counter create hit Free for All?: Lessons from the Rand Health Insurance Experiment - Download Free eBook
Ads Banner
Hot Best Seller

Free for All?: Lessons from the Rand Health Insurance Experiment

Availability: Ready to download

In the most important health insurance study ever conducted researchers at the RAND Corporation devised all experiment to address two key questions in health care financing: how much more medical care will people use if it is provided free of charge, and what are the consequences for their health? For three- or five-year periods the experiment measured both use and health In the most important health insurance study ever conducted researchers at the RAND Corporation devised all experiment to address two key questions in health care financing: how much more medical care will people use if it is provided free of charge, and what are the consequences for their health? For three- or five-year periods the experiment measured both use and health outcomes in populations carefully selected to be representative of both urban and rural regions throughout the United States. Participants were enrolled in a range of insurance plans requiring different levels of copayment for medical care, from zero to 95 percent. The researchers found that in plans that reimbursed a higher proportion of the bill, patients used substantially more services - indeed, those who paid nothing used 40 percent more services than those required to pay a high deductible - but the effect on the health of the average person was negligible. In addition, participants who were assigned at random to a well-established health maintenance organization used hospitals substantially less than those in the fee-for-service system, again with no measurable effect on the health of the average person. This book collects in one place for the first time results previously dispersed through many journals over many years. Drawing comprehensive, coherent conclusions from an immense amount of data, it is destined to be a classic work serving as an invaluable reference for all those concerned with health care policy - health service researchers, policymakers in both the public and the private sectors, and students.


Compare
Ads Banner

In the most important health insurance study ever conducted researchers at the RAND Corporation devised all experiment to address two key questions in health care financing: how much more medical care will people use if it is provided free of charge, and what are the consequences for their health? For three- or five-year periods the experiment measured both use and health In the most important health insurance study ever conducted researchers at the RAND Corporation devised all experiment to address two key questions in health care financing: how much more medical care will people use if it is provided free of charge, and what are the consequences for their health? For three- or five-year periods the experiment measured both use and health outcomes in populations carefully selected to be representative of both urban and rural regions throughout the United States. Participants were enrolled in a range of insurance plans requiring different levels of copayment for medical care, from zero to 95 percent. The researchers found that in plans that reimbursed a higher proportion of the bill, patients used substantially more services - indeed, those who paid nothing used 40 percent more services than those required to pay a high deductible - but the effect on the health of the average person was negligible. In addition, participants who were assigned at random to a well-established health maintenance organization used hospitals substantially less than those in the fee-for-service system, again with no measurable effect on the health of the average person. This book collects in one place for the first time results previously dispersed through many journals over many years. Drawing comprehensive, coherent conclusions from an immense amount of data, it is destined to be a classic work serving as an invaluable reference for all those concerned with health care policy - health service researchers, policymakers in both the public and the private sectors, and students.

