More than an asset, technology is an indispensable ally of the insurance industry.
Technology in all its forms, from tools that streamline operations to innovations that offer people a reprieve from undergoing a series of operations. Technology that saves lives without sacrificing limbs. Technology that spares insurers the cost of irreversible procedures. Technology that spares people the price of a lifetime of little or no mobility.
I refer, specifically, to technology that helps diabetics who suffer from foot ulcerations.
As a scientist, who also happens to be a diabetic, I refer to an epidemic we can prevent: an epidemic we must prevent, because the technology is available, the intelligence accessible, the results attainable; an epidemic we cannot dismiss unless we shut our eyes, unless we cannot see—unless what threatens us also blinds us—because it is otherwise impossible not to notice the cripples and amputees among us.
The good news is that we can reduce these risks, thanks to technology of the sort like MOTUS Smart Powered by Sensoria: an Optima Molliter boot, with a dedicated patient mobile app and Microsoft Azure cloud technologies. The boot has interchangeable, different density insoles to relieve pressure from the area of ulceration, thereby improving blood circulation and clinical outcomes.
Consider this breakthrough a giant step for not only diabetics but insurers, too.
Consider this breakthrough an even bigger leap for all mankind, as the alternative is neither fiscally sustainable nor morally sound. Not when the global cost of diabetes exceeds $1.3 trillion. Not when a surgical saw can no more cut costs than it can be something different than it is. Not when insurers can cut costs by covering what covers—literally—patients’ feet.
If insurers champion smart technology, the benefits will be universal.
From making health insurance more affordable to giving patients the chance to walk in a way they can ill afford to ignore, from bettering the reputation of insurers to changing people’s lives for the better, technology is invaluable.
Consider, then, diabetics as the most visible beneficiaries of new technology and renewed support from insurers.
Consider these advances for what they are, the product of extensive research and development.
Consider, also, what it means for insurers to earn the trust of clients: to keep and strengthen this trust by acts of prudence and policies that act to inspire doctors, scientists and entrepreneurs throughout the world.
Take these things into consideration, to be sure.
More importantly, take these things as an invitation to lead. Take the time to do these things well, so we may all move forward together.
The World Health Organization has published a five-year strategic plan focused on 10 major threats to global health in 2019, among which there are noncommunicable diseases, such as diabetes, cancer and heart disease – collectively responsible for over 70% of all deaths worldwide. Through the plan, the WHO wants to try to ensure that 1 billion more people will benefit from access to universal health coverage, 1 billion more people are protected from health emergencies and 1 billion more people enjoy better health and well-being. It’s a very optimistic goal, and different types of solutions are required; all means available should be employed.
Where do insurers come in? They are part of the healthcare system and could do more to contribute to global health. Though it may seem perhaps altruistic to associate insurers with such a noble scope, the reality is that they, insurers, would also benefit.
Forward-looking companies have understood this opportunity, so there are examples of insurers that have started to use the “Insurer as Partner” approach, which implies an active role in prevention rather than just being reactive and paying claims when an undesirable event occurs. This new potential role of the insurer has been made possible, in great part, by what we now call “connected insurance,” which encompasses IoT (Internet of Things), wearables and other monitoring devices. The re-shaping of the insurance industry has already begun, and it will continue based on new technologies.
Connected health insurance can become profitable for insurers as it allows measurement of the risk for a specific client and thus allows the presentation of an improved, better-priced value proposition that may also improve general health. The insurance company can’t possibly make it on its own and will have to seek partners from both the technological innovation sphere and medical providers, keeping in mind that its role in the health system is changing from “payer” to “pivot.”
Discovery’s Vitality Program
To better grasp the actual benefits for clients and not just for insurers that adopt such an innovative approach, let’s look at the South African insurance player Discovery, which can be considered the benchmark when it comes to engaging members and improving their quality of life. Its Vitality program has created a system that not only raises the loyalty of customers but improves their lifestyle and overall health.
