• I started to answer individual comments, but they deserve their own post. So start by reading our piece yesterday on the new NEJM study. On to your questions/comments/shrieks:

1) This was a HUGE study. How can you say otherwise?

According to the paper, 12, 229 people responded to the surveys and were analyzed. So, yes, for outcomes that effected everyone (think financial hardship), it’s likely they were super-powered. But for many of the outcomes that were more murky, that’s not the case. Take A1C for instance. Only 5.1% of the control group had an A1C>=6.5. Let’s assume that the starting prevalence was the same in the intervention group. That means that only 624 people (312 in each group) actually had a high A1C in the study. That’s not anywhere near as big. Especially when you’re talking about an indirect intervention like insurance as opposed to actual health care.

2) You can’t do a power calculation after the fact!!!

I’m not asking for a post hoc power calculation. I want the a priori one. You see, with only 600 or so participants with an A1C in the high range, I want to know what they were thinking ahead of time.

If my study is too small, then even if I see a difference that I think is meaningful, I might not be able to prove that it is statistically significant. So when I’m designing a study, I decide what is a clinically meaningful result. I then figure out what I can likely expect in terms of variability in the individual readings I might measure. Then I figure out how many subjects I need in order to know that if I get the clinical results I expect, they will be “detectable” by my analysis. That’s the calculation. If my sample is too small, then even if I find a clinically meaningful result, it might not be statistically significant.

3) You don’t understand statistical significance!!!

I assure you I do. When your point estimate is clinically meaningful but your results are not statistically significant it usually means that the variability was larger than expected, there really was no effect, or you were underpowered to detect the difference. See (2). I can’t tell which of these are true because I don’t know if the study was powered to detect the point estimate differences they found.

(I should add here that some are upset by the fact that our post said p=0.07 is close to significant. I (Aaron) am more of a purist when I’m using frequentist statistics, so I would agree and not say that. Austin is more of a Bayesian and doesn’t think that’s quite as blasphemous. But I recognize this is a Shibboleth for people who think they truly understand statistics, so I’m acknowledging it.)

4) Obamacare promised us it would save tens of thousands of lives a year!!! He lied.

Stop. This was Medicaid for something like 10,000 people in Oregon. The ACA was supposed to be a Medicaid expansion for 16,000,000 across the country. If 8 people’s lives in the study were saved in some way by the coverage, the total statistic holds. No one measured that. This is silly.

5) You’re using financial hardship and other stuff as a smokescreen.

No, I’m not. The reason I have insurance, and likely you do as well, is to protect me and my family from financial ruin. When I get sick, I don’t sit at home and let the insurance take care of me. I get off my butt and use the health insurance as the means by which to get health care. Medicaid is about access. It’s just the first step in the chain of events that leads to quality.

That said, I still maintain that we have never subjected Medicare or private insurance to this standard. Just Medicaid?

6) The results were bad anyway. Blood pressure moved a point down? That’s nothing!

Average blood pressure was a bizarre thing to measure. You have to remember that most people who get health insurance are healthy. They’re not going to get “healthier”. The average blood pressure in the control group was 119/76. That’s normal! You would only expect that it might improve in those with a high blood pressure. So I might have looked for an effect in those patients with hypertension. 16.3% of people in the control group had a systolic over 140 or a diastolic over 90. In the Medicaid group, that dropped to 15%. If they wanted to look at average pressures, why didn’t they single out the hypertensive people? I don’t know.

I’ll update this as more things occur to me.

@aaronecarroll

item.php
• I can tell you why it isn’t appropriate for the poor

. “Financial protection is the reason most Americans who can afford health insurance buy it. If that rationale is good enough for everybody else, why isn’t it good enough for the poor?”

I have to pay for my own, they should pay for their own. If I can’t afford something, I don’t buy it and I can’t afford to pay for my own health insurance and someone elses too. I don’t want to pay for someone elses health insurance. Simple statement.

If Aaron Carroll and Austin Frakt want to pay for other people’s insurance, I am fine with that but don’t pose it as a federal tax to take my hard earned money for your misguided charity.

