work on a healthcare-related hack, as it overlapped with multiple sponsors resources and software tools that we could use. Since the current car insurance landscape was shifting more towards incentivized personal monitoring, we thought it would be a good idea to bring this concept into the healthcare realm. In essence, we wanted to incentivize healthier lifestyles while both benefiting the insurance company and the end user financially.
What it does
Our project revolves around a reward system to incentivize healthier lifestyles. We developed a rewards points system that gave end users redeemable points based on two different general actions:
1) The end user "checks-in" to a HealthyHotspot™, which is a term we coined for a location that promotes a healthy lifestyle. This includes, but is not limited to gyms and other participating sporting events. The end user "checks-in" by simply opening the Rewards App on his/her smartphone and scanning a QR-code easily accessible at his/her current location. By "checking-in", users are able to gain rewards points due to their conscientious actions promoting a healthier lifestyle.
2) The end user takes a simple image of the current snack(s) they are eating. From this image, we are able to determine if the food they are eating is healthy or not, and provide them with rewards points if they are eating healthy foods. By taking an image of their current snack choice via the Rewards App, we are able to assign them rewards points to their personal account.
The Rewards App that our team made adds value to not only Anthem's Healthcare Marketplace, but also Anthem's business strategy as a whole. For comparison, personal monitoring applications on drivers' smartphones incentivizes safer driving habits, which in turn aids the car insurance company as they are paying less overall for that driver's mistakes. In return, the driver is compensated for their safe driving by lowering his/her rates bit by bit, which will save them money in the long term. Applying these same principles to healthcare companies, we can see that there is a huge market to tap into. However, we also understand that there is a big void in personal monitoring devices when discussing healthcare because of regional privacy laws. By leveraging non-sensitive personal data, such as frequency of gym visits and snack routines, our team was able to incentivize healthier living and personalized monitoring without compromising security and without accessing useless, sensitive data such as a Social Security Number and more. With this data, not only can we deduce if someone is living a healthier lifestyle, but we can run analytics on our newly acquired data and observe trends within demographics, regions, times throughout the day, among other attributes. This be applied towards more effective targeted advertising as well as shifting current marketing techniques. Some healthcare companies are looking into investing into this new market, but it is relatively untapped.
In addition, we wanted this technology to be easily implemented, as we wanted to provide other businesses the least amount of reasons to reject our offers to put our technologies in their storefronts. Our system does not necessarily have to be directly integrated with other consumer-facing technologies. Rather, with a simple printed QR-code marking a store or event ID, we can easily acquire data from end users walking into a gym, for example. Our Rewards App requires little-to-no previous infrastructure to run effectively, making it very attractive for other businesses to invest their time (and very, very little resources) into.
To achieve the simple and user-friendly environment that the Rewards App provided, we used Google Cloud Services to automate necessary components of our technology. To parse the store or event ID, which is in the form of a QR-code, we used Google Cloud Vision API to extract the important information from the image. From there, we referenced the store/event ID into Google Firebase to define the number of rewards points earned from visiting that store/event. For the food identification segment of the project, we used Google Firebase Machine Learning Vision API to label the food image that the end user took, and classified whether it was a healthy food or not. If it was indeed a healthy food, we added several rewards points to the user's account. All of this was implemented within an Android Application using Android Studio.
The biggest challenge of this project was enabling the Google Firebase Machine Learning Vision API as it required a new Gradle version and support for newer AndroidX libraries, which we were not familiar with. This meant refactoring most of our code base within the last 12 hours of the hackathon. As a group, we learned plentiful about Android application development as well as using not-so-well documented machine learning frameworks. With this experience now under our belts, we will be better equipped to pursue another related machine learning project in the future.
In the future, we envision this technology to be available to all Anthem customers and to be used by many. This product helps both Anthem as well as its customers by saving each money. Since we are relying on cloud-based services, there are very few hardware restrictions for our technology's successful usage, thus making it accessible to a wider audience. Plus -- who wouldn't want to get paid for going to the gym!