« So Are Real Average Wages Up? | Main | And One Last on Brooks »

July 24, 2007

Insurance Matters

One of the common objections to universal coverage is that insurance coverage doesn't actually improve health outcomes all that much. This objection, unsurprisingly, is generally made by people with health coverage. It's also not very true.

More evidence for the importance of health coverage comes from a study in The New England Journal of Medicine this month, which tracks what happens when the previously uninsured become eligible for Medicare. It turns out that they -- surprise! -- need a whole lot more care than their demographically similar, but previously insured, brethren!

They have conditions that need to be treated and managed, exhibit higher rates of hospitalization and doctor's visits, and generally cost a whole bunch more money than those who've been availing themselves of the health care system for years. As the NEJM dryly concludes, "The costs of expanding health insurance coverage for uninsured adults before they reach the age of 65 years may be partially offset by subsequent reductions in health care use and spending for these adults after the age of 65." Add in what they cost the system in catastrophic health incidents before they become eligible for Medicare, and it begins to look like keeping the uninsured population uninsured is a much better deal for the insurers than it is for taxpayers, hospitals, or just about anyone else.

July 24, 2007 in Health Care | Permalink

Comments

There is a discussion of health care coverage issues over at this blog:

http://www.barackobama.com/issues/healthcare/?discussion

Posted by: Jennifer | Jul 24, 2007 9:02:37 PM

Since you bring it up Ezra I thought I'd share this excerpt of a painfully bad interview in Health Affairs. I don't know the OECD data well enough but is she really just cherry picking examples? And what is up with the claim about preventative health care. I thought that was what we were worst on. Or is she saying that it's the best when people get it? Yet that would contradict her claim that in a private HC system the rich don't get better treatment.
http://content.healthaffairs.org/cgi/content/full/hlthaff.26.5.w552v1/DC1

Galvin: Data from the Organization for Economic Cooperation and Development (OECD) shows that quality, according to current measures, is generally far better in the Western European countries than it is here, and this holds for single-payer systems, like the U.K. and Canada.

Herzlinger: I wouldn't say that. I wouldn't say that women who have breast cancer in the U.K. fare as well as women with breast cancer in the U.S.

Galvin: Well, that happens to be an example of where it actually is better in the U.S. But, on the whole, wouldn't you say the jury is still out on whether a single-payer system can control quality better?

Herzlinger: No, no. The OECD and Commonwealth data showed that the U.S. had better preventive care, diabetes care, and childhood leukemia scores in addition to breast cancer. And measures of diseases like asthma and diabetes, which have significant genetic components, must be adjusted for the composition of the population before they begin to make sense.

Posted by: charles | Jul 24, 2007 9:50:23 PM

So where exactly does it show their outcomes are worse or once they are enrolled better for that matter. It mentions nothing of the sort. Not a single mention of outcomes. It talks about utilization i.e. doctors visits, hospitalizations and cost. More utilization does not necessarily equate with better care. Hell anybody could have predicted their outcomes and saved their time and effort. If you give out free tickets to the tilt-a-whirl, do you think more people are going to ride it?

Posted by: Dingo | Jul 24, 2007 10:01:16 PM

Ezra,

I was with you the whole post through until you sullied it at the end... before I get to that.

Substantively and politically, the essence of this post is something that should be strongly highlighted by UHC advocates. Carol Moseley-Braun, when she ran for President, made this one of her top 3 talking points on health care. Specifically, covering the uninsured very likely leads to little to no additional system cost-- existing treatment for uninsured is hidden in insured prices and their current underutilization would be offset by lower long-term costs in Medicare as described above.

But then...

keeping the uninsured population uninsured is a much better deal for the insurers

The reality is the uninsured population could represent an unrealized 20% market potential. As Mike Gravel advocated for last night, universal vouchers, (great idea) this would be something beneficial to insurance companies. Its only a single-payer solution to UHC that is detrimental to them.

Amazing how much there is to learn from these fringe candidates...

