(Image source: Munich Re.)
In the life insurance industry, wearable devices have largely been a means of achieving greater policyholder engagement—much like the role Internet of Things (IoT) applications have played to date in personal lines property/casualty insurance. It’s no small thing for life insurers to find a way of engaging with far greater frequency with their policyholders, and it’s a breakthrough for them to do so in a life-affirming way, associating the relationship with their own, shall we say, vitality, rather than their mortality. Wearables also—again, like IoT devices—encourage behavior modification that is likely to reduce risk. Homeowners insurers have recognized this behavioral dynamic for decades, giving premium breaks for policyholders with approved security systems. The next level is to be able to use data from IoT devices as reliable underwriting criteria, and a new study by Munich Re (Munich/Princeton, N.J.) does just that, using wearable device to effectively segment mortality risk.
Munich Re conducted the study using a U.S. population-based dataset provided by Vivametrica (Calgary, Alberta), a health analytics company that provides measurement of mortality and chronic disease risk using digital biomarkers developed from personal sensor data. Munich Re performed classical actuarial analysis on the dataset and applied survival analysis and machine learning techniques to gauge predictive correlations between physical activity and mortality.
Effective Means of Segmenting Mortality Risk
The wearable sensors that supplied the data measured the level of physical activity through step count and minutes of moderate to vigorous activity per day. Munich Re’s analysis found that steps per day can serve as an effective means of segmenting mortality risk even after controlling for age, gender, smoking status, and various health indicators.
Reliable wearable sensor data opens up an avenue for effectively underwrite life insurance without the intrusive process of obtaining bodily fluids, according to Patrick J. Sullivan, Senior VP, Integrated Analytics, Munich American Reassurance Company. “With the advent of wearable technology and smartphones that record physical activity we gain an important new source of data,” Sullivan comments. “The challenge is then, how do we quantify the relationship between activity and insurance risk so carriers can feel comfortable eliminating fluid testing and start to price for healthy behaviors. At Munich we’ve put a team of actuaries and data scientists on this because we see so much promise for getting more people the right coverage with a simpler process.”
The idea that a healthy life style is likely to prolong life is an obvious one; the problem is being able to accurately measure a given individual’s physical activity. Munich Re accepted the intuitive premise but faced a two fold challenge, according to Michael Taht, Executive VP, Research, Analytics and Underwriting, Munich American Reassurance Company: how best to define activity and then where to find data to support this hypothesis.
“The research that we have done with Vivametrica allows us to concretely confirm our hypothesis,” Taht affirms. “The research findings can now be incorporated into both the initial life insurance underwriting process and the ongoing relationship between the life insurer and the life insurance client. As a reinsurer, we are able to use our expertise to make this a reality.”
Remaining Hurdles to Adoption
Since the advent of wearable devices, insurers have speculated about the potential use of the “instrumented individual” to enable “continuous underwriting,” according to Tom Benton, VP of Research and Consulting, Novarica (Boston). Munich Re’s study signals a beginning for an increasingly accurate understanding of wearable devices’ use for life insurance underwriting, but hurdles remain.
Benton suggests that the most immediate issue is one of accuracy of data as measured by devices—many of which have notoriously been poor at measuring heart rate, for example. Another is persistence: adoption of wearables has been significant; people will put them on, but they may not keep them on. “Someone might get a device as a gift and wear it for a while, but then find it uncomfortable or not particularly useful,” Benton says. “There is a well-documented abandonment rate.
Incentives from a user’s life insurance carrier is likely to mitigate abandonment, but accuracy-related concerns may still impede the progress of insurers’ use of wearable-derived data, according to Benton. “Any problems with accuracy can lead to challenges in court,” he notes. “Also, regulators will always be wary about anything relating to data privacy, so it remains to be seen whether they will approve underwriting based on wearable device data.”