• Also, it doesn’t state how statistically significant the mental health benefits are or even the financial benefits. These effects may take years to quantify.

I am 100% sure that if the same dollars “extorted” from the American public to pay for Medicaid we used for strictly mental health purposes and to fund training programs for the poor, you would not only get better value but you would get better mental health and financial outcomes as well as stimulate the economy with skilled labor.

• You pay for their health care when they get sick either through:
taxes funding government support for unfunded care in hospitals payments for unfunded care in hospitals or

Higher hospital charges/insurance to cover higher hospitals bills because hospitals have to cover those losses or

taxes funding insurance subsidies.

And you’re going to pay for the people who can’t afford to pay for necessary health care unless we let them die in the gutter of an easily cured disease, like appendicitis.

• Or

You pay for their health care by living in a society with a lower GDP because it does not have a healthy and able workforce. There is a reason why developed nations offer universal care and while some of it is an ethical decision it is also an intelligent decision economically. It is no surprise that the largest and most powerful economies are the nations with the healthiest people (and please before you mention China their economy is so big because they have so many people not because they have a high GDP).

• “Are there no prisons?” asked Scrooge.

“Plenty of prisons,” said the gentleman, laying down the pen again.

“And the Union workhouses?” demanded Scrooge. “Are they still in operation?”

“They are. Still,” returned the gentleman, “ I wish I could say they were not.”

“The Treadmill and the Poor Law are in full vigour, then?” said Scrooge.

“Both very busy, sir.”

“Oh! I was afraid, from what you said at first, that something had occurred to stop them in their useful course,” said Scrooge. “I’m very glad to hear it.”

“Under the impression that they scarcely furnish Christian cheer of mind or body to the multitude,” returned the gentleman, “a few of us are endeavouring to raise a fund to buy the Poor some meat and drink, and means of warmth. We choose this time, because it is a time, of all others, when Want is keenly felt, and Abundance rejoices. What shall I put you down for?”

“Nothing!” Scrooge replied.

“You wish to be anonymous?”

“I wish to be left alone,” said Scrooge. “Since you ask me what I wish, gentlemen, that is my answer. I don’t make merry myself at Christmas and I can’t afford to make idle people merry. I help to support the establishments I have mentioned: they cost enough: and those who are badly off must go there.”

“Many can’t go there; and many would rather die.”

“If they would rather die,” said Scrooge, “they had better do it, and decrease the surplus population.”

Good ole’ Ebenezer, providing the foundation for Libertarian public policy since 1843.

• Fine, then stop driving on the roads that I paid for. Go get your own damn police. And quit calling my fire department.

• The fact that you pay for your own health insurance is part of the reason it costs so much. The rest of the world has far lower administrative costs that we do. I’ve seen it estimated that between two and three million people are employed shifting medical costs to someone else. This is individually rational and socially insane. A national system of some kind would eliminate it.

• Ro:

“I have to pay for my own”
Actually, insurance has the healthier subsidizing the sicker, so either someone is paying for you or you’re already paying for others (unlikely you’re right on break even). Ditto for roads and other government services, and rich states supporting poor states.

• Without help from Cannon and other critics of Obamacare, any objective person who read the report would interpret its conclusion as an endorsement of Medicaid expansion. What’s confusing to most people is this statement in the report: “Medicaid coverage generated no significant improvements in measured physical health outcomes in the first 2 years”. What are the “measured” physical outcomes? “Measures included blood-pressure, cholesterol, and glycated hemoglobin levels; screening for depression; medication inventories; and self-reported diagnoses, health status, health care utilization, and out-of-pocket spending for such services.” That’s it. It’s understandable that the authors used these limited measures – they are easily identified and recorded – but I don’t believe it possible to make a broad generalization about Medicaid expansion from them. Yet, even with such limited “measures”, the authors still concluded that “Medicaid coverage significantly increased the probability of a diagnosis of diabetes and the use of diabetes medication, but we observed no significant effect on average glycated hemoglobin levels or on the percentage of participants with levels of 6.5% or higher. Medicaid coverage decreased the probability of a positive screening for depression (−9.15 percentage points; 95% confidence interval, −16.70 to −1.60; P=0.02), increased the use of many preventive services, and nearly eliminated catastrophic out-of-pocket medical expenditures.”