Discovery’s Vitality uses an “early warning” mechanism that can anticipate serious health problems and more expensive claims. It does so by using connected devices like the smartwatch. According to Discovery, Vitality Gold status members with heart disease have 41% lower risk claims than members with no Vitality membership. Vitality Gold members living with diabetes have 50% lower risk claims.
Another claim coming from a presentation by Discovery Vitality at DIA Amsterdam 2018 deserves our attention: There is an 18% reduction of hospital and chronic claim costs for the batch of Vitality members who use the Vitality Active Rewards (VAR) alongside the Apple Watch, compared with the group of insured who do not use an Apple Watch. VAR is a smartphone application based on fitness points, which is designed to encourage Vitality members to increase their activity by setting weekly personalized physical activity goals – and then rewarding users for achieving them. (Discovery specifies that its data is based on a cross-sectional view of the relative claims experience, and it is premature to show the improvement over time given the lower frequency of health claim events.)
Discovery says that Apple Watch owners enrolled in the program are 35% more active than prior to getting the watch. Since the VAR system was launched, there has been a 24% increase in physical-activity days and a 9% increase in meeting higher exercise targets. The data is telling, and the implications for ensuring healthy lives and promoting wellbeing are significant. Seen from this point of view, the transition to a “prevention-centered” approach is a pragmatic decision for insurers because, in time, the portfolio tends to change its structure, passing from a majority of “sick” clients to a majority of relatively “in good health” clients.
ICS Maugeri’s Mosaic case study
Let’s look now at a more specific issue within the spectrum of uninsurable diseases, that is diabetes. Diabetes is a chronic disease that occurs either when the pancreas does not produce enough insulin or when the body cannot effectively use the insulin it produces. In 2014, 8,5% of adults aged 18 years and older had diabetes. In 2016, diabetes was the direct cause of 1,6 million deaths, and in 2012 high blood glucose was the cause of another 2,2 million deaths. It is estimated that the number of diabetic patients worldwide will be 629 million by 2045.
Diabetic patients have a high risk to develop severe complications generated by the evolution of the pathology, such as peripheral neuropathy, retinopathy, nephropathy and cardiovascular diseases. It is well-known that insurers either do not cover diabetic patients or, if they do, require a high premium. This is due to the difficulty to measure the probability of the occurrence of risks associated with these clients. Therefore, the insurance sector is leaving uncovered a market that is becoming more and more relevant.
ICS Maugeri, a major group of hospitals specializing in rehabilitation medicine, has developed, in partnership with the University of Pavia, an instrument called Mosaic aimed at improving the clinical management of patients affected by diabetes mellitus type 2 (T2DM) that can calculate the risk of developing complications in different time scenarios. Mosaic uses AI and machine learning that is based on algorithms able to learn patterns and decision rules from data. Based on the results expressed in one of their published research papers, the team has been able “to predict the onset of complications (retinopathy, neuropathy, nephropathy) at different time scenarios: at three, five and seven years from the first visit. The final models are thus able to provide up to 84% accuracy in predicting the probability for a diabetic to develop the three main complications and are easy to apply in clinical practice.
In insurers’ terms, this shows that risk associated with diabetes can be estimated; furthermore, given that clinical evidences show that a proper management of the diabetic patient (intensive pharmacological treatment etc.) can lead to a significant reduction in the possibility of developing complications, the risk itself can be managed and reduced. The question is: How can insurers make sure that diabetic patients follow the required therapeutic path? It’s a difficult job. Diabetic patients are required to follow a rigorous clinical, diagnostic and therapeutic path to manage and control their pathology and try to limit or slow the consequences of this chronic disease. This path involves periodic medical checks and diagnostic tests as well as continuous and intensive drug therapies, requiring significant effort for patients and their caregivers. Most of the time, the scheduling of such periodic checks must be autonomously managed by the patient, resulting in a progressive reduction of adherence to the required clinical paths.
