Tag: CER

ObamaCare’s Priceless Warm Glow

Ed Kilgore says ObamaCare opponents don’t care about cost-benefit analyses:

many of them just can’t bring themselves to even notice that…Obamacare with its Medicaid expansion, health care exchanges, and regulatory mandates [does] actually provide health coverage to people in exchange for the money and the “liberty” surrendered.

Speaking of, what is the exchange rate between liberty and “liberty”?

But about those benefits. What benefits do broad-based expansions of health insurance, like ObamaCare, actually provide? Aside from giving Kilgore a warm glow, that is.

It turns out there has been only one—one!—scientifically rigorous study of that question. The Oregon Health Insurance Experiment found Medicaid coverage confers modest improvements in self-reported health and financial security. The first batch of that study’s results appeared more than a year after Congress enacted ObamaCare. And there remains to this day absolutely zero evidence that Medicaid or other broad-based expansions of health insurance buy us the most health and financial security per dollar spent.

Then again, the Oregon Health Insurance Experiment did not attempt to measure the value of the warm glow that Kilgore and others derive from Medicaid and ObamaCare, one that appears to be worth trillions of dollars of other people’s money.

‘The Moment It Produces Useful CER, PCORI Is Toast.’

An excerpt from a Politico Pro article (paywall) on the “Patient-Centered Outcomes Research Institute,” President Obama’s comparative-effectiveness research agency:

The point of comparative effectiveness research is to compare two or more different ways of treating the same condition to see which one works best. The idea is that if definitive best practices can be established, they will be widely adopted by providers and may be preferentially reimbursed by payers. Cheaper treatments that are effective would be favored.

It may sound harmless — like common sense, even, to the uninitiated — but it’s a menacing prospect to some pharmaceutical companies and medical device-makers who are concerned that their products may wind up on the wrong side of the ledger.

For this reason, Michael Cannon, director of health care studies at the Cato Institute, says good comparative effectiveness research is almost suicidal.

“The whole point of [comparative effectiveness research] is to find out what doesn’t work,” Cannon said in an email. “Every time the government has tried to do CER, the guys who provide the stuff found not to work successfully lobby to have the offending agency defunded. I see no reason to think this time will be any different. The moment it produces useful CER, PCORI is toast.”

And that’s just one source of opposition.

Other sources of opposition include patients who don’t like restrictions on their health care subsidies, thank you very much.

(The “suicidal” bit is confusing, and wasn’t my language. So to clear up any misunderstandings: comparative-effectiveness research is good. Markets both produce and employ it. Government is so incompetent that it cannot reliably produce CER, much less make use of it. Markets are smart. Government is stupid.)

All That NAEP Tells Us Is Things Ain’t Good

Yesterday, another round of scores on the National Assessment of Educational Progress – the so-called “Nation’s Report Card” – came out. They revealed flattened 4th-grade math achivement between 2007 and 2009, and a two point (out of 500) increase in 8th grade.

So what do these bits of data portend? Ask the experts:

“The trend is flat; it’s a plateau. Scores are not going anywhere, at least nowhere important,” said Chester Finn, president of the Thomas B. Fordham Institute, according to the New York Times. “That means that eight years after enactment of No Child Left Behind, the problems it set out to solve are not being solved, and now we’re five years from the deadline and we’re still far, far from the goal.”

Next, former National Center for Education Statistics commissioner Mark Schneider concluded that “either the standards movement has played out, or the No Child law failed to build on its momentum. Whatever momentum we had, however, is gone.”

And then there’s Michigan State University professor William Schmidt, a leading national-standards proponent, who opined in the Baltimore Sun that “there is a hardly any change. There is hardly any difference. How could we as a nation let that happen?” His solution to the problem: National standards, of course.

So what do I think about all this? As a long-time critic of NCLB, I am glad to see people seizing on the latest results and declaring the law a failure. It helps to advance my goal of ending the greatest federal education intervention to date, and I think NCLB supporters kind of deserve these attacks on their law. They have repeatedly given NCLB credit for positive things the evidence couldn’t come close to supporting, and it’s nice to see them on the defensive after all their overreaching.

That said, just as previous NAEP results couldn’t prove that NCLB was working, the latest NAEP scores don’t prove that it is not. We simply don’t have sufficient information about the myriad other variables affecting education to do that.