• Two thoughts:

First, on your point #1 on sample sizes, note that the study is effectively even smaller than you’re saying. Thinking of the high A1C sub-sample as containing 624 people is correct for the intent-to-treat analysis (i.e. for estimating the effect of offering insurance).

But (net) take-up was only about 25 percent. Thus, for the IV analysis (which provide the estimates of the effect of insurance per se), the point estimates and variance are scaled up by a factor of 4=1/0.25. This means that the “RCT-equivalent” sample size is only 156=624/4. That’s really, really small, even before you get to the fact that insuring people does not automatically get them the appropriate care (your point about direct versus indirect interventions).

Second, in thinking about costs and benefits here, it’s really important to be clear about the fact that the (social) costs of the expansion is the incremental care consumed by these individuals. So the key question is whether that additional care consumed (additional prescriptions, office visits, preventative screenings, and maybe a tiny bit of additional inpatient care) is “worth it” in terms of the health benefits.

Trying to decide the “is it worth it” question on the basis of the Oregon study is crazy when we have a clinical literature actually powered to quantify the benefits of particular medical interventions. For the increases in utilization observed, I think the general message of the clinical literature would be that “yes” that care is worth it.

The only thing that would make appealing to the clinical literature suspect is if Oregon was giving clear evidence that the health benefits are not what we would have expected given the change in utilization. But for all the cases I’ve back-of-the-enveloped, the health responses are right in the range you’d expect based on the utilization changes and our prior estimates of treatment efficacy. They’re just insignificant because of the power issues.

Bottom line: Oregon tells us that Medicaid gets people more recommended care. It tells us (virtually) nothing about the efficacy of that care, so if we thought that recommended care was a good thing yesterday, we still should today. Properly interpreted, therefore, it’s a big win for Medicaid.

• I agree, but I was trying to be conservative and offer the intention to treat numbers. With insurance as the intervention, that’s probably the way to go.

• I can see the case for focusing on the ITT in this setting (although I disagree with it). Regardless, Baicker et al. made a choice to feature the IV (and consign the ITT to the appendix). Given that, I think there’s a strong argument for discussing the “effective sample size” that corresponds to the confidence intervals in the paper (which is the smaller one).

Anyway, we clearly agree on the substance here. I think the only disagreement is presentational.

• Thinking about intention to treat might often be appropriate, but not here. A major reason for non-winners to not get Medicaid is that they weren’t actually eligible for Medicaid. Medicaid eligibility wasn’t checked for lottery entrants, oniy for lottery winners who then applied. The “intended to treat” group would be lottery winners who were actually eligible, and we just don’t know who they were for the non-recipients.

• I’m sorry to continue this line of comments, but I just realized that my “effective sample size” calculation above was incorrect. The IV scaling increases the point estimate and the standard error by a factor of 4, but the _variance_ by a factor of 16. So the “perfect-compliance-RCT-equivalent” sample size for the IV analyses of the effect of insurance on outcomes in the high A1C subgroup is actually even smaller: 39=624/16.

That said, I was also pondering my argument to Aaron that one should give “perfect-compliance-RCT-equivalent” sample sizes that correspond to the confidence intervals reported in Baicker et al, Table 2. That table included the full population of 12,229 people (not just the people with the relevant condition at baseline), so the “perfect-compliance-RCT-equivalent” sample size for the IV analysis of the effect of insurance on the full-population outcomes is 763.7=12,229/16.

In any case, I’m sorry about my error above.

• I think Matt and Aaron are on top of the issue here. The effective sample size is too small to detect meaningful differences. In the case of total cholesterol the rate was reduced by 17% (from 14.1% to 11.7%). If that kind of drop is not detectable by the study then I think it is a problem. Unfortunately, the researchers were constrained to the number of people in the lottery and likely limited by survey funding for the number that could be included in the analysis. But, it might have been better to not release results unless they had enough sample to detect meaningful differences or, at least report the minimum difference they could have detected.