Posted by: wisewon | Jul 24, 2007 10:06:55 PM

What logic gets you to the conclusion that insurers don't want to insure the uninsured? How is it a good deal for them?

I think the only way you come to that conclusion is to assume that the premium for these people would not cover the costs. And yet one of the founding principles of insurance is to make premiums reflect medical costs through the actuarial craft. Even CMS and state Medicaid agencies practice this craft, and peg the PMPM (per member per month) premiums they offer to managed care based on medical claims experience so that the companies can take in enough in premiums to pay out the claims and cover admin as well as a profit in most situations.

In other words, insurers are losing revenue and profit by not insuring these people. Medical costs are high enough that individuals without insurance don't find it worth it to buy individual policies at a price insurers can offer and still make their customary 3-5% profit. I don't blame people for not taking the insurance in the individual market, which is an abomination. I've gone without insurance for a couple of years myself. But this situation isn't a "good deal" for insurers. It's suboptimal for everyone, because we're stuck with a market structure that sucks.

As you know, Ezra, Germany, France and the Netherlands all have systems that work well and leave a role for a reconstituted insurance industry. We need to find our own path to a system like one of these. Nothing else is necessary for a rational system. Nothing else is possible given the politics.

Posted by: jd | Jul 24, 2007 10:07:37 PM

Counting down to a conservative explaining that taxpayers should have to pay more for Medicare rather than deprive doctors the paychecks that come from treating people who've gone too long without insurance.

Posted by: Amanda Marcotte | Jul 24, 2007 10:13:03 PM

The only peoploe who truly believe that health outcomes don't improve if you have medical insurance are people who are either very young, very healthy, and/or ideologically blinkered. Clearly they have had very little contact with the actually existing health care system in the US.

Again, for a small example, look at recent studies concerning non-compliance with prescription drug medications that ran in the New England Journal of Medicine in the last year or two.

Posted by: KTLN | Jul 24, 2007 10:21:28 PM

The only people who truly believe that health outcomes don't improve if you have medical insurance are . . . people who have looked at the only large, double-blind study of health service utilisation, the 1982 Rand study.

One wishes that there had been other such, but so far, the Rand study remains the gold standard, and it showed that relieving patients of the cost of their care pushed up utilisation by more than 30%, but resulted in no improvement in medical outcomes. Data-wise, it's hard to trump double-blind with population studies.

Posted by: Megan McArdle | Jul 24, 2007 10:45:34 PM

Megan,

How old are you and have you or a dependent of yours ever had or needed major surgery? Have you ever met anyone who couldn't get necessary care because they were uninsured? I'd suggest you live in the real world for a while before you cite a 1982 study as the gold standard.

Posted by: KTLN | Jul 24, 2007 11:23:13 PM

Megan,

You're hanging on to a 1982 study of the health care system? It's not like anything could have changed in the intervening years, right? For example, say, the prevalence of meds for chronic diseases?

A recent study by Kaiser (the HMO, not the public health organization) showed that when cost sharing for meds went up for Medicare patients, Rx use went down and hospital admissions went up. I think they even saw mortality increase before they pulled the plug on the plan.

Maybe it isn't double-blind, but it isn't clear why this would matter. Life often isn't double blind. People know when they've got a relatively big or small deductible, and so do doctors and insurers. If this knowledge influences behavior, then it does so in real life and a double-blind study could actually DISTORT the results we can expect in the real world.

In any case, I'm not even sure what it means for the study to be double-blind. How can people be ignorant of the size of their deductible/cost-sharing and still respond to it? Or do you mean that they don't know that there are other people in a study with a different amount of cost sharing? Big deal. They knew if their cost-sharing was relatively large or small, same as people do generally, and same as they have in numerous other studies.

Posted by: jd | Jul 24, 2007 11:36:37 PM

Wisewon -- you're right. Insurers would like to cover this group -- well, save for the really ill -- if someone would pay them to do so.