30 review for Free for All?: Lessons from the Rand Health Insurance Experiment

  1. 5 out of 5

    Jeff Cliff

    A borderline must-read for those of us who either being born into Canada/NDP households or through otherwise cultural reasons are predisposed to say "privatized health care is a fucking stupid idea that will hurt access to healthcare". This was, for its time, one of the largest and most expensive scientific investigations in history. It's an experiment that at least as of 2007 was the best and biggest experiment of its kind that was done with an eye for actually collecting data on what the A borderline must-read for those of us who either being born into Canada/NDP households or through otherwise cultural reasons are predisposed to say "privatized health care is a fucking stupid idea that will hurt access to healthcare". This was, for its time, one of the largest and most expensive scientific investigations in history. It's an experiment that at least as of 2007 was the best and biggest experiment of its kind that was done with an eye for actually collecting data on what the impacts of free/cost-shared healthcare actually was. ACA and data from healthcare systems around the world combined with *much* cheaper dataprocessing than the 1970s have probably allowed for the testing of its hypotheses on a wider scale but it's not clear how much testing and learning has been going on, and of that how much of the data is as directly comparable as the RAND data was. Those of us into funding more science and taking science seriously as a source of good ideas and of testing bullshit claims should actually pay attention to its data and to what it implies insofar as changes in our own worldview may be required. Especially as healthcare gets eaten more and more by software, and as the political situation of 'who gets to decide how healthcare can work' gets decided at the global level. After all, if we do *not* use science in healthcare, we risk letting it go to the charlatans, and likewise risk pushing those smart enough and concerned with health crowd *into* peddling woo. The overview given by Robin Hanson is more or less accurate, and for those not interested in going through a book, his blog posts will cause you to get the gist of it. There's much in here to soak up into your mnemosyne deck/memorize if you are not in the healthcare field, hints for what the 'normal' is for healthy individuals before the obesity epidemic hit in large. Likewise, it would be interesting to look at, say, Fitbit's dataset through the lens of their numbers -- it'd be interesting for someone with access to their data to publish a book like Dataclysm that looks at the RANDHIE data to see if the normal, healthy american standard that they used either was a) incorrect in how normal/healthy it was, in practice b) whether we have, by and large, grown more unhealthy over time. Somewhat harrowing some of this was data on death. tl; dr don't smoke & get your blood pressure checked at least once. The RAND study was limited to how pre-ACA, pre-peercover, pre-AI, health insurance systems can impact citizens who use it, but reading this book kind of made me rethink my career choice. Especially in light of some of the data at the end of the book(that it wasn't until the 1920s that there was a more than 1:1 chance of leaving a meeting with a doctor healthier than when you showed up), one natural question is 'should we subsidize more nurses/doctors instead of healthcare itself', which in and of itself has interesting boundary properties as question, after all, becoming a doctor takes a lot of work and it's unclear how many doctors we could have even if we tried to maximize such a thing. But this kind of question starts to be more important, once you get into the nitty gritty details of 'how to make people healthier' since, as we'll see below, some of the answers don't turn out to be as simple as we'd like. But there is an intrinsic problem with the 'unknown unknowns'(p174-180) - we don't know what doctors can help with, we don't know whether what we are dealing with is something we should go to the ER for(even sometimes nurses screw this up, I can attest), and sometimes we don't know how to hook up with a nurse who can do the work instead of a doctor. We don't know what to google. Part of the problem of not solving the last mile of these unknowns is a bigger pressure on ERs. Solving that problem might end up doing a LOT more to promote health outcomes than the 'free healthcare' part for the average person, it turns out. And that problem can be solved at least in part in software. As a work of big *science* there was some interesting details(~p233). It reminds me very much of the attempts of at least one of my employers to create company-wide idea generation system. As they were doing experiments on live/working human beings who had to do their job whether or not the research was going on, there was some conflict between 'research' and 'action' "teams" That these conflicts and other problems would come up as easy to describe hard to resolve problems - the earlier you get action & research teams involved the better. Other noteworthy * The action team has less stake in the success of experiment * It's good idea to foster and cultivate mutual respect between different teams. * Don't suppress disagreement * Frequent (weekly) project progress reports/communication. Have hierarchy & have heads of hierarchy be in communication with eachother. But stay agile, elicit & use feedback, make sure there's a decent division of labour & plenty of lateral communication. This sort of thing is *way* easier to do post-internet. * managers can help by making teams(ie, so that people don't just stick with who they know best/are most comfortable with). * routine - as opposed to ad hoc--mechanisms must be designed for resolving differences. Such mechanisms broaden the perspectives of both parties & consistent action diffuses conflict and tensions before they get out out hand. Their approach to research, whether effectively limited by computational constraints or not, was that of a Linear waterfall. More about that when we get to ScienceMart. "Skills" as something one has that allows for an experiment to occur, there was something in their description of skills that sounded so...arbitrary and alien. One unfortunate aspect of real big science in the US is...lawyers. When we start talking about healthcare, unfortunately we're talking about legal liability. If the US starts having a healthcare crisis it's undoubtably also going to be a crisis of legal liability. How much this would be true in Canada is unclear to me...but it'd be something to look into should we ever have a crisis here. (Oh wait. We are having a crisis here! Hmm) The RAND experiment also looked into non-physical health and had some data on how meaningful these other alternative conceptions of health actually are, and how to think about health in general. Mental health behaved *a lot* like other kinds of health, for example in its stability over time/treatment(p=.6, p193). Social health is less valid as health than mental health but still clusters around truth. Social contacts have a causal effect on general health. One of the things that was measured was "disability days". As someone who suffers from allergies and other minor health issues periodically it sure sounds like a day without internet is starting to bleed into 'health', and in general network connectivity is starting to have health-like properties (and more with time). Should the RANDHIE ever get replicated I would wonder if it would make sense to take this into account. So now here's the shocking parts. For the average person there were (with minor exceptions) no benefits from free care. Exceptions including some dentistry, blood pressure and glasses(dentistry and glasses aren't even covered in Canada!). There were segments of the population that free healthcare did help, such as the poor and the chronically sick - but according to their data, an *effective* system like the US Medicaid would actually conceivably cover them. This is not to say that socialized healthcare is a bad thing, since there's also the 'back end' to be concerned about - canada pays less than the US does for the same drugs due to its single-payer system, for example. But it certainly changes the framing of debate a little, when we aren't as concerned about whether not people would be just as healthy, deal with about as much pain(p20*), have roughly the same treatment at their own dime or have the government pay for it though, or become healthier since it's roughly the same in both cases. I know, this is kind of shocking, but their data is good. There were plenty of limitations of their data, though, in that there may be situations/contexts that it may not apply. Access in rural areas. There were only 6 areas that were tested -- and though they were varied, I don't think they had anything quite like Rural Saskatchewan represented. The case for a CCF-style free healthcare system in rural saskatchewan might be quite strong relative to their data. Their data does not seem to deny