Within the Mosaic project, the patient is monitored with the help of wearables and telemedicine; this allows the team to (i) personalize and update pharmacological treatments, (ii) identify and update the diagnostic path to be performed to monitor and reduce the risk of complications and (iii) identify on-time criticalities that may require timely investigations. Therefore, this approach allows a significant risk control and, potentially, risk reduction, allowing the insurer to update the premium yearly.
How do Discovery Vitality and Mosaic fit in together?
Discovery Vitality uses gamification, reward systems and tracking devices to steer clients toward a healthier life style. Imagine if Vitality would be integrated with Mosaic’s technology for diabetes patients. This would mean that suddenly diabetics would become insurable, and the client base would increase.
Vitality has already proven that a reward-based system can help improve behavior, so probably it would also work as an additional incentive for diabetics in keeping them effectively engaged with their prescribed treatment. Taking for granted that diabetics will follow a program step by step and change their behavior toward a desired goal is not something anyone should do. Even if the real stake for diabetics is their own life expectancy, which should be motivation enough, the reward element could be a good and fun extra incentive for reaching health goals.
As estimated costs with lifestyle-related conditions (including diabetes) will be 47 trillion by 2030, insurers, the healthcare systems, clinical providers and patients could all benefit in some way from such a program. We would like to see this implemented in the short term at a larger scale than the test made by Mosaic, and it would also be interesting to look at how this approach could be extended to other chronic diseases.
By contrast, unlike the more transparently dishonest wellness industry, the diabetes industry’s “outcomes” can only be challenged, rather than simply invalidated on their face. And no way I’m offering a reward. (I’ll make an even-money bet, though – same rules.) There could be actual savings from these programs, but these five examples of biostatistical sleight-of-hand suggest that those actual savings, if any, are far more modest than claimed savings.
1. Conflating verb tenses
Here is a claim by a diabetes prevention vendor showing ROI on its program. One would be excused for thinking that these results had actually been achieved and validated, given the choice of verb tense in the graphic:
Looking harder at the language, note that the phrase is “recoup their investment,” not “were validated by the Validation Institute as having recouped their investment.” Yet the verb “saved” is in the past tense.
And the heading says: “How quickly employers recoup,” whereas the only article cited in the footnote analyzed Medicare patients, whose chronic disease are far more advanced.
So only a small minority of pre-diabetics will develop diabetes in five years. And then, of course, it would take years for avoidable complications to develop even if no one “manages” the newly diabetic employees to avoid them.
How, then can $1,338 of expenses per participant be claimed to be avoided by keeping these employees from getting sick in a measly two years when most of these employees aren’t going to be sick five years from now even without an intervention?
Comparing participants to non-participants
Speaking of “participants”…
Let’s be very clear: Whenever you see the word “participants” in a study report, the claimed outcome is vastly overstated. Participants will always outperform non-participants, regardless of the intervention. The National Bureau of Economic Research proved this using a randomized control trial. Further, on three other occasions, biased researchers trying hard to prove the opposite accidentally showed that 100% of their apparent “savings” were attributable to participation bias, meaning 0% to the program. (The bias won’t always account for 100% of the claimed savings, of course.)
The best example of this bias? A Koop Award-winning wellness program accidentally revealed that participants hugely outperformed non-participants even when there was not a program to participate in. In this slide, note the X-axis. The groups were separated in 2004. The program was implemented in 2006. During the two years between separation and implementation, the participants “saved” almost 20% vs. the non-participants by doing nothing.
This isn’t a secret. Participation bias is well-known to insiders in the diabetes industry. Yet every single diabetes vendor ignores this bias (or “matches” participants to some medical charts), while most also fail to disclose the dropout rate – and the fact that most employees who drop out of programs do so because they aren’t getting results.
Projecting participants’(!) short-term weight loss into the future
Essentially all of them do this, too. Very large-scale studies have shown that only the smallest percentage of people who lose weight keep it off. There is no reason to think that somehow a few diabetes vendors have unlocked the key to long-term weight loss that has eluded the rest of the world and all academic researchers, especially when the vendors don’t follow employees for the long term or count dropouts.