Which leads me to a much bigger problem: People using ambiguous  NAEP scores to push their favorite reforms. Some “standards and accountability” proponents, for instance,  argue that achievement improvements came as a result of states implementing standards-and-testing regimes during the 1990s. And William Schmidt suggests that the latest NAEP results demonstrate a need for national standards.

Now, NAEP simply cannot be used in any reasonable way to justify the national-standards assertion. We’ve never had national standards, so we obviously can’t measure their outcomes with NAEP. We can, however, attempt to use NAEP to assess the assertion that the push for state standards and testing in the 1990s drove real improvement. We can attempt, that is, but any conclusions will be riddled with problems.

Let’s start with the NAEP exam from which yesterday’s results came, the “main” mathematics exam. After that, we’ll look at some “long-term” NAEP results.

Take a look at the chart below. It is the 4th-grade trend line for the main math NAEP, with vertical lines separating what I’ll call the pre-accountability period (1990-1996), the state-accountability period (1996-2003), and the NCLB period (2003-2009). The numbers below each period are the per-year changes in scores for the periods above them.

RegMath41996

What do the numbers show us? At first blush, the message seems to be that scores were improving slightly faster in the pre-accountability period than the vaunted state-accountability period, and appreciably faster than under NCLB. But this breakdown may beg more questions than it answers.

At what year, for instance, should you peg the start of the state-accountability period? I chose 1996 because by then, according to a count by Hanushek and Raymond (Table 1), twelve states – a pretty large number – had some sort of accountability mechanism in place. I could, however, have chosen 1992, because by 1993 three states had such mechanisms and the accountability movement could certainly be said to have been underway. Similarly, NCLB was enacted in 2002 – is it right to start the NCLB period in 2003? Obviously I couldn’t start the period in 2002 because there is no data for that year, but why not 2005? After all, though enacted in 2002, NCLB took a few years to be fully implemented.

Let’s make an adjustment. I’m going to keep the start of the NCLB era at 2003 because that’s pretty close to the enactment year – though it could very well produce misleading results – but will move the end of the pre-accountability period, and hence the start of the state-accountability period, to 1992.
RegMath41992

Now what should we conclude? Again, it appears that the pre-accountability period had the best results, but this time by a much bigger margin. That said, that period included only two years – hardly sufficient data to identify a trend. Also, the NCLB period fared better than previously against the state-accountability years.

Of course, the main NAEP gives us data for less than two decades. So what does the long-term NAEP show for 9-year-olds (roughly 4th graders) in math?

Here’s a long-term trend chart, which like the main NAEP charts above is broken into periods with score-change-per-year noted below. I’ve broken it into the period before the 1983 publication of the landmark A Nation at Risk report, which scared people silly about the schools; the post-ANARbut pre-state-accountability period; the state-accountability period; and the NCLB period.
longterm96

What do these results show?

In contrast to the main NAEP scores, the greatest improvement on the long-term test occurred during the state accountability period, and the second greatest under NCLB. But again, this raises more questions than it answers: Why the difference between the main and long-term results? When is best to start each period? Does it make any sense to start a period with A Nation at Risk? Should the NCLB period start in 1999, well before the law was enacted, or in 2004, two years after it’s passage? And the questions go on.

Like we did with the main NAEP results, let’s once again look at a different start date, 1992, for the state-accountability period, this time on the long-term exam.

longterm92

Once again we get a whole different story. Now it is the pre-accountability period, not the state accountability period, that shows the best outcomes. And so the ambiguity continues…

All of this, of course, goes to show that NAEP results cannot be used with any confidence to conclude that any particular reform that occurred within the NAEP time span worked better or worse than any other reform within that span.

That said, there is one thing that we canuse NAEP to demonstrate very powerfully: As Andrew Coulson’s chart below vividly illustrates, if moving the achievement needle as measured by NAEP is the goal of education spending, then we have really been getting robbed! Moreover, to the extent that standards-based reforms have been a major national phenomenon since the 1990s, it is impossible to conclude that they have done very much good. Indeed, if we are to conclude anything, it is that it is time to focus on reforms that are completely different from the top-down “solutions” that have given us so little, and taken so much.

andrew-coulson-cato-education-spending

CER: A (Slightly) Different Perspective

My colleague, Michael Cannon, makes several good points about comparative effectiveness research (CER), both in his letter to USA Today and in his excellent paper on the subject. I strongly agree with him that we should not reflexively oppose CER—much of health care spending is wasteful or unnecessary, and it makes sense, therefore, to test and develop information on the effectiveness of various treatments and technology, giving consumers tools to evaluate the value of the care they receive. There is also a case for the use of CER in taxpayer-funded programs like Medicare and Medicaid. Taxpayers should not have to subsidize health care that has not proven effective, nor can Medicare and Medicaid pay for every possible treatment regardless of cost-effectiveness.

