1、MILLIMAN REPORT The role of wearables in private medical insurance March 2020 Joanne Buckle, FIA, FIAI Tanya Hayward, FIA Natasha Singhal Kishan Desai MILLIMAN REPORT Table of Contents 1. EXECUTIVE SUMMARY . 1 2. WHY ARE INSURERS THINKING ABOUT IMPLEMENTING WEARABLES? . 2 2.1 RISKS TO STAKEHOLDERS .

2、 3 2.2 INSURANCE FIRMS ALREADY USING WEARABLES . 4 3. WHAT DO “CONSUMERS” THINK? OUR SURVEY RESULTS . 5 3.1 SURVEY OBJECTIVES . 5 3.2 DEMOGRAPHICS . 5 3.3 DEVICES, TRACKING ACTIVITIES AND STEPS . 6 3.4 STEP COUNT . 8 3.5 SHARING PERSONAL DATA WITH INSURERS . 8 3.6 INFLUENCE OF WEARABLES ON ACTIVITY

3、AND HEALTH . 9 3.7 FEEDBACK FROM USERS . 11 4. HOW RELIABLE IS REAL TIME DATA? . 12 4.1 FRAUD AND ABUSE . 12 4.2 FUNDING CONSIDERATIONS . 12 4.3 ACCURACY OF RESULTS . 13 4.4 CONSIDERATIONS FOR USE OF WEARABLES DATA . 14 5. WEARABLES AND HEALTH RISK ASSESSMENTS (HRAS) . 15 5.1 HEALTH RISK ASSSEMENTS

4、(HRA) . 15 5.2 HEALTH RISK ASSESMENTS AND WEARABLES . 17 6. CONCLUSION . 20 HOW MILLIMAN CAN HELP. 21 MILLIMAN REPORT The role of wearables in private medical insurance 1 March 2020 1. Executive Summary Mass market wearable technology is relatively new, and it has evolved dramatically in recent year

5、s. Ever since the rollout of electronic devices capable of measuring and recording various types of health data, private medical insurers have been carefully eyeing the potential usage of these devices. In this paper we discuss the implementation and practical uses of wearables in the private medica

6、l insurance market. We review why insurers are considering using wearable devices and how these devices might benefit insurers business models, as well as some of the pitfalls to consider. Wearables provide real-time data points that we expect to be detailed and accurate. However, is this really the

7、 case? We look at the reliability of real-time data, potential areas of fraud and abuse and whether wearables data can really be considered reliable. We also look at funding considerations when incorporating wearables into an insurance product, alongside other key considerations for the use of weara

8、bles data. We found that, whilst wearables data can help insurers gain additional insight into the general fitness levels of its policyholders, the additional data collected might not necessarily improve upon existing claims cost prediction techniques. As part of our research, we conducted a market

9、survey designed to understand consumer opinions on their interaction levels with wearable devices and their thoughts on the use of wearables in insurance. We share the findings of our market research and discuss some of the key conclusions. An interesting outcome from this research was that even tho

10、ugh a significant proportion of our respondents regularly tracked their health data, and worked within the insurance industry, views on the role of wearables data within insurance varied quite widely. We consider the range of devices used by survey participants, the type and frequency of activity ca

11、ptured and their views on the use of this information in determining the premium level for their insurance policies. Finally, we consider whether the types of data available from wearables really provide predictive value in healthcare. As most health risk assessments (HRAs) are generally accepted to

12、 be evidence-based in their scoring, we conducted a sensitivity analysis on HRA data elements that are parallel to the types of data that could be obtained through wearables (e.g. tracking exercise and activity levels). We varied the different input metrics in three independent HRAs to see which met

13、rics had better predictive value for the overall outcome. We use these findings to assess whether corresponding measurable data provided from wearable devices are likely to enhance the pricing methodologies of private medical insurers significantly. Our findings indicate that, although wearables may

14、 encourage members to increase their activity levels, many of the key factors that influence HRA scores (and hence members overall health levels) are not captured by wearables. MILLIMAN REPORT The role of wearables in private medical insurance 2 March 2020 2. Why are insurers thinking about implemen

15、ting wearables? The use of wearables in insurance is typically centred around three main objectives: 1. Improve claims cost prediction a. Insurers can use wearables data to supplement their underwriting processes and pricing models. 2. Make people healthier and reduce healthcare claims costs a. Incr

16、eased awareness about healthy lifestyle behaviours and increased physical activity is expected to improve members health and eventually reduce overall healthcare claims costs. 3. Strengthen competitive position a. As the popularity of wearables increases, insurers may be required to offer them as pa

17、rt of their regular wellness offerings to remain competitive. Although these objectives sound reasonable, it is important to question whether wearables do in fact add power to claims cost prediction beyond what traditional and other big data measures contribute, and if wearables are in fact capable

18、of making people healthier. Offering wearables to insurance policyholders certainly presents new opportunities not previously available to insurers. However, these opportunities are not without their pitfalls, as shown in Figure 1. FIGURE 1: OPPORTUNITIES AND PITFALLS OF USING DATA FROM WEARABLES Ca

19、pability Opportunities Pitfalls Collect real-time data Large amounts of real-time data to track activity and health indicators of individuals Storing, processing and creating business value can be tricky. Risk of collecting incorrect or misleading data. Regulatory and data protection concerns. Prici

20、ng rating factors New rating factors not otherwise available. Enhance predictive power. Complex black box pricing models. May not provide additional insights compared to traditional methods because the underlying science is still not clear (i.e., we dont know which factors are most predictive of mor

21、bidity with any real level of precision). Potentially expensive to incorporate. Underwriting using additional data New underwriting criteria not otherwise available e.g. having credible and reliable wearables data may be a useful source of information when classifying policyholders as a standard or