Rhetorical and arithmetical sleight-of-hand
This single set of claims from a diabetes vendor looks quite impressive at first glance:
Now look at it again, paying special attention to the underlined words:
Once again, there is that word “participants.” That is just the tip of the invalidity iceberg. Six variables were tracked…and yet 27% of active, motivated participants weren’t able to reduce any. Randomly, three should decline. And many of the other 73% of participants could reduce only one…and this is considered successful?
Further, these statistics look like averages on first glance –but they are not. They are examples (“improvements such as”) of reductions that maybe a few participants achieved. I can guarantee that, absent statins, virtually nobody reduces triglycerides by 29%.
Regression to the mean
Diabetes vendors often split the population into high and low utilizers and claim credit for reductions in high utilizers (whom they manage) while counting the utilization increase in low utilizers toward their “savings” — as though last year’s high utilizers also would have increased had it not been for the program.
In this case, there are two giveaways that the 59% decrease in admissions is totally or mostly regression to the mean. The first giveaway is observing which diagnostic categories account for the bulk of an employer’s admissions. This is the top 10 list, in descending order. (“Del” means “delivery.”)
To begin with, the majority of admissions on this Top Ten list – and about a third of all employer-paid admissions — are birth events, not affected by a diabetes program.
You don’t see diabetes on this Top Ten list. That’s because diabetes itself in the last year for which complete data is available accounted for less than 1 admission per 1,000 commercially insured people (126,710 admissions in about 150 million privately insured people). Diabetes admissions don’t even crack the top 25. Because total employer-paid admission rates are about 50-60 per 1,000, eliminating every single diabetes event would decrease admissions by – get ready — about 2%.
Reducing admissions by 59% would require wiping out not just every diabetes admission but also almost every admission not related to childbirth. The vendors might argue that temporary weight loss and eating better reduce other admissions, too. However, the only non-childbirth events in the top 10 are septicemia, joint replacements and pneumonia. Good luck crash-dieting your way out of those.
The other giveaway that this seemingly impressive “decrease” is regression to the mean is that the non-program-members (the vast majority of the population here) regressed upward to the mean. There is no reason to think that admissions in the average employee population are going to increase 4%. Over time, inpatient admissions in the commercially insured population are falling.
Using a selection methodology that is partly dependent on having high claims in the baseline assures both this apocryphal 55% “decrease”– and the equally apocryphal 4% increase in non-member admissions.
For instance, about a third of all heart attacks occur in people who did not have a pre-existing CAD diagnosis. Therefore, if you “manage” patients with diagnosed CAD, you will show a one-third reduction in heart attacks in that population, simply because you didn’t tally the heart attacks in the cohort you didn’t manage.
Then you’ll separately tally the employees without a pre-existing document CAD diagnosis, note the increase and say: “See how fast heart attacks increased in the population we didn’t manage.”
The right answer, of course, is to add the heart attacks in both cohorts back together. Naturally, you’ll find no reduction at all.
Coming soon: What is the Solution?
The next installment will cover how you should measure outcomes to avoid being taken advantage of and to see what really does happen in your population when you implement a diabetes prevention program.
The gist of my presentation compared the current efforts to launch rockets into space with our healthcare system.
In the space race, there are two major players at this time, United Launch Alliance, composed of Boeing and Lockheed Martin, with decades of experience and strong government relationships, and SpaceX, the Elon Musk company.
ULA is like our current healthcare system — big names, big contracts, major impact on and strong relationships with our federal government — and the rockets cost a lot of money. In fact, ULA could also be compared to the Cancer Moonshot, which also has big names with strong government relationships and has big bucks. The Cancer Moonshot approach of using big data analytics, biologics and CRISPR to edit out the genetic defect are all needed and are all great ideas. They are also shiny objects, and they will most likely cost a lot of money.
“You never change things by fighting the existing reality. To change something, build a new model that makes the existing model obsolete.”
