May 15, 2003
From the desk of Jane Galt:Finally!
We get our first email offering statistical evidence of decreased quality of care!
We tried to hire a Chicago Economics PhD named Joseph Doyle last year. (He went to the Sloan School at MIT). He wrote a quite convincing paper that there was a reasonably sizeable effect of your insurance status on the probability of death from motor vehicle accidents. The paper is called ""Does Health Insurance Affect Treatment Decisions & Patient Outcomes? Using Automobile Accidents as Unexpected Health Shocks" (I can't find it on the web, but it's referenced in this paper)
It's really a very clever paper, and Doyle has all the proper levels of "scientific" modesty (to mention another interest of your) in describing it. The problem, of course, is that the uninsured aren't like other folks, so it's really difficult to figure out whether worse outcomes in the health care system are due to their lack of insurance or from other stuff. Doyle's ingenious idea was to look only at motor vehicle accidents where the guy is brought in unconscious. The uninsured die more often. There are obviously hosts of objections one could make to my description of this (and believe me, we brought them all up when Doyle came to talk to us) but he carried out a number of tests to try and ferret out what caused the result. We are a group of economists who were highly skeptical of his conclusion, but this was a case where the care of analysis won most of us over.That said, the effect wasn't huge, it dealt only with mortality, not morbidity, and it really focused on trauma care versus, say cancer treatment, where Doyle admits he can't even imagine how one would study such a thing in an unbiased way. But I just wanted to let you know that there is at least one non-anecdotal rigorous study (earned him a PhD from Chicago) which does suggest that the uninsured get somewhat worse care. Keep up the good work.
That's the kind of evidence we like to see! Of course, we'd want to see more studies, and to find out how he controlled for prior health and severity of injury, but we presume that those are exactly the sort of thing the economists asked about.
See, guys, there is disconfirming evidence out there! Now let's broaden the scope, since the original question of "Who hasn't treated their children for lack of funds" seems to be too narrow. Anyone have stories about people who couldn't get cancer treatments because they lacked health insurance?
Posted by Jane Galt at May 15, 2003 10:23 AM | TrackBack | Technorati inbound linksSomething of a hasty thought here ...
>>Of course, we'd want to see more studies, and to find out how he controlled for prior health and severity of injury
There would be no need to "control" for severity of injury in such a study. Severity of injury would be randomly distributed across the sample.
A brief googling yeield:
None of these are conclusive and I haven't reviewed them in detail - the final one is the most interesting from the perspective of the effect of people putting off non-critical care until it becomes emergency.
I would expect to see similar problems with hypertension. A simple GP visit and drugs (not to mention diet and other things) can reveal high blood pressure which, if left, can lead to organ damage and myocardial infractions.
"There would be no need to "control" for severity of injury in such a study. Severity of injury would be randomly distributed across the sample."
Not necessarily. Suppose, for instance, that uninsured motorists are more likely to drive older cars not fitted with airbags. Then to the extent that airbags prevent serious injury, uninsured motorists will be more severely injured, on average.
The same argument goes for vehicle qualities other than airbags as well. Consider a Volvo wagon and a Geo Metro. Which is safer? Which do you think has a higher fraction of uninsured motorists driving them?
That rather depends on whether the sample size was large enough to provide a statistical universe, doesn't it? Also, the uninsured tend to be young, and I imagine that they tend to be driving smaller, older cars with worse safety records.
Julia Sweeney's brother had lymphoma that went undiagnosed until it was already extremely advanced and was incurable. Why? Well, he was uninsured and relied on emergency room visits for his health care. Emergency room doctors, seeing that he was not acutely injured nor HIV+, would scratch their heads, suggest that his symptoms were possibily a flu virus or maybe CFS, and send him on his way.
So yeah, that counts as a denial of service.
Isn't there also a control missing for the possibility that those who drive uninsured may not necessarily be as healthy, generally, as those who do carry insurance?
I think it's been shown that poverty is the greatest risk to life of all (John Stossel did a piece on this called "Are We Scaring Ourselves To Death"). So, for somebody who doesn't carry insurance because they can't afford it...might they already have a greater chance of dying younger, or not surviving injuries do to general poor health, etc?
I don't know, maybe its a circualr argument...I suppose one could argue that one of the reasons the poor are generally less healthy is because they have no insurance!!
Michael -
How about just saying that the poor tend to fall into two categories: young people newly on their own (who might well be expected to be less risk-averse than their elders) and chronic losers in the game of life. The second category consists largely of people that have never learned to plan ahead in life. They'd be less likely to wear seatbelts, as well as the factors Jane mentioned above.
Either way, I'd love to see how he tried to control for these effects. Trauma, in particular, seems to be the least likely place for there to be any kind of difference in health care.
>>Not necessarily. Suppose, for instance, that uninsured motorists are more likely to drive older cars not fitted with airbags. Then to the extent that airbags prevent serious injury, uninsured motorists will be more severely injured, on average.
Hmmm ... maybe you have a point. But the people had to be severely injured enough to be carried unconscious into hospital in the first place. This still looks like grasping to me.
