Skip to Content
Author's profile photo Jonas Wang

Predictive Analytics in the Health and Life Insurance Industry

Several months ago, we built a prototype that integrates a Microsoft band with HCP and shared what we did on this blog post. And this time we are back with more new exciting developments to share. (WOOHOO!)

As the popularity of wearable technology continues to grow, it also creates massive amounts of health and fitness data. Many companies have expressed their interest in this data. Apple released their own platform, Apple Health, that collects various data from fitness apps and wearable companion apps on your iPhone and attempting to put the data in one place. Google Fit is Google’s answer to Apple’s Health app. It uses the sensors built into your device to automatically track activities like walking, biking and running. All of this data is meaningless unless it provides insight regarding your health. Predictive analytics is the way to go. Insurance companies have started looking into ways on how to utilize this data. One of the ideas is to analyze patterns in past events and model future outcomes or behaviors. SAP has existing products that provide predictive analytics. Therefore, we continue to improve our prototype by integrating these products.

Predictive Analysis Library (PAL)

One of the ways HANA increases performance is by executing complex calculations, like predictive computations, in the database instead of at the application server. SAP has grouped functions for particular topics together into the Application Function Library (AFL). The predictive functions have been grouped together in the Predictive Analysis Library (PAL) which is part of the AFL.

We used the forecast smoothing algorithm from PAL to generate predicted daily average heart rate with historical data. We then created a new chart on our UI5 application to compare the predicted data with actual data.

Screen Shot 2016-08-15 at 9.33.23 AM.png

Sunburn Calculator

The Microsoft Band features a UV (ultraviolet) sensor that can identify the UV index level of its surrounding environment. Similar to other sensor data, the iOS app collects the UV index level from the Microsoft Band and sends them to HCP. Then the UI5 application visualizes the UV index level in real-time.

Screen Shot 2016-08-16 at 10.43.48 AM.pngScreen Shot 2016-08-16 at 10.31.39 AM.pngScreen Shot 2016-08-16 at 10.32.24 AM.png

Having the data on our hands is not enough unless we can analyze them. So we added a “Daily UV Exposure Summary” page that displays UV exposure time on different UV index levels. It also calculates the UV exposure time limit based on the user’s Fitzpatrick skin type. When the user reaches his/her UV exposure time limit, the UI color will change to warning red and a new button called “Send Warning” will appear on the page automatically. The insurance companies can potentially use this button to send a warning to the end user.

Screen Shot 2016-08-09 at 1.22.10 PM.png

Screen Shot 2016-08-09 at 1.30.47 PM.png

Financial Service Claims Management (FS-CM)

Once the end user receives the warning on the mobile application. The end user can use the warning information (UV exposure time) to schedule an appointment with a dermatologist and also send a claim to FS-CM, which is a standard SAP product that can handle claims for insurance companies.


Integration with Business Object Cloud (BOC)

SAP BusinessObjects Cloud (formerly SAP Cloud for Analytics) is a new generation of SaaS that redefines analytics by providing all analytics capabilities (BI, Planning, Predictive…) for all users in one product. It is built natively on SAP HANA Cloud Platform to simplify access to a public cloud experience that customers trust.

We followed this tutorial and successfully shared the UV data from HCP to BOC. The visualization tools provided by BOC were by far the best we had seen. And we were able to make various charts and a geo map on BOC.

For example, we made a trend chart that visualizes one user’s weekly UV exposure time. As you can see, the user had more UV exposure time during the weekend. This could possibly imply that the user stayed outside more often during the weekend.

Screen Shot 2016-08-15 at 3.43.37 PM.png

We also pulled the weather data from and made a comparison bar chart to explore the relationship between weather and UV exposure time.

Screen Shot 2016-08-15 at 4.12.29 PM.png

The geo map is the most cool-looking one. We selected a sample of ten people from two different locations (Newtown Square and Palo Alto) and compared their high UV exposure time in January.

geo map.png

New Architecture:

Overall, we have made a few improvements: heart rate prediction with PAL, FS-CM integration and analyzing data with BOC.

Screen Shot 2016-08-18 at 11.23.34 PM.png

Looking Ahead

Due to time limit, we could only implement these new ideas so far. We would like continue to improve our prototype with more new ideas and hopefully turn it into an official SAP product. One of the ideas we are looking forward to is SAP Ariba Network integration. This can create the possibility of letting end users select a dermatologist they prefer or even order sport gears from SAP Ariba Network.

Assigned Tags

      Be the first to leave a comment
      You must be Logged on to comment or reply to a post.