Also, it should be noted that while 42% of those selected in the lottery enrolled, 18% of those not selected in the lottery eventually enrolled. And, those who were selected only had ~25% of the months with coverage (6 of 25 months) compared to 10% of months with coverage (2 of 25 months) for those not selected. These are not huge differences in the dose (15 points) compared to an RCT with a 100 point difference in dose.

• Where are you getting your formulas for effective sample size calculations with IV? I don’t know this formula, and I’m curious to read about it.

• Forgive me if this is a dumb question, but what’s the appropriate way to measure if this is a good or bad program? Your line about “It’s just the first step in the chain of events that leads to quality” seems spot on, but that’s seems like the sort of thing that can’t be easily determined, to say the least, from a two-year study. And even if it can be determined more easily, what do we look for? Slimmer people? Lower blood pressure? Less smoking? Fewer cases of diabetes?

• I’m getting a little frustrated with people over point (5). We don’t expect car insurance to reduce car accidents do we? My understanding is that most studies of both private and public health insurance find little-to-no effect on health outcomes. The fact that there were some positive effects of medicare on mental health outcomes is pretty impressive. But that’s not the purpose of insurance.

• Apples and oranges. Car insurance doesn’t cover oil changes, tire rotation, regular maintenance etc. and therefore doesn’t contribute to better “car outcomes.” Accidents are “catastrophic events” — in the same way that a cancer diagnosis is a catastrophic event — we obviously don’t expect health insurance to reduce the rates of cancer prevalence (though it may increase the rate of diagnosis).

Our conception of health insurance is that it should cover basic things like physical exams, blood pressure screenings, etc. that fall under “preventive care.” Better access to preventive care is assumed to improve health outcomes. Thus, if insurance gives better access to preventive care, it should improve health outcomes.

• I agree that we expect more from health insurance than we do from car insurance. However, if we’re talking about financial hardship health insurance functions exactly like car insurance.

One of the things we want from health insurance is a reduction in the financial hardship experienced by folks with health events. Since we want this from health insurance it is entirely reasonable to examine the question. Based on what little I’ve looked at it looks like the Medicaid expansion successfully prevents financial hardship.

On an unrelated note there are private products on the market that cover car maintenance in the same way health insurance covers preventative care. They’re usually called warrantees. I suspect these don’t increase the insurance companies profits enough for the companies to ever include them in coverage, but there isn’t anything preventing insurance companies from offering these products as part of coverage.

• Sure, health insurance can prevent financial hardship — but that’s predicated on an event causing financial hardship. In a comment on TIE’s other post I mentioned that there is probably selection bias into Medicaid — that is, only a portion of the people that were selected in the lottery chose to enroll. Those are likely the ones who already have unmet health care needs, and I would wager that they’re not representative of the average individual eligible for Medicaid in Oregon.

I think if the analysis were expanded to look at all of the people eligible for Medicaid (including the ones who didn’t take Medicaid when it was offered) the impact on financial hardship would be less significant.

If the goal is to prevent financial hardship through insurance, catastrophic coverage (which is cheaper than Medicaid) is likely a more efficient way to go. I think it’s fair to say that financial hardship occurs more from “catastrophic-like” events than ordinary doctor’s visits etc.

I’m not sure that warranties could be considered to cover “routine” maintenance. We’re talking about pretty major things like changing timing belts or engine failure — though I’ll admit I’m not a mechanic/engineer/car insurance actuary or claims specialist so take what I’m saying with several grains of salt.

• Re item 5: If you take the alternative to Medicaid to be “no Medicaid” then your point stands. But what if the alternative to Medicaid is cash; that is, the total aid to the poor doesn’t go up or down, but rather is paid more in cash than in kind benefits?