However, I am more skeptical in general about CER than he is for several reasons.

  • First, “quality” and “value” are not unidimensional terms. In fact, such concepts are highly idiosyncratic with every individual having different ideas of what “quality” and “value” means to them, based on such things as a person’s pain tolerance, lifestyle, feeling about hospitalization, desire to return to work, and so forth. For example, a surgeon may tell you that the only way to ensure a cure for prostate cancer is a radical prostectomy. But that procedure’s side-effects can severely impact quality of life - so some people prefer a procedure with a lower survival rate, but fewer side effects. Who is better suited to determine which of those procedures represents “quality” and “value,” a government board or the person directly affected?
  • Second, comparative effectiveness research too often has a tendency to gear its results toward the “average” patient. But many patients are outliers, whose response to any particular treatment, for either good or ill, can vary significantly from the average. This matters little when the research is simply informative. However, if the research becomes the basis for more prescriptive requirements, for example prohibiting reimbursements for some types of treatment, the impact on patient outliers could be severe.
  • Third, comparative effectiveness research can create a time lag for the introduction of new technologies, drugs, and procedures. The FDA, for example, has already caused delays in introducing drugs that have resulted in unnecessary deaths. Depending on how the final program is structured, comparative effectiveness research could create another layer of bureaucracy and testing between the development of a new drug, for example, and its introduction into the health care system. One only has to look at the difficulty in expanding Medicaid drug formularies to see how this could become a problem.

The advocates of government-sponsored CER clearly intend for it to be used as a basis for rationing care, not just in government programs, but for private insurance as well.

Cannon points out that government-sponsored CER is likely to be corrupted under pressure from special interest lobbies and politicians. I couldn’t agree more. Government-sponsored CER, therefore, is liable to yield the worst of all possible worlds, not only rationing, but rationing that is based on special interest lobbying rather than science.

Health care, is of course, a finite good. Therefore, it will always be rationed in some fashion. But, it is far better if the rationing agent is the consumer himself, rather than the government or any other arbitrary agent. The private sector is already undertaking CER. To the degree that consumers, insurers, and providers make use of this information, that is a good thing. If consumers don’t like how an insurance company, for example, uses CER in determining its reimbursement policy, he or she can choose a different insurer.

Government-imposed fiat rationing allows for no such choice. Therefore, we should oppose any government involvement in CER, and any efforts by the government to use CER to restrict reimbursement, especially in private insurance plans.

LTE re CER in USA Today

I’ve got a letter to the editor in today’s The USA Today on comparative-effectiveness research:

Commentary writer Kevin Pho misrepresented my views on comparative-effectiveness research (CER), which is the analysis of which medical treatments work best (“Unbiased research for doctors is good medicine,” The Forum, March 26).

Pho wrote that “drug companies, medical device makers and think tanks such as the libertarian Cato Institute have expressed concerns that health care rationing and denial of certain treatments or drugs would follow” taxpayer-funded CER.

In the Cato Institute study linked to in the piece, I write that rationing is the intent behind such research, but I do not express concern that it will lead to rationing. Indeed, I express the opposite concern: that taxpayer-funded CER will not eliminate low-value services, just as it has failed to do so in the past.

Pho uses AARP executive Bill Novelli’s words to suggest that Cato, as well as drug and device makers, use “scare tactics” to oppose taxpayer-funded CER. Far from engaging in scare tactics, my paper makes precisely the same observations that Novelli does.

By contrasting Cato to CER “champion” Hillary Clinton, Pho also gives the false impression that libertarians support CER less than those who support taxpayer funding.

Yet two themes of my paper are that CER is crucial and that removing government obstacles to private production would provide a much more stable stream of research — and broader use of that research — than taxpayer funding would. I think that makes me the champion of CER, not Clinton.

At a minimum, it is misleading to suggest that libertarians oppose CER.