Megan -- This is partially true ad partially untrue. But it's really untrue with the group examined in this study: cardiovascular patients. We've had a host of studies showing the increase in cost-effective, positive health outcomes from treating them with such drugs as statins.

Moreover, the Rand study didn't leave anyone uninsured. It created a group with more upfront cost-sharing. That's, uh, not what we're talking about here. And the one group the Rand study did show health outcomes decline for with cost-sharing was the poor -- exactly the group most likely to be uninsured.

Not to mention you don't believe what you're saying. You, after all, are insured. You seek medical treatment for a variety of disorders you listed on your blog yesterday. Clearly you do believe that insurance and care will improve your health outcomes. Otherwise, you're wasting quite a bit of money...

Posted by: Ezra | Jul 25, 2007 12:40:17 AM

But what happens to the newly insured who avoid these unpleasant, costly conditions caused by lack of access? Are they never so afflicted? Some of them, yes, because they get hit by a bus, or otherwise die young. But most of them now get the life-span/health-span extending preventive care and, later on, still develop those fun chronic and terminal conditions as their bodies deteriorate.
Provision of health insurance can't save money merely because it provides health care. What it can do, however, is make the amount we do spend more rationally directed at improving health outcomes. Efficiency, not savings, is the benefit.
In other words, we cannot shy away from the acknowledgement that UHC will entail rationing. More and better health care, but rationing nonetheless.

Posted by: RW | Jul 25, 2007 12:42:37 AM

Boy, when some people get the bit between their teeth they bite so hard it leaves marks.

Did RW just say that we need to spend money more rationally, and that means we'll have rationing?

Or does he just end every statement with a routine warning of rationing to come, like the dire warning on the cigarette pack?

The whole "rationing" thing depends on the idea that everyone will want to get a colonoscopy, or maybe have a kidney removed, because it is "free". In reality, the vast majority of people avoid seeking treatment until a friend has told them there's no alternative (studies have been done).

In fact, this is the main reason people die from heart attacks- they delay calling for help until it is too late. You might delay too if you thought it would cost you $1000 out of your own pocket to enter the ER and ask to be seen.

But the real question here is- where's Jason? We can hardly expect to break 100 in the comments without his troll wit. I fear the lad is ailing.

Posted by: serial catowner | Jul 25, 2007 8:42:39 AM

Megan, link please?

Posted by: Barry | Jul 25, 2007 8:54:25 AM

The idea is not that there are no treatments that work. The idea is that there is substantial iatrogenicity of healthcare. So, for example, if you are a cardiovascular patient, and you are insured, you become more likely to get statins. Great! You also become more likely to get a cardiovascular stent, which gives you the opportunity to get morbidity and mortality enhancing infections. Not so great! Total effect: arguably a wash.

That's why measuring the amount of treatment received, rather than M/M, is not necessarily a good guide. An uninsured cardiovascular patient may get all the stents, etc that they weren't getting before; but that doesn't necessarily mean they're healthier.

As for my usage of doctors, my most intensive period of doctor utilisation was when I was uninsured. I might wouldn't have had whatever they call that procedure when they stick a tube down your throat, because anaesthesia etc. are very expensive; I had that shortly after I started working at The Economist. (To be clear, I hadn't been nursing symptoms, waiting to be covered; the ulcer/cancer-like symptoms first appeared a week before my start date). But I paid a pulmonologist, etc, $400 a pop out of my own pocket. I went to see a doctor about every month or so, and it was still cheaper than buying New York State's cheapest health plan under the mandated service and community rating regulations. What I didn't do was consume extra treatments or tests. If they wanted to do something expensive, like a chest x ray or an MRI, I said "I'm uninsured, is that necessary?" and they said no, why don't we use this much cheaper thing. My health care usage during my two and a half years uninsured was a model of what conservatives are thinking when they think of consumer-directed healthcare, and not because I was trying to prove a point.

