An accidental experiment that saved 700 lives | Hacker News Reader

An accidental experiment that saved 700 lives

statmodeling.stat.columbia.edu

96 pointsluu102 comments

> the first rigorous experiment to find that health coverage leads to fewer deaths, a claim that politicians and economists have fiercely debated in recent years

Living in Europe I must have missed this debate. Is it not self-evidently true? What are the arguments against it?

Most studies on the topic shows that having health insurance in the US doesn't impact people's health much or at all.

Everyone agrees this is counterintuitive, but that's been the data so far. This study adds some weight to the other side.

> Most studies on the topic shows that having health insurance in the US doesn't impact people's health much or at all.

I can't speak to "most" having not seen all of them, but studies have shown it matters.

https://www.theguardian.com/us-news/2017/jun/24/us-healthcar...

These came up a year later:

https://www.ncbi.nlm.nih.gov/pmc/articles/PMC2323087/

https://news.gallup.com/poll/268094/millions-lost-someone-co...

Seems there has been some prior evidence for supporting the idea that health insurance (the only way to get affordable healthcare in the US) matters.

What are examples of studies that show it has little to no impact?

No no. They say -mortality- isn't affected. Like, people don't die more often without health insurance. There are all sorts of measured benefits of health insurance, quality of life, lower prevalence of medical debt, etc. It's just that a lack of it doesn't lead to people dying sooner.

Which kind of makes sense; the cost even with insurance ensures people don't go to the doctor as often as they should, until something is concerning them. And if it is sufficiently concerning, even someone without insurance goes. And if it's going to prove lethal without treatment, even those without insurance will probably still get the treatment, since better to be alive and in debt the rest of their life than dead.

Someone else linked https://en.wikipedia.org/wiki/Oregon_Medicaid_health_experim... and it's a good example; it calls out that those with coverage are less likely to have medical debt sent to a collecitons agency, which implies that those without coverage are still seeking treatment...they just are going into debt for it.

The Wikipedia article also says:

"Approximately two years after the lottery, researchers found that Medicaid had no statistically significant impact on physical health measures"

In the US you still get emergency medical care without insurance but then have to deal with the bill. Sometimes that means bankruptcy, sometimes its forgiven. The elderly (65+) have universal care through the government account for most mortality.

So with those combined, it means the impact of coverage on mortality is small enough that its hard to measure. Of course avoiding financial catastrophe is also a good reason for some kind of universal coverage even if the direct health benefits are low.

Right, but none of the two methods you mention will get you treatment for any cancer your get in your 40s, right?

Is the argument that before 65 people just don't die from anything medical, except things emergency room will treat?

But then there's also the secondary effects, like losing your house because of medical bills. And we know that even with universal health care (every civilized country except the US) health is associated with money.

Another reply asserted that studies say it doesn't help much, but... that's just so counter intuitive to me that the the studies that were replied there saying actually it does help can be summarized by "well, duh".

But I've thought "well, duh" about false things in the past, which is why I want to know.

At the end of 2017, Congress passed legislation eliminating the health law’s fines for not carrying health insurance, a change that probably guarantees that the I.R.S. letters will remain a one-time experiment.

I still get a form from my employer proving that I had health insurance, so even with a financial penalty there's no reason that the letters can't continue to be sent out to the uninsured to remind them of options.

From paper: “Of the 4.5 million households who met the criteria for inclusion in this pilot program, 3.9 million were randomly selected to receive the intervention.”, “The final sample consists of 4.5 million returns, corresponding to 8.9 million individuals.”

Those in the treatment group were 1.3[1] percentage points more likely to enroll in coverage in the year following the intervention than those in the control group: about 50,000 extra households (~100,000 extra people) were signed up due to the letter.

So, about 200 died because they didn’t receive the letter (But that was offset by the 8 billion lives that were saved by something else that didn’t happen that day.).

[1] “those in the treatment group were 1.3 percentage points more likely to enroll in coverage in the year following the intervention than those in the control group, a 2.8% relative increase. On average, each letter increased coverage among this group by 0.14 months during 2017, or one additional year of coverage per 87 letters sent. We document larger effects among individuals who lacked any coverage during the prior year and among older non- elderly adults.”