There are many individuals in favor of this approach, and this study would tend to support it. What if what the person really wanted was a more reliable car, or extra hours of day care, or private school tuition? The ability to buy the things that someone really wants to buy should improve “soft measures” much more, especially if combined with widespread availability of community rated plans with no preexisting condition exclusions, so that someone who wanted health insurance against catastrophic problems could get it.

In which case, shouldn’t premium subsidies in social insurance plans, e.g. PPACA, just be replaced by cash grants to the poor – which could be done with expansions of existing programs such as EITC? What is the advantage to giving subsidies only for a health insurance that does not appear to improve health very much? It is difficult to argue that cash doesn’t reduce financial hardship 🙂

• The problem is that human beings are unable to handle effects with low probability and high impact outcomes.

Give a poor person (a little) more money and they are likely to spend it on other more immediately important things, like new shows for the kids etc. In the end they will still get sick and society is still on the hook. Laws exist to change peoples behavior with their and society’s best interest in mind (e.g pyramid like Madoff).

For those of us like Ro above i have only disgust, and thank you for the Scrooge quote.

• If the study found that Medicaid recipients lived 6 months longer, then I would agree that there would be a need to “nudge” recipients to take health insurance.

But this study points out that Medicaid does not really improve heath outcomes, so now we are just talking about money and happiness, The problem with being poor is not the low probability, high impact risks, it is rather that many risks that would be high probability, low impact for someone in the middle class are high probability, high impact.risks.

For example, if your child is sick, you and I can afford to take time off from work without it being a big deal. For a poor worker dependent on getting shifts, this might get them stuck with the label of “unreliable.” Or another big fear is car trouble. Again, it is not a low probability risk; it happens all the time if the best car you can afford is a couple thousand dollars. And again, it can have seriously adverse consequences for that individual’s employment situation.

Who am I to say that they need to be protected from a \$10,000 hospital bill if they get a heart attack when their roof is one big storm away from caving in? People in general might not be good at dealing with low incidence risks, but poor people do not even have the resources to deal with normal, everyday risks. A bit of straight up cash could go a long way there.

Further, I still think the best health care reform would be a universal, catastrophic coverage layer and whatever you want to do beneath that, so catastrophic care would be dealt with in my best of all possible worlds. But that is not necessary for this discussion.

@SB in STL: your analysis seems correct, if sadly cynical, Nevertheless, there exist voices of compassion on all sides, and, as you point out, they will ebb and flow into and out of power at any given moment, it is important to always leave the door open.

• If you can’t afford to deal with any unanticipated expenses, you allocate resources based on whatever is more urgent. If your car won’t start you pay the mechanic and put off the check-up.

Since America is unwilling to let the poor die in a ditch when a trip to the hospital could save them, health care can always be put off, until it’s a crisis, in which case, the hospital will treat you regardless of whether you can pay.

So, if you look at things from a poor person’s perspective, buying health insurance is low priority. This is the point of a universal mandate.

• Hi Nick,

The obstacle to replacing in-kind programs with straight cash subsidies is primarily political–the self-anointed “makers” that comprise the majority of at least one of our esteemed political parties are even more loathe to give straight cash to THOSE people than to pay taxes to support programs of which they themselves will eventually get a cut. For example, through eventual participation in those programs (Medicare), or through the business opportunities those programs create (Managed Care Medicaid and private exchange-based plans run by for-profit, publicly traded entities).

In other words, a cash transfer program such as you describe is straight-up welfare for poor people (or the 47%, if you prefer), whereas with in-kind programs, you get a mix of poor-folk welfare and corporate (or rich-folk) welfare. Of course in more genteel times (like the 1960s and 70s), there were prominent academic/political conservatives championing ideas, like straight cash transfers, as a more effective and dignified means of helping the less fortunate, and they actually got a fair hearing in the court of policy-making (Milton Friedman, for example). Unfortunately such voices have all but disappeared from contemporary conservative political discourse (and those who do advocate such policies are marginalized from power by the party apparatus). Quite sad, to me anyway.

While I’m reluctant to post slogans on a “serious” blog, its worth keeping in mind that, when it comes to the spending of tax dollars on social programs in the US of A, “Left v. Right is how the Top divides the Bottom.”