SpaceX is the upstart that is doing just that. It is based on Elon Musk’s original vision to re-energize the public to space exploration by putting a greenhouse on Mars. This initial vision has become the goal of “enabling people to live on other planets.” He quickly discovered that he could not do it with the rockets developed because they cost too much. So what did he do? He devised a new system/rocket and removed the waste, the waste of throwing away the rocket, resulting in lower costs and making his dream reasonable. Well, he did that and much more. His launches are considerably cheaper than those of the big guys of ULA. His Falcon Heavy is only the latest example:
“The launch contract will cost the U.S. Air Force $130 million, far less than the $350 million average cost of United Launch Alliance’s Delta IV, previously the heaviest lifter in the U.S. arsenal.”
So what does that have to do with healthcare and diabetes in Mississippi?
In 2015, Mississippi ranked first in the nation for overall diabetes prevalence, with more than 333,000 adult Mississippians living with the disease; that’s more than 14.7% of the adult population
Diabetes accounted for more than 1,000 deaths in Mississippi in 2015
In 2013, direct medical costs (e.g., hospitalizations, medical care, treatment supplies) accounted for about $2.4 billion, of which Medicaid spent almost $1 billion.
MS has an estimated 30% of adults with pre-diabetes, creating the potential that more than 600,000 Mississippians are on the path to develop type 2 diabetes
Yet we know that perhaps 80% of type 2 diabetes is preventable. We also know that an estimated 30% of healthcare is waste, fraud and abuse. So that’s roughly $800 million in waste, etc. that if freed up from the system could be applied to the social determinants of health that are driving this disease.
Imagine that, the money to solve the problem is locked up in the system itself.
Why not create a grand mission just like Elon Musk’s mission to Mars. A mission that people can work toward, as they do the incremental changes needed to create the new system to make it happen. Lifting a quote from President Kennedy, I said:
We chose to eradicate every case of lifestyle-related type 2 diabetes and pre-diabetes in the state of Mississippi, for no more than we are spending today on healthcare. We chose to eradicate every case of lifestyle-related diabetes and pre-diabetes, not because they are easy, but because they are hard; because that goal will serve to organize and measure the best of our energies and skills, because that challenge is one that we are willing to accept, one we are unwilling to postpone, and one we intend to win.
It’s a heavy lift, no pun intended, and it will take decade(s), but it can be done. It will require new systems, a long-term approach and a lot of small changes to get there. If we created the system to do this with diabetes, we could then apply it to the rest of the preventable issues, for we will have developed solutions for diet, exercise, patient engagement, adherence, appropriate medical care, rural care, urban approaches, personalization and on and on.
For years, insurance companies have taken steps to improve the life insurance underwriting experience in the hope of removing obstacles and decreasing not-taken ratios. To that end, some have forgone the traditional exam altogether in favor of simplified issue. But the truth is, consumers still aren’t flocking to life insurers, and the results of these efforts have been incremental.
Force Diagnostics has taken a different approach. We’ve developed a consumer-centric process featuring rapid testing that delivers results in 25 minutes. Tests are performed outside of the home in retail clinics and pharmacies, and results are immediately transmitted directly to the carrier’s underwriting engine for immediate processing. Because of the speed to results, innovative insurers and reinsurers could offer an accurate quote for life insurance to their consumers within 24 hours. And with the benefit of testing with fluids (HbA1C for diabetes, cotinine for nicotine, lipids for cardiovascular risk and the presence of the HIV virus, as well as body mass index and blood pressure), insurers may offer the majority of their products quickly and with assurance.
Once the calculator is downloaded, you may select a typical life insurance policy from a dropdown menu and enter assumptions that reflect an existing underwriting process. The calculator then shows a comparison on underwriting costs, internal rate of return (or IRR) increases, issued policy increases and the potential effects on persistency. At the end, total costs per app are calculated, as are total profits.
There is tremendous value in improving the customer experience throughout the underwriting process.