Edited: corrected numbers - mistook households for individuals.

Not to be a nihilist, but 200 seems like a ridiculously and surprisingly low number for impact of such an expensive and broad program. Does it really validate what the author posits?

Bear in mind nearly all of the people receiving letters should be <65. I quickly skimmed the paper and it seems like the authors don't have enough data to come up with a quality adjusted life-year type of statistic like the NHS uses.

Note that only those influenced to buy insurance would show any possible benefit.

On second thoughts, that should be 100: I stuffed up mixing households versus individuals.

“The final sample consists of 4.5 million returns, corresponding to 8.9 million individuals. Individuals in the sample were randomly assigned to receive a letter (86%) or to a control group (14%). One letter, addressed to the taxpayer(s), was sent per return. Hence, randomization was conducted at the household level.”

So 86% of 4.5 million households were sent a letter, and that led to a 1.3% percentage point increase in signups compared to the 14% of households that were not sent a letter. They are then saying that those 1.3% of extra signups led to saving ~700 lives. There is the risk of a selection bias because we don’t know why those extra 1.3% signed up versus those that didn’t.

The actual paper uses months, since everyone dies, so zero lives can be saved in the long term.

As a counterpoint, the Oregon Medicaid health experiment was also an RCT but found the opposite.

https://en.wikipedia.org/wiki/Oregon_Medicaid_health_experim...

> Approximately two years after the lottery, researchers found that Medicaid had no statistically significant impact on physical health measures, but "it did increase use of health care services, raise rates of diabetes detection and management, lower rates of depression, and reduce financial strain."

How is that the opposite of "An accidental experiment that saved 700 lives"?

An intentional experiment that saved no lives?

I fear for things mostly managed in primary care like diabetes, hypertension, etc. the period of this study is just not long enough to get adequate amount of mortality data.

> The experiment, made possible by an accident of budgeting, is the first rigorous experiment to find that health coverage leads to fewer deaths,

It, of course, does not prove this. It proves that within this specific set of circumstances, within the specifics of our current health care system, that increasing coverage lead to fewer deaths. You can't just automatically extrapolate that to health coverage in general.

And there exists at least one field in Scotland in which there exists at least one sheep, at least one side of which is black.

Look, this was literally a randomized controlled study of the entire US population. You know, the gold standard for generalizability. And what other point of comparison would possibly be relevant besides "the specifics of our current health care system?" What more do you want?

I'm not complaining about the study methodology. I'm just saying this isn't the right takeaway.

Just because a study is randomized and has a good sample size doesn't automatically prove every hypothesis of the study. You have to actually look at what the data is saying and isn't saying.

And nothing about this data says anything regarding some universal truth regarding the benefits of some universal thing called "health coverage".

Well, let's see:

* More people obtained health insurance after receiving the letters than those who didn't.

* Among the people who received a letter, fewer people died than those who didn't.

* Since we're agreed that the study was properly randomized, there were no other differences between the treatment and control populations.

What conclusion are we to draw, if not that having health insurance prevents deaths?

Literally the only differences are that the treatment group got a letter, more of them purchased this thing which then existed called "health insurance" than in the control group, and fewer people in the treatment group died over the course of the study. Are you going to sit there and tell me that people who were already less likely to die went out and bought the thing that at the time of the study was called "health insurance" just because they got a letter? That doesn't pass the smell test.

Are you trying to tell me that health insurance is somehow different now than it was before? Also doesn't pass the smell test.

I am at a loss for what you are actually trying to say. Could you please just say it, rather than being oblique to the point of opacity about it?

The point of the individual mandate was never that it was good for the individuals, but rather that it avoids ‘adverse selection’ where healthy people see the price and decide they can’t afford it but the mom with MS who has a kid with spinal deformation thinks ‘what a bargain!’

This is the same argument for any tax. Tollbooths, for example, have quantifiable deficiencies, but feel more fair... to some.

