“There is fundamentally no disincentive in how someone’s health data can be used to determine their insurance premiums. Of course, somebody who is doing more effort, walking more, exercising more, gets more benefits versus somebody who is doing less,” said Vishal Gondal, founder of wearables company GOQii when asked if use of health data can make insurance premiums discriminatory. He also noted that health data can potentially help us in moving away from traditional healthcare practices, and adapting to teleconsults, and help insurance companies in developing tailor-made health insurance for their customers.
Gondal was speaking at MediaNama’s discussion on the impact of the PDP Bill on the healthcare sector held on June 19. This discussion was supported by Microsoft, Google, TMT Law Practice and the Telemedicine Society of India. Comments have been edited for brevity and clarity.
Watch Gondal’s bit from our discussion:
Addressing issues of bias, discrimination, and accuracy
1. What if algorithmic bias leads to discriminatory insurance premiums?
When asked, by Mozilla’s Udbhav Tiwari, how companies like GOQii mitigate unavoidable biases and discrimination in the outcome that is inferred from people’s health data, Gondal claimed that the issue of bias doesn’t even arise. “The way we think is how we can incentivise you for healthy behaviour. There are no negative points for anything. In fact, we have a system called GOQii Cash, it’s like virtual currency; and everything you do, you get cash in the game. This cash can be used to get discounts to buy healthy products,” he said.
If someone smokes, they are bound to pay higher premiums, Gondal said. “So, the discrimination is not because you are spending money on smoking, it’s because it is a known behaviour which causes poor health, and which is why no one will treat a smoker equal to a person who doesn’t smoke when it comes to insurance premiums,” he said.
He went on to claim that what looks like discrimination is actually a way to “encourage” people to take up a healthy lifestyle. “There is enough scientific data and enough studies which have linked a lack of exercise, sleep and diet to non-communicable diseases like diabetes, blood pressure, and hypertension, which are the biggest killers. So, on the same basis, the only difference here is instead of discriminating, it is encouraging you. The insurance company is telling the consumer to sleep, exercise, and be healthy. So, it is actually a way to incentivise positive behaviour,” Gondal justified.
“Some of our users exercise two to three hours a day and possibly are equally doing strenuous workout as daily wage labourers. The data is just looking at your heart rate, it doesn’t differentiate between whether you are a daily wage labourer or whether you are running a marathon for forty-two kilometres. So, the strenuous activity is not connected to this. It’s just looking at data.” — Vishal Gondal
The only time when it can be discriminatory is if an insurance company wants to look at your genetic data to decide a premium for you, Gondal said.
2. How can we be sure of authentic data from wearables?
If data from wearables is used to determine insurance premiums, then the accuracy of the hardware itself will have a bearing on the outcome, a member from the audience noted. In response, Gondal said, “You have to remember that these are not medical devices, and nobody is claiming that they provide accurate medical information. These are, at best, devices which show you trends or which show you a good indication of how you are”. “If someone was walking only five thousand steps in January, did that trend go up from five thousand to six thousand, to seven thousand? If that trend has gone up, it will automatically impact your weight, your other parameters,” he explained.
“We have created what is called the health scoring model, and the health scoring model incorporates both, qualitative and quantitative aspects of various points and puts together a score. Again, we will keep bettering the system, but the system is not just looking at absolute numbers, it is basically looking at what was your baseline and how you improved it. That improvement is what is important, because currently insurance companies know nothing…We are building a mechanism where the insurance company is equally incentivised to be part of your health journey,” he said.
What about patients’ privacy?
Gondal said that GOQii deals with such massive datasets while following relevant privacy laws of specific regions, or other sectoral regulations. “In the absence of specific laws, what GOQii does is either follow the HIPAA guidelines or the GDPR guidelines. If we end up working with insurance companies, that sector has its own set of guidelines. So, the good news at least from our perspective is that since we are working with entities who already have their own privacy guidelines, we are following that, and as we go into international markets, we follow guidelines there. “However, the challenge remains on how will the data be governed, who will be giving consent to share this data. In the case of GOQii, we take consent from every consumer we work with because they are part of our app and they also have the option to stop giving consent. So, those are questions which I hope the Personal Data Protection Bill will answer,” Gondal said.
What if someone is identified based on the data they provide? When asked if companies can try and identify people who are providing them their health data by virtue of using their products, Gondal clarified that GOQii is not “catching people and putting devices on them or sensors on them. Consumers are doing self-tracking, and that tracking is done on our platform. Then the consumer decides whether they want to give this data to the insurance company. But, the data in itself is anonymised,” he said. “Anyone can download our app, create an account using any email ID or number,” he claimed, adding “we don’t really need to know who it is”.
“What we are more interested in is what is your BMI, the coach has to give you inputs on your health lifestyle, we would need health information. So, depending on what you are looking for from the platform, that kind of data has to be provided.” — Vishal Gondal
Data storage and portability: Gondal said that data storage is a part of their entire system. “You cannot prevent storing, of course you can ask to delete it later on, but we do store it,” he said. “Depending on what portability platform users want, we can do it,” Gondal said, when asked if their systems make it easy to port data, which is a requirement under the PDP Bill. “Currently nobody has wanted the data to be ported, so we haven’t. But at the right point of time we have the tools to port the data into any format,” he added.