Posted by: Megan McArdle | Jul 25, 2007 8:57:44 AM

I agree with you, btw, that the rand study is not a perfect substitute for being uninsured, but such a study would be awfully hard to do . . . are you going to keep people in the control from taking jobs with insurance? But given that most of the uninsured do in fact get some healthcare, I think it's hard to know whether the marginal healthcare they got on insurance actually improves outcomes. Of any group, I'd expect that would be truest of diabetics, not CVs.

Barry, here's an abstract: http://www.rand.org/pubs/research_briefs/2006/RAND_RB9174.pdf

Posted by: Megan McArdle | Jul 25, 2007 9:20:34 AM

From http://www.rand.org/pubs/reports/2006/R3055.pdf,
page 12 as it is numbered in Adobe Acrobat
(typing, because it appears to be a scanned document):

"CONCLUSIONS
We drew three main conclusions about the influence of free care on the health statues of HIE adults. first, free care did not affect the major health habits associated with cardiovascular disease and many types of cancer, even though those habits (expectially smoking) were at levels where substantial health benefit from behavior change was possible. Second, free care had at most a small effect on any of five general heatlh measures for the average enrollee. confidence intervals were wider for subgroups of persons with low income or initially in poor health; therefore, we cannot rule out clinically meaningful changes in particular subgroups. Third, people having specific condictions with well-established diagnostic and theapeutic procedures (myopia, hypertension) benefited from free care, and these improvements appeared to be greater among the poor."

Looks a bit more mixed than Megan suggested.

I'd also point out that one randomized, controlled study does not outweigh a whole sh*tload of quasi-experiments, which is what the adoption of various health care systems by many nations is.

Other sources:
http://en.wikipedia.org/wiki/RAND_Health_Insurance_Experiment

One summary of quasi-experimental studies (not at the national level) is at:
http://www.annals.org/cgi/content/full/146/11/814;
it lists a number of studies demonstrating improved healtcare outcomes (and a few studies with null results).

These studies generally used changes in government policies as 'quasi-experiments'; some used instrumental variables.

For those who are econowhackjobs, I'd like to remind them that such methods are in general use in economics, which means that questioning them is econoheresy.

One stu

Posted by: Barry | Jul 25, 2007 9:29:04 AM

Megan: "I agree with you, btw, that the rand study is not a perfect substitute for being uninsured, but such a study would be awfully hard to do . . . are you going to keep people in the control from taking jobs with insurance? But given that most of the uninsured do in fact get some healthcare, I think it's hard to know whether the marginal healthcare they got on insurance actually improves outcomes. Of any group, I'd expect that would be truest of diabetics, not CVs.

Barry, here's an abstract: http://www.rand.org/pubs/research_briefs/2006/RAND_RB9174.pdf"

Thanks for the link, Megan.

What I find very interesting about your response is that there are whole fields of social science where randomized experiments are difficult/unethical/to expensive to do. However, these fields have found some ingenious ways to cope with this. One such field is economics. The studies mentioned in the 'annals' link use such methods.


Posted by: Barry | Jul 25, 2007 9:32:53 AM

Megan,

You were lucky. You didn't get cancer, hit by a bus, suffer a heart attack or many of the other afflictions this life offers. I have not yet hit 50 and I am astonished at the number of people I know who have had cancer. It's pretty damn hard to economize in these circumstances. I had a 4 year old with cataracts of all things, which required four surgeries, each a five figure procedure.

I have worked in a hands on way with health plans for over twenty years and I've seen decent health coverage be the difference between people having access to treatment and protection from financial disaster and I've seen others who because of the lack thereof fell into penury or simply were unable to get needed care.


As for your study, again, this is essentially pre-AIDS. What about the people with AIDS who are now being able to live relatively long and healthy lives due to the availability of drugs that are so expensive that insurance is a virtual prerequisite?

You liberterians remind me of the armchair Marxists I knew in college. All theory, very little real world experience.