In the long run, it's trying to avoid the adverse selection death spiral where the price keeps rising to match the costs, and squeezing more people out of the pool.

Has this been happening now that it was eliminated? Costs seem the same as before for me.

My premiums went up 10% this year, to over $2,400/month for the family.

That’s normal inflation for health care.

For comparison, in New Zealand, assuming you are earning NZD180k (well into the top 1%[1]):

Your taxes are about NZD70k in total. About $12k of that goes towards heath[2].

The most expensive family plan (2adults 3kids) to top up your cover (beyond the normal care everyone gets by the state) is NZD700 per month (the major health insurance provider in NZ[3]).

[1] https://www.ird.govt.nz/about-us/tax-statistics/revenue-refu...

[2] https://www.treasury.govt.nz/information-and-services/financ...

[3] https://join.southerncross.co.nz/quote

Thanks for this break down. Very interesting to see how the health insurance cost is similar to the US but the payer is different.

In the US, if you are lucky, your employer pays the bulk of the premium. If you aren’t, you try and find a cheaper ‘death and dismemberment’ plan (only covers tragic events ) so you aren’t bankrupted if something really bad happens.

The terrible part is for those folks, there is no state plan to fall back on. Insurance and providers are more than happy to bill you into crushing debt.

Was this ever a real problem, before or now after the individual mandate?

"Empirical evidence of adverse selection is mixed. Several studies investigating correlations between risk and insurance purchase have failed to show the predicted positive correlation for life insurance, auto insurance, and health insurance. On the other hand, "positive" test results for adverse selection have been reported in health insurance, long-term care insurance, and annuity markets."

https://en.wikipedia.org/wiki/Adverse_selection

Before Obamacare adverse selection was managed by the "preexisting condition" system.

Much like how you can't sign up for fire insurance when your house is already burning, you couldn't get insurance only after you need care.

Obamacare did away with "preexisting condition", and the mandate was meant to replace it.

With nothing at all acting as a backstop, AFAIK, I assume this is now baked into the ever exploding "cost disease" of US health care.

> Obamacare did away with "preexisting condition", and the mandate was meant to replace it.

The Obamacare "fix" already existed in several states. You believe that such a policy has certain consequence - can you identify those states by looking for that consequence?

Note that pretty much every state had "if you keep your care, you can change providers", so the problem was only for folks who dropped coverage and then wanted back in.

If you have a relevant condition, dropping coverage seems like an odd thing to do.

Some states had high risk pools. Limited coverage, limited access, high premiums.

I was unclear. While some states did have high-risk pools, other states had "can't turn down" rules much like Obamacare.

Can you identify the "can't turn down" states by looking for the predicted health outcomes of such a policy?

In one sense we were a can't-turn-down state. No limit on the premium, though, and note the response I received from an insurance application. The can't-turn-down option wasn't meaningful.

That's some serious twisting of words to claim that straight up denying people health insurance is "managing" anything.

No, it's just using words.

You can describe things without stating an opinion about the things!

No, sorry. Implying that adverse selection is "managed" by excluding people from having health insurance implies that the system actually works. Since large numbers of people are, by definition, excluded from having health insurance under such a system, it de facto does *not* work.

This is not an opinion; it's a fact. A system that, by design, fails those who need it the most in order to syphon money off from those who need it the least is a failure.

You seem really confused about the English language.

It used to be that if you had health issues you simply couldn't buy insurance, period. My former employer failed in the housing collapse, trying to get private "Please feel free to reapply when you have a diagnosis." I'd love to have even a real diagnosis, but I do not expect I ever will.

A person who thinks they are healthy but goes for an annual checkup just because it is covered by insurance is still benefiting from this.

unless they get unnecessary treatment and/or unneeded testing.

Ha… I’d love a plan that covered an annual checkup. Never seen one of those as an option.

Weird! So many things cost less, if caught early. A lot less.

Hey I’m on the same page. Doesn’t make any sense to me.

All health plans in the USA must fully cover an annual physical and other preventative services (like shots and some kinds of screenings) under the ACA.