this. Of course, after the PCs took a wrecking ball to saskatchewan the NDP wound up unwinding a significant portion of healthcare that was out in rural saskatchewan anyway so with even the NDP ceding that, at least as long as we keep voting for stuff like NAFTA at the federal level, and as long as we keep electing PCs provincially...we will probably not have free rural healthcare for very long, at very high standards so it's kind of a moot point. There was some discussion of mental health, and adverse selection of employers willing to offer healthcare coverage for mental health. I wonder how much this has changed, as awareness and acceptance of mental health has changed over the decades. There does seem, at least in the US context, to be a public policy issue in that not enough employers were offering mental health coverage as part of their health coverage, and that this wasn't intrinsically due to the expense of helping people afflicted by mental health problems be healthy, it wasn't intrinsically due to the lack of 'mental health' being a meaningful 'health'-like property worth treating seriously, but rather due to game theoretic concerns. Gender and families were touched on, and they had some data on women...but it seemed like they weren't really all that concerned with splitting up their data into impact on men and women. Was this because there wasn't much difference or was it because they didn't look? Without replicating their experiment it's kind of hard to tell. Interesting to note, though, is that in the discussion of 'how to pay for healthcare' they predicted a greater tax burden on families...this in turn predicts a higher divorce rate if the burden gets high. Oh look what happened since? Drugs were touched on near the end, and there wasn't really much of a difference in drug use, including but not limited to opioids between free & non-free plans. People figure ways of getting their drugs either way. The idea of putting cost barriers to deter people at a moral level from using pain killers to alleviate pain turns out to be just as ridiculous as the idea that free opioids will turn everyone into addicts, at least according to their data, at least in the near future/decades. Generally people treat drugs as a means of being healthy or dealing with pain. There are some exceptions but they are exceptions, not the rule. There were two classes of care found, with examples, of things that could be provided for free, that there would be (in some cases substantial, like dentistry) benefits to those who were for whatever reason not willing to pay for the care themselves. The two classes of these were those with clear benefit above and beyond the cost of providing it (eyeglasses, esp to the poor) and those which it was clear there was a benefit but it was unclear the cost was worth the benefit(hypertension). For those interested in following this line of thought, there's this thread. There is some danger in looking at one part of a complex system, such as healthcare, and trying to pry apart what will happen if you start changing aspects of it. It is worth considering that they acknowledge this but at the same time the way to deal with complex and weird systems is not necessarily to merely let them run and hope they never break. Experiments like the RAND health insurance experiment help us understand both ourselves sub-parts of the bigger picture that might be in need of closer attention(*cough* free glasses for every man, woman and child in the world who need them). Though left for others to debate, there is kind of the question of *how* you might use their results to shape public policy. In the case of the current US, it's entirely thinkeable that ACA was compromised to death, even if it was well intentioned and would have worked as a single-payer system or without handouts to the insurance industry and similarly any attempt at RANDHIE style reforms might similarly be doomed to death by a thousand cuts to stuff pork into. Similarly, how do you get average people to pay for *anything* for the poor, who actually do need the extra help to be healthy/not die? It might very well mean that, though it might in principle be more effective, healthy, efficient dollar wise to have more private care...supporting whiny average people and unsupportable trolls might be a *cost* of supporting the poor and the chronically sick. Also, there may be something to be said for the value of living in a society, like canada, where healthcare is simply taken for granted as part of our social fabric. But at that point we start getting into faith in healthcare, not science. And that starts to be dangerous territory for rationalists. Take a look at this study, for example: "Second, we obtained important experimental evidence supporting our prediction that the more an external system of control is deemed benevolent—that is, perceived to hold the goal of serving a system’s constituents to the best of its ability—the more likely it will be relied on during times of low perceived personal control. Indeed, those participants who did not a priori deem the Canadian sociopolitical system as representative of their interests did not increasingly defend it following the personal control manipulation. This suggests a key, theoretically sensible moderator. We would to remind the reader, however, that we are not suggesting that perceived system benevolence is the only moderator of this phenomenon. Other factors, such as perceptions regarding the system’s ability to competently and effectively meet its goals, are also likely to play a moderating role." This reminds me of Meredith Whitney's view of goodwill inherent in the banking, financial and real estate sectors in the US prior to the great recession - that goodwill can go away in a hurry. Speaking of goodwill going away in a hurry, on the longer term there was some worry expressed that it's unclear how sustainable the medical insurance businesses in the US were. This is probably doubly true post-ACA. Sure they are making a lot of profit, now...but in the long run it's not clear that they have to be. Also, it's not clear that they would serve the average citizen, they might instead ratchet up to only serving the wealthy and powerful and deny coverage to those who realistically should deserve it(say, firefighters who rescue people during a terrorist attack or something). Maybe they will, but it's not *guaranteed* that they will. There are market failures. And *those* failures are very much the sort of thing that inspire people into free care systems. Insurance companies should be very aware of this, and publicly go the extra mile to avoid even the perception that they might not cover people. But there is a broader problem, as mentioned above -- there's no guarantee that we have our incentive & other systems set up to produce enough doctors and nurses. It's entirely plausible they could collapse, leaving us without doctors, too, if we get our incentives wrong. Or maybe our doctors/nurses would fly away and go to the US as they have a tendency to do post-NAFTA, and tend to do much more frequently in other countries(like the Philippines). There are fundamental contradictions inherent in capitalism that this study couldn't possibly approach with its data - in the long term they will have to be addressed because it is likely that all of these problems (some back-end others not) will manifest, but in the short term it's probably a pretty good approximation of how things could work under moderate changes. Another limitation is that their data was on a *significantly younger* cohort than the one living today. This might call into question the validity of their findings on boomers as they age. When I was a university student reading Marx I realized that north america and the world more generally was heading towards political change. That the boomers, in their youth, were interested and capable of significant autonomy as members of a large cohort and that the political system that surrounded them, much of which was designed to restrict the animal nature of man restrained them significantly. When the boomers built the post-war health system some of the excesses of paternalistic views on what the state's responsibility actually was was tempered by a desire, as a generation, to throw of the chains that had been built for them, from the mental asylums to other structures of state power. However in the 2010s this same generation is now old and becoming dependent upon the healthcare system they built to even survive. Old folks homes are filling to the brim, and we are entering an age that the government(who owns the healthcare system in canada) has absolute authority over an increasing % of the population. Likewise in the US, the structures that control the healthcare system are gaining the same kind of political power. In either case, the healthcare system can become a means of justifying as absolute of state power as have ever existed. It may become necessary, within our lifetimes, to dismantle it. But once we do this, we may need to remember what was the most important/least important parts if it's to be recreated.