Posted by: Klein's Tiny Left Nut | Jul 25, 2007 9:33:29 AM

Barry, KTLN: the Rand study is not the only study that finds no benefit from additional health care; only the best. A more recent one would be this study from the NEJM: http://content.nejm.org/cgi/content/full/354/11/1147/T2

It shows that as long as you visit a doctor at least once every few years, the difference in outcomes between having no insurance, private insurance, or government insurance is trivial.

It's pretty strange to be lectured on not having real world experience, when I was a sickish person with no health insurance for 2.5 years, and you, apparently, have never been uninsured.

Nonetheless.

Cancer is a weird case. While some cancers are near-curable, others have their survival times lengthened by days or weeks by the insanely expensive treatments that are thrown at them. So I agree I could have spent a lot of money; but depending on the cancer, it's not clear that I would have had a better outcome. Cancer is also weird because it's so expensive that many patients blow their lifetime caps, and yet the hospitals continue to treat them. And pharmaceutical companies generally have humanitarian funds for chemo for uninsured patients. So it's not clear what the actual effect would have been on me; obviously, I'd have had to declare bankruptcy, but since the odds are good that any kind of cancer a woman in her late twenties would have gotten would have been an extremely nasty, aggressive, fatal cancer, who cares? I'd be too dead to worry about getting a mortgage.

Hit by a bus, I worried about. But the State of New York, in its infinite wisdom, had decreed that I was not allowed to have high-deductible insurance, and I could not afford the monthly $450 HMO that was available to me.

The point is not that there are no treatments that work. It's more subtle than that. Cutting edge healthcare may save your life--and undoubtedly, through trial and error, creates new "standard" treatments that will save many other lives. But at any given time, having gold-plated health coverage may not improve your health.

Posted by: Megan McArdle | Jul 25, 2007 10:11:59 AM

People in their twenties get curable forms of cancer too. My best friend had testicular cancer at 27, my upstairs neighbor had thyroid cancer at 29, both very curable forms of cancer with good treatment. And good treatment and good insurance tend to go hand in hand.

Posted by: Klein's Tiny Left Nut | Jul 25, 2007 10:53:50 AM

I have always been covered because I actually planned for the worst. I planned for what I would do if I got cancer, heart attack or hit by a bus for over 5 years. Catastophic policies are cheap. Mine over that period was less than 40 bucks a month. Which is what some of those without coverage spend on cigarettes or ringtones for their 1800 minute cellphone/MP3 players in a month. Priorities in how you spend your money, what a concept! If they can't afford that then they qualify for medicaid.

Posted by: Dingo | Jul 25, 2007 11:22:15 AM

I agree Megan. Gold plated health insurance in the late 80s would have paid for a extremely expensive Bone Marrow Transplant, but all that would have done is lessen the chances of your survival compared to an insurance company that didn't pay for it. You can never equate more care with better care unless you look at the outcomes of your intervention in all fields of medicine including cancer.

Posted by: DIngo | Jul 25, 2007 11:30:15 AM

Megan: "Barry, KTLN: the Rand study is not the only study that finds no benefit from additional health care; only the best. A more recent one would be this study from the NEJM: http://content.nejm.org/cgi/content/full/354/11/1147/T2
It shows that as long as you visit a doctor at least once every few years, the difference in outcomes between having no insurance, private insurance, or government insurance is trivial.
"

thanks, Megan. Although that 'as long as'....


"It's pretty strange to be lectured on not having real world experience, when I was a sickish person with no health insurance for 2.5 years, and you, apparently, have never been uninsured."

Fascinating. An MBA from a world-class university, but no reading comprehension.

Megan, I was lecturing you on promoting _a_ randomized experiment over a large number of other studies and sources of information. And rubbing it in a bit, because you looooooooove to play the economeister game, which is a field consisting almost entirely of observational studies (please note the world 'almost').

"... and you, apparently, have never been uninsured."

Now you're just making things up. Not only have I been uninsured, but there's nothing that I've ever posted on the internet which would lead anybody to the conclusion that I've never been uninsured.

Posted by: Barry | Jul 25, 2007 11:45:53 AM

The comments to this entry are closed.