From hhs.gov:

>Most plans must over [sic] a set of preventive services – like shots and screenings – at no cost to you.

>For example, depending on your age, you may have access to no-cost preventive services such as:

And goes on to list many things but not an annual check up. And also notice “most”, “depending on your age”, and “may”.

So if you schedule a visit to a doctor without a specific problem and they give you your yearly shots and do routine tests, and all of that is covered by insurance, what do you call it?

A plan I’d like to have offered to me.

Would you share a bit about your efforts to find such a plan? We other commenters are experiencing cognitive dissonance reading your comments, because in our experience, all the plans we've been on since ACA, cover an annual exam with no questions asked..

> And also notice “most”, “depending on your age”, and “may”.

You say that like you think those are being used as weasel words. But they're just allowing for a more concise summary that does not need to recite every nuance of the requirements, which rightfully are not written as one size fits all. Factors like your age truly do matter to the question of what preventive care makes sense, and it would be a bad public policy to require insurers to cover care as preventive in situations where it does not have preventive value.

Ok, but “most” is not “all” like gp (or maybe ggp) claimed.

It's close enough. That "most" truly is "all" when discussing plans offered through the Healthcare Marketplace.

And since you still haven't provided any clarification about your claimed inability to find such a plan being offered to you, I don't think I need to bother digging up details on whatever obscure products require the statement about preventive care to be qualified. If you're being at all honest about your experiences with health insurance, you can more readily identify some exceptions than I can.

Every plan that was offered to me through the marketplace stated that annual checkups were out of pocket but I could count them against my deductible. I don’t know how to be more clear than that.

> I don’t know how to be more clear than that.

You really aren't putting much effort into being clear and convincing, especially given how you've been shown that people are finding your claims quite hard to believe. You are still not offering any specifics such as which state these plans are offered in, what level(s) of plans you looked at (catastrophic/bronze/silver/gold), what (if anything) in the way of preventive services they do purport to cover, what services you expect from an annual checkup that would not be covered.

There are lots of obvious ways you could have been much more forthcoming to bolster your credibility, but instead you've been more or less evasive.

Silver (and bronze) plans in a southeastern state.

I would like to go to the doctor and not have to pay out of pocket for my annual checkup. I understand that the checkup may raise tests and procedures that would cost, but first visit fees and annual checkup fees and some level of basic blood work shouldn’t be out of pocket like they were in the plans offered to me. (I’d even be fine with a copay of $10-25)

Beyond that I don’t actually care if you believe me. This is what happened to me

They do, but beware fun fees: I've spent hours of my life disputing a psych evaluation charge (under 5 minutes of a pediatrician asking a kid how they are feeling, without any paternal prompting), which the insurance company claimed it wasn't covered in the free physical. So you go to a supposedly fully covered appointment, and come back home to a bill you have to pay, or argue with.

It's not just the insurance company in that case though, the doctor's office would have submitted an additional billing code for the appointment (which for something like you describe should probably be treated as fraud).

The original idea with insurance providers was that they were more educated than their customers, and since they were paying, they would be incentivized and better capable of negotiating with medical providers.

Of course, this has proven not to be the case. Part of the reason is that insurance providers are legally prohibited from making profits beyond a certain proportion to what they actually pay for medical services. So now, they are actually incentivized for _higher_ medical costs.

Other than as part of an overall plan what would be the point? The cost would be the cost of an annual checkup. Other than if you could pay it with pre-tax dollars what would be the point? Insurance is about risk, not about certainty.

If you read the paper, it is not an accident at all. It was an intentional program where 14% of households fined under ACA received a letter to tell them how to enrol. A little sad the US Govt fined the rest but didn't provide this info to them.

Obamacare was likely the only politically viable solution at the time. I wish we would use it as a stepping stone to more comprehensive coverage for all US residents.

In retrospect, Obama seems to have made a terrible mistake with the Obamacare framework. He made the original sin of acquiescing that there would be no single payer, thereby relinquishing his only true leverage over an obstinate Senate clinging to a skin Democratic majority.