  2. 4 out of 5

    Olivia

    a giant research paper describing the objectives, design, results, and implications of the ~20-year long experiment - lots of citable material, i.e. increased cost sharing decreases use without adverse health effects on average persons but with adverse health effects on the poor

  3. 5 out of 5

    Andrew

  4. 4 out of 5

    Robin Hanson

  5. 4 out of 5

    Paul

  6. 4 out of 5

    Ben

  7. 5 out of 5

    Mary

  8. 5 out of 5

    ℕℰЅЅᎯ

  9. 5 out of 5

    Jason Furman

  10. 4 out of 5

    Gwern

  11. 4 out of 5

    Eugene Joseph

  12. 4 out of 5

    Homoionym

  13. 5 out of 5

    Mickayla

  14. 4 out of 5

    Sara Ann Joehnk

  15. 5 out of 5

    Joseph Garner

  16. 5 out of 5

    Susan B

  17. 4 out of 5

    David Silva

  18. 4 out of 5

    Fred Falk

  19. 4 out of 5

    Sean

  20. 4 out of 5

    Feng Lin

  21. 4 out of 5

    Michele Thornton

  22. 4 out of 5

    Diane

  23. 4 out of 5

    Lauren

  24. 4 out of 5

    Caitlyn Mcneil

  25. 4 out of 5

    BookDB

  26. 4 out of 5

    Pablo Stafforini

  27. 4 out of 5

    Tony Boyles

  28. 5 out of 5

    Rainier Moreno-Lacalle

  29. 5 out of 5

    Rachel

  30. 4 out of 5

    Sudip

Add a review

Your email address will not be published. Required fields are marked *

Loading...
We use cookies to give you the best online experience. By using our website you agree to our use of cookies in accordance with our cookie policy.