Thanks,
SB

• Does the above analysis include the EITC?

• You have a point about offering cash. However, from a strict utilitarian perspective, Medicaid is probably more efficient for two related reasons. First, medicaid is quite simply the cheapest healthcare available–it easily underprices any private insurer, and has much lower loading costs as well. Second, dollar-for-dollar, the utility gain to a risk averse individual from reducing that risk is greater than the utility gain from the cash equivalent. That is, if we wanted to match the same level of utility that enrollees currently get from Medicaid, we’d have to pay them a much higher cash payment than their expected medicaid costs.

• Also, basic cash grants to individuals are absolutely politically unfeasible, while giving health insurance (which has a specific purpose) is far easier to sell to Congress and the public.

• 99% of docs will take Medicare. Less than 50% will take Medicaid. Thats the big difference.

• If people had an expected life span of only two years, then this Medicaid experiment would have been much more informative.

• This study points up the problem of how large — and how carefully controlled for crossover — population studies of health care issues need to be to achieve adequate statistics, for the simple fact that such large numbers of the population are not effected by the problems evaluated and the small numbers of people who are effected degrade the statistical results.

The current abysmal state of knowledge about breast cancer management is a good example. Despite the apparent large size — in the 50,000 to 150,000 patient range — of the pioneer studies about breast dancer management, they have all collapsed due to the small numbers of women who actually got breast cancer. This allowed the studies to be confounded by small amounts of crossover and by such things as the higher rates of death from other causes in the Swedish Two County Study treatment group (does mammography increase your chances of dying in a car crash or from stroke?) or the inadvertent sabotage of the Canadian study by doctors and nurses who transferred patients with known masses on initial screen to the treatment group since otherwise then-current OHIP policy would not pay for mammograms — the clinical staff had no idea that ten to fifteen extra cases of advanced breast cancer would throw such a large study into the trash can.

It often takes millions of patient-years to make a general population study work. The Oregon study is just starting out, and is handicapped by all the problems noted above and more. Carried out long enough — and without crossover — it might show us more when time passes, but the idea of continuing to deprive large numbers of people of insurance will probably create ethical issues that will lead to the study being aborted long before the relatively small samples can run their course.

• Affect/Effect. Edit.

• Proportion with BP over 140 drops from 16.3% of patients to 15.
High HDL drops from 14.1% to 11.7%.
HgbA1C drops from 5.1% to 4.2%.

I don’t know how they did their analysis, but I am stunned that this did not result in significance. The point estimates are actually pretty different. That’s a 20% drop in poorly controlled glucose. Not bad!

Furthermore, the fact that all three measures dropped is not a coincidence. I understand why they would treat them independently to be fair and simplify analysis, but considering all three together is extremely suggestive and relevant in this case.

And why are they even wasting time worrying about the average BP in a healthy population? As you mentioned yesterday focusing on the percentage of people at unhealthy levels (the three measures described above) is clearly the correct measure.

• What Aaron may not have realized in asking for analyses to be restricted to the ‘afflicted’ hypertensives/diabetics is the powerful effect of ascertainment (which was demonstrated by increased diagnosis). On average, ascertainment alone picks up the more marginal/borderline patients, ‘improving’ apparent outcomes without any practical differences from healthcare delivery. This is a well known epidemiology bias. The population level approach undertaken addresses that.

• That would only matter if you were comparing attributes of the affected group with and without Medicaid.

What we’re proposing is to compare the percentage of patients affected with and without Medicaid.

Increased diagnosis/ascertainment would work against a positive finding in that case.

Although I doubt it would be much of a problem anyway for something like BP>140.

• We live in an imperfect world here.

Looking at just the numbers on either side of some magic threshold for high blood pressure doesn’t tell us nearly as much as the actual reductions by people near that range, because a drop of 20 points in a severely hypertensive person might not move them over the line (but still be worth a lot!) whereas a drop of 2 points in a borderline case might move them over the line but not be worth thinking about.