The compromise was not a good one. It never made great sense for any of the parties. It was a victory of pragmatism over function, and no Republican voted for it anyway.

And yet now this accidental experiment exists because of Obamacare which settles a longstanding debate in the US as to whether healthcare coverage actually saves lives -- while Europeans go "Well, duh? Why was this ever a question?"

Yeah, but had he not compromised and instead just pushed through a single payer solution ("medicare for all"), it would have passed along the same lines, given us the same data, and have been harder for Republicans to gut since it would have given no individual mandate, or high priced low coverage marketplace plans, to attack.

He didn’t support a single payer system or a public option, he just made mouth noises to that effect. Look at the recent failure in California to bring a single payer bill to a vote, despite the Governor’s ostensible support and Democrats’ 75% super majority in the Senate. Perennial incompetence isn’t a plausible explanation; rank corruption is.

Edit: to add a bit of evidence that Obama specifically is corrupt, here is an article titled ‘Barack Obama Got Paid $400,000 for a Speech to Health Care Execs. Here's What He Told Them’:

https://money.com/barack-obama-speech-400000-health-care-con...

How is that anything but a payoff?

I don’t get it. The same article said he made $60 million from a book deal. $400K is a nothing burger. He would have made money from the speaking circuit regardless.

I don’t understand, you say $400k is nothing, but then say he would have made money from the speaking circuit regardless, as if he didn’t do just that, and as if taking vast sums from powerful corporate interests whose interests he’d manifestly served during his time in office isn’t the definition of corruption. I’ve read claims that he makes about 50 speeches per year, which seems plausible, given that Bill and Hillary Clinton made 729 speeches between 2001 and 2016 for a total of $153 million ($210k average).

See: https://www.cnn.com/2016/02/05/politics/hillary-clinton-bill...

I want to go back to the $400k speech to the healthcare industry. First, the notion that this isn’t corruption because he made so much more money in a less unseemly manner is pretty curious. Does it not then follow that the very rich, so long as their original fortunes are not ill gotten, are incapable of corruption in principle ? Second, $400k is precisely the annual salary of the President(!). For maybe an hour of work. It’s not just egregious, it’s super on the nose. Third, imagine he’d put every bit of political capital at his disposal into the fight to create a single payer system, and against the odds (and the money of those executives to whom he gave the speech) succeeded. Do you think the healthcare industry (or the part that’d still exist) would still pay $400k for an hour of his time?

He had no political capital against the lobbyist who had the democrats in their pocket too. Clinton also tried passing health care reform with a Democratic majority and failed miserably.

At least Obama got something passed. Now, even though I personally would pay a decent chunk of change for health care if I decided to go into independent consulting, at least it’s an option now. It wasn’t before.

He would have made money on the speaking circuit from some group of people regardless of what he did.

Your theory is that Obama pushed through a complicated health care plan that affects millions of Americans and acts as the centerpiece of his legacy, and it was all to set up a $400k speaking fee a year after leaving office?

He makes it clear he’s the sort to play ball, so he gets the money for the campaign. He plays ball, so he gets paid. And it’s not $400k, it’s $10-20 million per year indefinitely. See my other comment—-the Clintons made $150 million on speeches alone between 2001 and 2016.

As President, no matter what you do, somebody benefits. Also, somebody will be willing to pay you six figures for speaking fees. Sometimes those parties are the same.

I guess I just subscribe to Hanlon's Razor on this one. I think he should have advocates for single payer, but I think it was a strategic error, and not a selfish manipulation of the system.

Obama cares about his legacy as much as anyone else would in that situation.

A couple thoughts:

  As President, no matter what you do, somebody benefits. Also, somebody will be willing to pay you six figures for speaking fees. Sometimes those parties are the same.