22、substandard risk. Potential for fraudulent methods used to achieve high activity levels. Discrepancy between devices may produce different conclusions for different measures. Lack of evidence that high levels of activity recorded by wearable devices can be associated with better risks. MILLIMAN REPO

23、RT The role of wearables in private medical insurance 3 March 2020 The expectation is that insurers can positively impact their members lifestyle behaviours with wearables, which should result in improved health and lower healthcare claims costs. Figure 2 illustrates a potential pathway of a members

24、 journey along a positive lifestyle behavioural change journey. FIGURE 2: THE IDEAL OUTCOME FROM IMPLEMENTING WEARABLES ON POLICYHOLDERS However, actual experience could turn out quite differently from the expectation. A study by Rand Health on US wellness programmes in 20131 found that, while it is

25、 possible that the use of wearables can improve health and claims experience, further considerations are required to incentivise members. Two major findings from the study included: 1. Take-up rates for wellness tests are less than half for eligible employees and that less than a fifth of employees

26、provided with follow-up actions carry them out. 2. When wearable technology is used in conjunction with loss-framed incentives, there are lower activity levels and higher financial costs to members. Conversely, gain-framed incentives can lead to higher activity and greater discounts for members with

27、 no additional financial costs incurred. It is important to recognise that the use of wearable technologies alone is unlikely to be sufficient to drive real change in lifestyle behaviours and impact members health. A comprehensive wellness programme that focuses on additional health-related activiti

28、es with appropriate incentive structures is more likely to achieve the desired effects. In a separate Milliman publication, we discuss the considerations for implementing and evaluating wellness programmes.2 2.1 RISKS TO STAKEHOLDERS For wearables to be integrated into a viable insurance product, ri

29、sks faced by key stakeholders need to be addressed to achieve stakeholder alignment. As shown in Figure 3, for insurers to achieve alignment with the relevant stakeholders, they will need to consider their own objectives, as well as those of the stakeholders and regulators. They will also need to co

30、nsider the needs of their customers and identify their operational and technological capabilities to ensure that they can create a viable insurance product. Figure 3 is taken from the Institute and Faculty of Actuaries (IFoA) presentation Wearables and the Internet of Things.3 1 Mattle. S. et al. (2

31、013). Workplace Wellness Programs Study. RAND Health. Retrieved 3 March 2020 from https:/www.rand.org/content/dam/rand/pubs/research_reports/RR200/RR254/RAND_RR254.pdf. 2 Beveja, L. et al. (December 2019). How to Implement a Wellness Programme. Milliman Report. Retrieved 3 March 2020 from 3 IFoA Wea

32、rables and Internet of Things Working Party (18 June 2018). Wearables and the Internet of Things: Working Party Update. Retrieved 3 March 2020 from https:/www.actuaries.org.uk/practice-areas/health-and-care/disbanded-research-working-parties/impact-wearables-and-internet-things. MILLIMAN REPORT The

33、role of wearables in private medical insurance 4 March 2020 FIGURE 3: STAKEHOLDER ALIGNMENT FOR VIABLE INSURANCE PRODUCTS4 2.2 INSURANCE FIRMS ALREADY USING WEARABLES Many insurers are already making use of the technology in their insurance and wellness programme offerings. Figure 4 includes example

34、s of how some insurers are using the technology to incentivise policyholders. FIGURE 4: EXAMPLES OF CURRENT USE OF WEARABLES IN INSURANCE Aditya Birla Health Discounts for policyholders who record a specified number of steps using an activity tracker or attend gym sessions or have a health assessmen

35、t. The Vitality Programme Vitality members earn points and achieve a higher Vitality status when they undertake activities that are assumed to impact on health status. Higher Vitality statuses unlock higher rewards for benefits such as gym, travel and other discounts. AXA Offers a free Withings Puls

36、e fitness tracker. Participants receive discounts of over $100 on their insurance policies, as well as discounts off any Withings product purchases when they complete a certain number of steps. Oscar Rewards customers who track their fitness data gift cards when they reach their step goals. United H

37、ealthcare Rewards users with healthcare credits for reaching daily fitness goals. Qantas Assure Policyholders receive Qantas frequent flyer points if they lead more active lifestyles. Aetna Monitors daily activity and provides assistance in achieving personalised health goals. The app also provides

38、recommendations, nudges and rewards. Esurance SavorBand devices are offered which can capture information on food consumed, including recipes, cooking tips, and purchasing discounts along with other data. Beam Technologies Uses Bluetooth-enabled toothbrushes to reward good brushing habits with disco

39、unted insurance premiums and other rewards. 4 Ibid. MILLIMAN REPORT The role of wearables in private medical insurance 5 March 2020 3. What do consumers think? Our survey results 3.1 SURVEY OBJECTIVES We conducted a survey, shared on LinkedIn, to investigate what our contacts think about wearables i

40、n insurance in the context of their role as consumers by asking questions relating to: Demographic profile Current use of wearables Opinions on sharing wearables data with insurers Opinions on fairness of using wearables data for premium calculations Opinions on the use of discounts for using wearab

41、les on insurance policy renewal We gathered 70 responses from this survey. Highlights from our survey results are discussed in the sections that follow. 3.2 DEMOGRAPHICS We asked our respondents about their age bands, genders and job types. That information is presented in Figure 5. FIGURE 5: POSSIBLE RESPONSES TO DEMOGRAPHIC QUESTIONS AGE BAND GENDER JOB TYPE 18 18 24 25 34 35 44 45 54 55 64 65+ Female Male Prefer not to answer Other (please specify) Actuarial Data science Clinical practice Strategy Finance Other (please specify) Figure 6, Figure 7 and Figure 8 summaris