Let's not forget that what's even more profitable than 'pricing insurance according to actual activity' (which is a nicely sounding PR cover up) is 'dropping people with early symptoms of a disease before it gets found by their doctor'. Which is unethical, but legal (right?) and extremely, extremely profitable.
Keep all premiums, drop people with cancer before they find out about it. And yes, if it's possible to infer pregnancy from shopping bills, then inferring a early health problem from a detailed 24/7 tracker data will be much easier.
And if you want to opt out of this? 'Here are our rates without a fitness tracker discount, they cost 5 times more, ok?'
.. which is why a government guaranteed insurance with good overall coverage is necessary. severe market defects in health care furthermore make dictated prices for services/prescriptions necessary. Adjusting Obamacare for that you end up with the Swiss system - still expensive bur at least effective and IMO fair.
Thinking about it, increasing information asymmetry due to machine learning applied to marketing may make heavy regulations necessary in most other sectors as well. Imagine software that tracks every person and figures out what they need when, identifies a monopoly of opportunity for any given good and them jacks up the price just for that person individually - in real time, while the person walks up to it.
Have you looked at the BAT project? Seems to me that something along these lines could bring such a vision closer.
As in, your browser blocks tracking / identification by default (integrated TOR?). Sites can offer you a bit of tokens if you agree to share your identity / if you let them track what you're looking at. It's a bit of an extension of what BAT aims for, but I think their vision goes in such a direction.
I could see such a thing taking hold, just like AdBlock Plus and co. have become a thing that advertisers / publishers need to work around.
This should quickly be declared illegal, if not already.
With the amount of data that is being gathered on us, coupled with advances in machine learning and statistical analysis, in the near future, "they" will be able to figure out high risk scenarios before we even know about it.
If they drop coverage or increase premiums based on this, they are no longer an "insurance company". I don't know what we would call them.
> This should quickly be declared illegal, if not already.
Dropping coverage like that is illegal already.
> If they ... increase premiums based on this, they are no longer an "insurance company". I don't know what we would call them.
Quite the contrary - an insurance company, by definition, is expected to price insured entities based on the expected future payouts (ie, the levels of risk). More specifically, the sum of all expected future premiums that a patient pays should be slightly more than the sum of all expected future claims for that patient.
If you're expecting it to behave otherwise, then you're looking for something that isn't insurance - insurance is, by definition, a model that relates risk and price.
That's not insurance, you're basically describing a health savings account.
Insurance spreads the cost of all payouts by all patients as to spread out the risk amongst every patient. Not every patient will be profitable. The goal is to have enough healthy, profitable patients to balance out the sick individuals.
No, but no individual plan is ever expected to lose money. If that's the case, you're talking about something that isn't actually insurance.
You're illustrating the original point: when people talk about "health insurance", they're already talking about something that fundamentally is not insurance. If health insurance were insurance in anything but name, then increasing premiums based on expected future risk would be exactly how we would expect health insurance companies to behave.
"" More specifically, the sum of all expected future premiums that all patients pay should be slightly more than the sum of all expected future claims for all patients ""
You can not reduce the calculations to a single patient. If an insurer finds out that you will develop a certain type of cancer for which median treatment cost is 1M USD and lifetime support costs will be 30K USD/year, they should not be allowed to increase your premium accordingly.
> Shouldn't that read: "" More specifically, the sum of all expected future premiums that all patients pay should be slightly more than the sum of all expected future claims for all patients ""
No, it shouldn't. I mean, the thing you wrote is a true statement too, because of the distributive property of summation, but the point is that insurance does work on a per-plan level. If any individual's expected future premiums are less than their expected future payouts, then it's not insurance.
There's a common misconception that insurance is about pooling risk between patients, which is not really the case - it's true at an accounting level, but only as a secondary effect. The fundamental purpose of insurance is to smooth risk across states of the world, which is a completely different thing.
Pretend for the moment that spontaneous combustion of humans exists, can be predicted with some understood probability model, and is an insurable event. In that case, "spontaneous combustion insurance" is not about smoothing risk between Alice and Bob - it's about smoothing risk between universe A (in which Bob spontaneously combusts tomorrow) and universe B (in which Bob does spontaneously combust).
Spontaneous combustion is an easy example to use illustratively because it's a binary event, terminal (you can only combust once), and intuitively independent (by definition of spontaneity, Bob's combustion does not impact Alice's chances of combustion). This is a simplified version of the general insurance model, however, which is capable of handling non-binary events, non-terminal events, and interdependent events. In all of those cases, the risk smoothing happens between states-of-the-world, not between individuals.
> If an insurer finds out that you will develop a certain type of cancer for which median treatment cost is 1M USD and lifetime support costs will be 30K USD/year, they should not be allowed to increase your premium accordingly.
That's a value judgement that you're free to make, but at that point, we're no longer talking about insurance in anything but name.
"Aetna now reportedly has ambitions to offer it to a wider field — adding large swaths of new health data to pull from and giving the health insurance company insight into the activities of its customers."
I am against it not for privacy reasons but instead because it is an expensive option for what it offers. there are many lower cost options that can perform the monitoring they claim. People get mad at the cost of insurance, well here is one major reason why.
This would almost be as bad as when the local police forces and post offices near us got Segways. that and when schools were bamboozled into buying iPads for all students.
Apple is padding their bottom line, nothing more. Aetna is going along because it must make some higher ups feel good by association.