Presidents have no business accepting speaking fees from anyone. They receive a lifetime pension and various other benefits worth millions of dollars after leaving office. Even if these fees were accepted in good faith, they still have the appearance of corruption, which undermines public confidence in the political system, and cannot but instill in the corrupt the belief that for corrupt acts in office they stand to gain tremendously. My understanding is that the presidential pension was created to avoid the embarrassment of a destitute former president, specifically Harry Truman. Harry Truman conveniently had an opinion on the subject:

  No young man should go into politics if he wants to get rich or if he expects an adequate reward for his services. An honest public servant can't become rich in politics. He can only attain greatness and satisfaction by service.

Further, Jimmy Carter rarely accepts speaking fees, and when he does he donates them to charity.

  I guess I just subscribe to Hanlon's Razor on this one. I think he should have advocate[d] for single payer, but I think it was a strategic error, and not a selfish manipulation of the system.

This is certainly not dispositive, but I wonder what you think about Obama’s extraordinary intervention in the 2020 democratic primary. Right before Super Tuesday he worked (successfully) to convince Pete Buttigieg and Amy Klobuchar to drop out, endorse Biden, and exhort their supporters to vote for Biden over Sanders. If Obama was so convinced of the need to reform the medical insurance system that he made doing so his highest legislative priority, and if he did believe that in theory a single payer system was preferable to the ACA, but made as you say, a strategic error in not at least advocating for it initially, then isn’t it somewhat curious that he would work so hard to ensure that Sanders, who was the only candidate who plausibly would have attempted to pass a single payer system, lost, and Biden, who not only did not support a single payer system, but said that if by some miracle it came across his desk he would veto it, won?

Biden was his VP and would support his legacy. Probably also the best chance to beat Trump.

Or yes, maybe Obama is the equivalent of a single issue voter.

For folks in the "experiment", if health insurance had a 1 in 1600 chance of saving your live each year, but costs $10K/yr, is it worth it?

> Inside was a letter stating that they had recently paid a fine for not carrying health insurance

What? The IRS can fine you for not having health insurance?

Yes, it was called the Individual Mandate. It was a feature of Obamacare which was later repealed.

https://www.healthinsurance.org/glossary/individual-mandate/

California has a fine for not having health insurance.

> made possible by an accident of budgeting

Would this have been legal/ethical to do intentionally?

Probably not, because how would you get consent from the test subjects without messing up the test?

It's so very weird reading this as a UK citizen. It feels like reading about some kind of dystopian future. Despite the UK government's best efforts to the contrary, even the poorest in society can still access health care locally at no cost... except the hospital car parking fees, that's a killer.

Not everyone has a car.

Most (all?) hospitals are easily reachable by public transport. I've never had a car and never had trouble getting to a hospital when needed.

I got a ride to the hospital.

The ambulance ride cost me $2000 out of pocket.

I'm sorry to hear that, but that kind of thing is not an issue in the UK, or indeed, many parts of the world.

I live in Switzerland. I had a skiing accident and got stretchered down to the main station, then an ambulance down the mountain to the hospital. Since at that point I hadn't filed my accident insurance claim yet, they sent the full bill to me.

Want to know how much it was, in the land of $25 pizzas? Under $300. And of course, I forwarded it to my accident insurance, so my out-of-pocket was $0.

Fix your system.

Also in Switzerland: an ambulance was called for me in the middle of Zurich, they arrived after driving only a few kilometers, performed an EKG, and didn’t transport me anywhere. I was billed over $700 for that, which I had to pay out-of-pocket as the theoretical 50% insurance coverage was below my deductible.

Most likely it didn’t have to pay any money because you either payed your full deductible for the year already, or most likely it was payed by the accident insurance and not the regular health insurance

Yes, the accident insurance paid... but the bill was sent to me first.

Oh, fair point. We'll add a fine for bikes as well. Thanks for helping us remedy this oversight.

Make sure you stay vigilant and keep fighting for the NHS.

Our public healthcare in Australia has been slowly eroded over the past decade (if not longer), and I'm starting to wonder if the next generation will even be able to enjoy "full" coverage at all.

This isn't propaganda or hyperbole. When my wife had a miscarriage earlier in the year, we were billed $150 for the privilege of being told we'd lost our child.

This should not happen in Australia, but it did. Don't follow in our footsteps.

Reply on news.ycombinator.com