Not at all -- there's a big difference between someone with a concussion, and someone whose spine was crushed.
I think the more salient question is not the difference in care within our system but the difference in care between our system and an alternate system. What might the difference in care be between, say, our uninsured auto-accident victims and auto-accident victims in Canada? And what is the difference in care for the average person (insured or not) between the two systems? Even if our system is not 100% perfect, a cure for that imperfection might be much worse than the disease.
Forget the Canadian system. Forget the British system. Forget the Micronesian, German, Chilean or Zimbabwean systems. The question, as rhetorically pleasing as it may be to point out the shortcomings in system X, is whether our system can be improved.
If one wants to compare US accident trauma treatment to the Canadian system, one may well be able to make the argument that our system is better. So what ? So we won't model our accident trauma system after the Canadian system. Who's not to say that the Dutch system or the Ecuadorean system isn't better. May be, may not be. But if the whole point of the exercise is to figure out whether our system can be improved, we won't get there by repeating over and over that our system is better than system X,Y or Z.
Or is not getting there the point ?
Cheers,
So Rofe, you think we should compare the US healthcare system to some hypothetical ideal that exists nowhere in the world?
I have a question. You would agree, I assume, that the British and the Canadians aren't stupid. They already have socialized systems, therefore they can change their systems to improve outcomes without those pesky HMO's and greedy private physicians getting involved. Can you explain to me why they haven't modified their systems to conform to your hypothetical ideal?
Jim
>>You would agree, I assume, that the British and the Canadians aren't stupid. They already have socialized systems, therefore they can change their systems to improve outcomes without those pesky HMO's and greedy private physicians getting involved. Can you explain to me why they haven't modified their systems to conform to your hypothetical ideal?
The proportion of private medicine in the UK has been on an upward trend for years and years. Canadians I can't speak for, other than to observe that the ones I've met have a bit of a tendency to whine.
Jim,
Here's my peeve. Whenever the issue of health care reform comes up, it's not too long before the stories come tumbling out about waiting 6 months for a heart exam in Canada or the UK. Those stories may well be true. But what do they prove ? That conservatives can pick out the softest targets in the health care world and use them to 'prove' that any reform in the US dooms us to the UK's NHS ?
Why don't I ever hear anyone talking about how crappy the Dutch system is ? Why doesn't anyone rant about how crummy the Norwegian system is ? Are those systems pitiful as well and I'm just not listening closely enough ? Or, maybe, those systems aren't so bad and therefore not such good 'evidence' ?
The only system I can talk about outside of the US is the German system, a mix of private and socialized care. It seems to work pretty well, though there's plenty of discussion here about costs that are too high, the need for reform, etc. Those topics may sound familiar. Is the German system better than the US system ? Seems to me it has some advantages, some disadvantages. But that goes back to my original point. Other than to the extent that, if we were serious about reform in the US, we might pick up on things the Germans do better, who cares ? Whether the German system is better or worse, we have to deal with the US system.
I think there's an old saw that many of us would subscribe to, "If it ain't broke, don't fix it." Why isn't the discussion about what things are broke in the US system (I think there are some), and if they are broke, how do we improve them ?
Despite your interpretation of my comments, I don't think there's any magical system out there that's perfect, nor is perfect the goal I'd aim for. But why does the discussion (largely from the right) immediately hit on the soft targets ? Seems to me that's avoiding the issue.
It's late; I have to run. Apologies from rambling.
Cheers,
hi robin goodfellow and rofe,
i can't help but think that you are arguing at cross purposes. robin g., as i understand it, is offering a counter-factual argument. if we changed our system, would we make things better or worse? r.g. is using canada and the uk as examples of counter-factual alternatives that we would need to look at. i think that is valid. on the other hand, i think that rofe makes the good point that this should not stop us from examining the system we have and doing things to improve it, and if needs be taking ideas up from socialized medicine countries that could help us improve this outcome.
Rofe, I think the reason you see opposition to health care change in the US come mainly from the right is because most suggestions for change tend towards increased government involvement and socialization. In Canada it's entirely the opposite. Change typically means increased privatization and the majority of the opposition comes from the left.
dsquared, there is very little private medicine allowed in Canada. Actually, that may make us a bad example - whatever form nationalized medicine might take in the US, it's unlikely to go as far as Canada has gone. And we do not whine. Stop picking on us.
The question is not whether any system can be improved, perfection is impossible therefore some improvement is possible for all systems. The question is: improved HOW? In what direction should we take our health system, what changes should we make to improve it? If we intend to make our system like some other country's system then we need to see whether that would truly be beneficial or harmful. If we intend to change our system in new, dramatic, untested ways then I think we need to be damned careful how much and how fast we change it. And whether or not we go too fast to determine if the changes are actually helping or hurting before we have a chance to change back instead of being forced to stay with a new system that's worse. Some people, a lot of people, propose to change our entire health care system by government decree in a very short span of time to some sort of socialized health care system. That, I think, is unwise based on the evidence. Just about every other type of potential improvement to our health care system falls under changes that will naturally occur in a competitive system. That's a pretty strong case, in my view, for largely leaving the system we have alone.