"Now, the story is that the overstock has to be got rid of somehow."
"Apple Watch sales are up 50%"
These two statements are not in opposition if lots of watches are selling at a deep discount. When I google "apple watch amazon" the top link is a used model, second is a discounted series 1.
It'll be opt-in and this isn't a new thing - there's several UK insurers that'll give you discounts and bonuses for life insurance if you log workouts with their apps. This just makes it easier for both sides.
I suspect they can detect more than just how active you are with this thing.
There are probably enough sensors to detect untreated sleep apnea for example. Or the pattern of arm movement, heartbeat, sitting in one spot, etc, associated with binge drinking. Or early onset of Parkinson's. Or perhaps false positives associated with these things.
I'm not personally willing to create the equivalent of an OBD-II port on my body and let an insurance company have unfettered access to the data. Insurance companies, at least in the US, have shown they aren't shy about finding ways to cut their costs at the expense of someone's well being.
honestly I'd support a program where the more you work out the lower your insurance premiums would be. Pretty hard to cheat workouts with Apple Watch when it has a heart rate sensor and all that.
It isn't a horrible idea in the abstract but it seems likely sex would trigger data similar to most workouts. Also, there are people who eat healthy but don't have time for the gym. Conversely people who go to the gym everyday but mostly to try to burn up their Big Mac combo meal. From an insurance standpoint, I think HA1C is the only measure we could like use. Things like BMI, etc. are too varied based on genetics.
You generally get what you measure, so be careful to measure what you want to get. Isn't elevated heart rate connected to better health outcomes no matter what physical activity it comes from? Why should the insurance company care if some of it is sex instead of, say, running?
>it seems likely sex would trigger data similar to most workouts
Accelerometers and GPS would allow them to narrow down specific types and patterns of sex as well. Lots of sex in cheap motels == high risk. Correlating customers together..this one has HIV, and appears to be having sex with this customer == high risk. Fun.
For health (which we're talking about), life, or car insurance (since a bunch of UK insurers also have "safe driving" apps that earn you points, etc.), yeah, sure, why not?
Because obviously they can use those to profit by dropping eg people they diagnose (or just infer statistically) with serious problems before they develop, or raise their prices, etc.
This is not about some extra win-win motivation to be healthier...
It's about turning risk (which is supposed to be what insurance companies are paid for) into rent-seeking...
That reminds me, back during the Tamagotchi craze (which coincided with the Pokemon craze) there was a virtual pet Pikachu that rewarded you for walking (it was a pedometer) and the screen would show Pikachu walking while you were. It was pretty cute. Well, it wasn't long before my little brother figured out how to "cheat" He put it on the washer. (I think dryer worked as well)
What I would be afraid of, here, is that if there are signs that big data can uncover that peel out some catastrophic disease from your workout data that would mean your insurance company knows you're very ill long before you do... It would extremely tempting not to do something about that knowledge that helps the company even if it hurts a soon-to-be-dead-anyway person.
But why not. If you're sitting up all night drinking energy drinks and smoking cigarettes while Uber drops off more McDonalds while I'm out running, working out, stretching regularly...
Isn't that how insurance works though? There are always going to be people who don't take care of themselves. Instead of trying to shame them like this, why not increase awareness? Why not reduce subsidies on corn, meat and instead (or at least in addition to) subsidize vegetables and fruits?
If I'm sitting at home, avoiding skin cancer and excessive wear & tear, while you're running around ruining your joints and breathing car exhaust...
...why should I subsidise _your_ insurance?
(really, America is so weird. Even in our rather terrible non-single-payer insurance system it would be illegal to base premiums on lifestyle. Insurance cost is decided by law, you may buy some extras, the end).
Next thing you know you will pay higher premiums when you decide to have children with someone of whom some algorithm says is a poor match with your genes. Why should I subsidise your "inferior" child?
Why should the young subsidize the old. Why should healthy subsidize the chronically ill. Why should anyone subsidize anyone else at all. You should only pay what you use and nothing more.
/s [this tag necessary because idiots seriously believe this]
If Aetna wanted to offer customers discounts on insurance for healthy activity, it would make far more sense to do it with something like a Fitbit. A purpose-built device for tracking activity and heartrate. The Apple watch doesn't do as good a job as the fitbit at fitness tracking, is bigger and heavier, is far more expensive, and gets far worse battery life. Not to mention the fact that the Apple Watch only works with iPhones meaning that it would be useless for the majority of US smartphone users whereas the Fitbit works with both iPhone (as much as Apple lets them) and Android.
Side Note: I switched from an Android watch to a Fitbit for activity and sleep tracking for many of the reasons outlined above.
my speculation is that these talks are specifically in regard to tracking some biometric that the current line of apple watches doesn’t yet have sensors for.
The secret Apple project for potentially
blood glucose monitoring would be amazing. Non-interstitial fluid / noninvasive sounds like a really hard problem that may require a very bright light source and/or a sensor more sensitive than used for heart rate. It's a worthy goal. Even more awesome would be a limited generic mass-spectrometer that could do more, arbitrary blood chemistry quantification in software without having to develop many/additional specialized sensors.
Keep all premiums, drop people with cancer before they find out about it. And yes, if it's possible to infer pregnancy from shopping bills, then inferring a early health problem from a detailed 24/7 tracker data will be much easier.
And if you want to opt out of this? 'Here are our rates without a fitness tracker discount, they cost 5 times more, ok?'