‘We want to build health outcome centric insurance plans,’ Gondal said
Healthcare datasets can also help insurance companies to assess population behaviour, and develop specific products accordingly, Gondal said. “The future that we are envisaging is that, how can we build health outcome centric insurance plans, because right now insurance companies are blind. They sell you an insurance plan and then they are just waiting and hoping and praying that you don’t end up becoming sick, because the claims are big,” Gondal said.
A ‘Health Score’ can determine insurance premium rates: He said that when consumers use wearables like the ones GOQii makes, they can be “scored”, based on the data that these devices capture. “Insurance companies’ entire risk management is broken. But with our platform and the data, consumers who are adopting a healthier lifestyle with continuous engagement can be scored, and just like credit scores are well accepted in the banking system, we believe that eventually health scoring will become a model to allow discounting of premium and rewards for healthy behaviour,” Gondal said.
GOQii is already developing insurance products as part of a sandbox: GOQii is working closely with the insurance regulator, IRDAI, and with other insurance companies as part of a regulatory sandbox, he said. “We are working with Star health insurance, Bajaj Allianz, Kotak and Max Bupa and there are a few other products in line,” he said. The regulator has defined the maximum number of customers, and has also put a cap on the total amount of insurance premium collected, he added. Right now these insurance products are available to to a “very small group of customers,” he said.
‘Health data can potentially bring premiums down’: The goal really is to “incentivise” people for leading a healthy lifestyle, Gondal claimed. “The inferences are going to be around sleep which is the base, nutrition, fitness, cognition and finally happiness, and we believe this has an impact on longevity…longevity has an impact on insurance, it has an impact on health insurance, life insurance and even as a government itself, we know India’s life expectancies is till 67-68, Japan and other countries are about 80-85. But those economies where they have people living longer but who are sicker are becoming a burden on the healthcare system,” he added.
Can insurers mandate health data for providing services? When asked what will happen if insurers mandate health data records for providing insurance, Gondal claimed that “they cannot mandate it, they can only create a program which incentivises it”. “If you do not invest in good health, then you are going to spend that money in bad health, for hospitals, doctors. What we are saying is, instead of spending that money on your treatments, spend it on healthy food, spend it on exercise, spend it on workout,” he added
Health data will make healthcare not only accurate but predictive: Gondal
Talking about a recently launched GOQii device, Gondal noted that it has sensors to track body temperature, blood pressure, heart rate, and sleep among other things. “We are collecting, on this device, about twenty-five thousand data sets per person per day. Eventually like I said, if everybody is going to be wearing a device of some kind, and that combined with your phone will become the primary means of diagnostics. Of course, these may not be medical grade yet, but eventually combining this with AI and machine learning will make it not only accurate but predictive,” he added. Personalised medicine, where a doctor will prescribe medicines based on people’s “particular parameters”, and “genetics” will become possible as more data points become available, according to Gondal. Food and nutraceutical companies can use people’s data to “tailor make products for diabetics or for vegans who are diabetic, and for other specific lifestyles,” Gondal said.
Moving away from traditional healthcare: Gondal said that as devices collect more data, it will end up transforming the role of a doctor and treatment methods. “The doctor’s role is being modified to become a guide and use the data to drive 360-degree care. So, it’s no longer about, I have a symptom and I have a pill. Now it’s about how you can, on a long-term basis, manage your healthcare. This is only possible with data, and that is what suddenly is getting unlocked”, he said. GOQii already has something similar in place, Gondal noted. “We call this the GOQii data pool where we are ingesting data from health wearables, lifestyle data from consumers, medical data, fitness data, and doctor notes. All of this, both qualitative and quantitative information is right now available in a pool with GOQii,” he said.
“So, what we are talking about is a new layer in the healthcare system. Currently the system was, you fall sick, you go to a doctor, doctor gives you medicine, you go back. Now, you are going to be continuously monitored. And even before a sickness comes in, your body gives many signs, and those signs will be potentially caught and that data will be analysed.” — Vishal Gondal
Health data can aid in medical research: “We are talking with the medical research fraternity, and already have tons of users with all kinds of health issues on the platform, and their anonymised data could have a massive impact on medical research and medical researchers can plug into it.” He explained GOQii’s partnership with a German company called Thryve which is using anonymised health datasets amassed by GOQii for predictive healthcare: “What Thryve is able to do is put predictive algorithms on top of our data. So, instead of you waiting to get sick or have COVID or any kind of conditions, these algorithms are trained to detect depression, heart disease, diabetes and so on. As the data becomes more and more better in quality, these algorithms are going to become that smarter,” Gondal stated.
Also in this series:
- #NAMA: Personal Data Protection Bill needs to work on basic definitions to cater to the healthcare sector