Another variable to control for would be which hospital the unconcious accident victim was carried in to. It seems likely that accident victims without health insurance might tend to live near hospitals with poorer records.
It would also be interesting to know at what point the hospital discovered the accident victim was uninsured. Do victims whose insurance status is unknown also die at a higher rate?
Some years ago, one of the insurance companies doing business in Michigan discovered that one-third of the people at fault in accidents with the company's insureds were not carrying auto insurance. This is a serious issue in Michigan where it is illegal to drive without insurance.
Although there are some uninsureds, they do not amount to a fraction of one-third of the drivers.
Point is that irresponsibility in one area of life might mean irresponsibility in others. Including driving like a moron.
How does that affect the results in the study cited originally?
Among other things, I am concerned about how long it is from reaching a hospital while unconscious to death. Does it take long enough for anybody involved in the care to find out the insurance status?
Does the admitting office do a quick study to find out what the guy's insurance is and then run up to the ER?
What, exactly, is the mechanism connecting the discovery of the insurance status and the level of care?
If there isn't any, then the result must be a matter of something else.
Improving our helath care system and improving health insurance coverage are not the same thing. One can argue that governmental interference in health care is the root cause of the problems in health care, not the solution. As an example, the setting of rates for services. A BLS transport by ambulance to the hospital can bill Medicare about $90. Your insurance will be billed about $200 for the same ride. That suggests, on the face of it, that the market price is somewhere in between.
Or, take the government mandating the number of hospital beds in an area. Instead of letting the market decide that hospital A has 100 beds too many, the government mandates that hospitals A-D all reduce their number of beds by 25. This may, in fact, put a small hospital out of business because a certain number of beds are necessary from a business perspective just to keep the doors open.
In many specialties, the primary patient is elderly, and on Medicare. By controlling the physician's ability to charge for his services, and not affecting his costs, the government forces the physician to increase volume and decrease unfunded services (like time spent with the patient).
Prescriptions are cheaper in Canada because the government sets the price. Again, the true market price is somewhere between the Canadian price, and that paid by the rest of us, because the difference has to be made up somehow.
A paid EMT locally makes $8-10 per hour and a paramedic $9-12. In other words, the guy saving grandpa probably makes less than the girl at the drive through window at McDonalds. McDonalds can pay more because they can raise their prices.
The paper that references the study merely cites a "statistically significant effect" (otherwise known as a "reasonably sizeable effect")... I hunger for numbers! :) What kinda sizeable we talkin' about here?
Frankenstein - freely choosing to go to the ER for care, resulting in a missed diagnosis, is not a "denial of service." No one turned this person away from appropriate care. Failing to have insurance is not the same thing as being barred from care. There is plenty of charity care available in the US from nonprofit hospitals and health systems. There is also Medicaid and dozens of other government-paid programs for the truly indigent.
It is a violation of federal law for any hospital to do a "wallet biopsy" on an emergency patient, and further violation for the hospital to refuse to take an emergency patient or to steer them to another hospital because they don't have insurance. I can assure you there is no variation in emergency care based on insurance status (unless the patient refuses care because they think they can't afford it, which is not, I repeat, a denial of service). If the uninsured people in the study are dying at a greater clip, it is either because of post-emergency care or some other variable that has nothing to do with health care.
>>statistically significant effect" (otherwise known as a "reasonably sizeable effect")
Not necessarily, of course. "Statistically significant" is really just a measure of how many data points you have. The tiniest imaginable difference could be statistically significant if you had a really big sample.
Assuming that poor peopled end up in lower quality hospitals is not necessarily a safe assumption. One of the best trauma ERs in the world is Ben Taub, a public hospital in Houston. if you get shot or seriously injurned in an automobile accident, that's the place to go. Of course, if you can afford it, you will probably want to move to a hospital with more pleasant facilities after you are stable...
Bolie IV
I've been a physician since 1974, and I still don't understand the US health care "system." I don't think there is one. The study of injured patients' mortality doesn't seem very relevant to the risks and benefits of insurance coverage. One problem is that most people don't need any medical care whatsoever for most of their lives. But when they need it, they really need it. Another problem is that there is no real "market" for health care in any widely understood sense. As an exercise, call some of your local hospitals and clinical labs and ask for their full retail price on the set of tests summarized at http://www.directlabs.com/cwp.php
I have done this. The common response is astonishment, followed by "we'll get back to you on this." Some hospitals never do. Others will figure out an answer in 4-5 days, but none of the 3 that did answer actually gave me a firm amount in reply, just a very vague (and very high) estimate.
Or call a surgeon, and ask for his and his hospital's total fee for an uncomplicated indirect inguinal hernia in a male. You will likely get no usable response. This has happened to my brother. Paying for many health care services is like going to a pharmacy with a prescription, signing a legally binding pledge to pay whatever the pharmacy decides to bill you for, and not knowing ahead of time what the amount will be. Maybe I should become an economist...
>>Maybe I should become an economist...
If you're not joking, drop me a line, as I can put you in touch with a number of people who might be interested in co-authoring a paper on that sort of data.
Comments are Closed.