However, looking at blood pressure rates only among hypertensives (for a snapshot in time study, not panel data) isn’t as helpful, because hopefully part of the effect is to move people out of t he hypertension range.

The best thing might be to carve up the blood pressure ranges into different threat levels and compare the distributions. That sort of analysis takes longer to explain than you could shoe-horn into a NEJM article, but wouod be worth doing in the longer version I expect they’ll write for, say, the Quarterly J of Economics,

On a related note, one thing that struck me is the number of control group people takng medication for blood pressure. Perhaps one explanation for a low Medicaid effect on this measure is that blood pressure meds are cheap enough for even the uninsured to get pretty easily, so there isn’t as much room for Medicaid to have an impact. I haven’t seen that explanation discused, which makes me think it’s probably wrong, but I’d like to know why.

• What I haven’t read a lot about in response to the Oregon study – and what I hope to hear – is how the results of this study change your views, if at all, about how Mediciad programs should be reformed.

• I’d love to know if anyone has created a list of people who

1. After Newtown, criticized the nation’s mental health care system, and
2. After this study, ignore the MH benefits and claim that Medicaid is worthless.

• As a person diagnosed with Type II diabetes I am puzzled about why diagnosis (most diabetes diagnoses are Type II) does not lead to significant drop in blood sugar. In fact that outcome is so unlikely that it should lead to an immediate investigation itself — before we can make anything of the effect of expanded Medicaid.

Medicaid — I have Medicaid too but also Medicare — has so little acceptance from normal medical practices that all it accomplishes — which is terrific! — is allow indigent patients — I’ve been there — to attend the country hospital clinics or go to the ER and not get huge bills they cannot pay — which catastrophically impact on their credit ratings (if they have any) and thereby job and housing opportunities. Medicaid basically asks already squeezed doctors to take pennies on the dollar — can’t do it.

A big question here — thinking about the unthinkable — is if Medicaid patients don’t get much lower level of care than normal in Oregon — has to be asked. Also unthinkable: are some (not most!) Medicaid patients in Oregon so dysfunctional that they don’t utilize care effectively — possibly same dysfunction that has some on Medicaid in the first place?

First and foremost question that needs to be answered: what could possibly lead to no effect on blood sugar among 300+ diagnosed patients?

• “The reason I have insurance, and likely you do as well, is to protect me and my family from financial ruin. When I get sick, I don’t sit at home and let the insurance take care of me. I get off my butt and use the health insurance as the means by which to get health care. ”

If that is the case then we should consider giving them cash so they are not so poor and can buy insurance. Some kind of basic income guarantee. In this debate we should always keep in mind that most people lose monetarily by having health insurance. So the poor might be correct in choosing the money over the insurance given an option. This is relevant because the latter might be cheaper leaving us better off than them better off.

Also:
To early to tell but this study might prove Robin Hanson correct.

• One more point:

Blood sugar and high blood pressure are just measurable symptoms. We only care about them if they lead to pain and suffering or death.

This is important because:

1.Symptoms are not health, for example high blood pressure controlled by medication still effects health negatively. Further hypertension may not lead to any adverse effects in some individuals. They may die of an unrelated cause having had high blood pressure but never a heart problem or stroke.

2. Some researchers, like Nortin Hadler, claim that hypertension and high blood sugar should not be treated until they are much higher than were we start to treat them now.

• I am confused about why the authors of the study would use such a small sample, choose such a narrow time line then measure outcomes that can take an extended period of time to correct. Diabetes, hypertension and depression are not like the common cold that will improve in a week if untreated and will improve in 7 days if treated. The time from start of treatment to noticeable results will depend on how long the patient has had the symptoms, and in this case since we are talking about the poor who have little or no previous access to health care they may have had these diseases and other pre-existing conditions or are such severe cases that it will take a multi-modality approach. Access to food, exercise and quality of living conditions will also effect the outcomes as well as compliance with medications.

It seems like a pretty thin file to indict such a large program, especially considering the consequences of not treating any of these people, consequences that will in turn effect public health, safety, learning, and whether these people can become productive members of society again.