KAYAK2U

January 19, 2010

Klein Boggle – Little Big Lies

A Klein Boggle is a certain non-orientable argument, i.e., a contention (a one-dimensional manifest) with no distinct "logical" and "reasonable" sides.

My newshounding has been lackadaisical of late, but I did manage to stop by the Post's Ezra Klein corner today, to find the dear girl still practicing the art of the big lie, to the usual acclaim of his dutifully credulous readers.  Here 'tis (with Klein's link to the PDF study):

I get the sense that some Democrats in Congress could use a refresher on the stakes in health-care reform.  According to a study conducted by researchers at Harvard Medical School, 45,000 people die a year because they lack health insurance.

He's told the "45,000 people die" lie several times.  He's been repeatedly told that it is a lie, and he continues to tell it, anyway.  I'm not aware that Klein even grants any sort of controversy with respect to the matter, which, one would think, an avowed health care policy expert ought to be able to take in stride, especially if he truly thinks he's on the right side of it.

But, of course, Klein is not on the right side of the erstwhile health care debate.  As many have pointed out, the study that grants Klein and others their plausible denialability was baldly flawed in its methodologies.

Here are some of the critiques of the study, in no particular order… 

David Dranove, a pro-reform academic type:

[T]he Harvard study uses a methodology that is inherently biased. (Note to the authors of the study: I did not say that you were inherently biased. I said that your methodology is biased. On the off chance that you see this blog and care to respond, let’s keep the discussion centered on methodology. I have had enough of your personal attacks.)

Now I have to get a bit technical. In regression and related analyses, a critical assumption is that the unobservable characteristics of the “control” and “experimental” groups are uncorrelated with the observables. Translation in this case – if the regression model does not include all possible factors that might predict mortality, and just one of these omitted factors is correlated with insurance status, then the reported coefficient on insurance status is biased. This is an onerous requirement for sure, but it must be met if bias is to be avoided. Without this full set of variables, and in the absence of a randomized experimental design, it is still possible to avoid bias by using advanced statistical techniques such as “instrumental variables” regression. But the Harvard study does not use this technique.

The implication is that their results are biased. We can even guess at the direction of the bias. They started with a few control variables and the estimated impact of being uninsured was much bigger than 44,000 lives lost. When they added additional control variables, the estimated impact fell. It is plausible to suppose that with additional controls, the estimated impact would shrink further, perhaps to the point of statistical insignificance. I don’t know that for sure. No one does.

David Hogberg at Investor's Business Daily:

Two of the authors, the husband and wife team of David Himmelstein and Steffie Woolhandler, are co-founders of the single-payer advocacy group Physicians for a National Health Plan. But even if they were completely disinterested researchers, the article cannot reliably link insurance status to mortality. The article itself says so:

Our study has several limitations. NHANES III assessed health insurance at a single point in time and did not validate self-reported insurance status. We were unable to measure the effect of gaining or losing coverage after the interview.

In other words, the researchers had no idea if respondents who were uninsured in 1993-1994 remained uninsured through the time that they died, or if they obtained insurance coverage during that time. (They also don’t know if respondents insured in 1993-1994 later lost insurance.)

Thus, it’s quite possible that the people who were uninsured in 1993-1994 and had died by 2000 may have had insurance and, presumably, greater access to health care sometime between 1993-1994 and their time of death. To put it gently, that’s a pretty serious flaw.

The authors follow that passage with this lame defense: “Point-in-time uninsurance is associated with subsequent uninsurance.” In other words, if you are uninsured at one point in time there is a reasonable likelihood that you will be insured later on.

To support that sentence, the authors cite a study in the journal Health Affairs titled “Health Insurance Coverage And Mortality Among The Near-Elderly.” This study examined insurance status of people aged 55-64 in 1992 and its effect on mortality through 2000. And this study did examine the insurance status of those uninsured in 1992 through subsequent surveys. But this is what the study found:

Among adults who were uninsured in 1992, the proportion of respondents who reported being publicly or privately insured rose progressively in the ensuing four surveys (46.6 percent, 58.4 percent, 66.1 percent, and 74.5 percent), as nearly half reached age sixty-five and became eligible for Medicare by 2000.

Thus, people who are uninsured at one point in time are more likely to be insured in subsequent years. It would be more accurate to say: “Point-in-time uninsurance is associated with a subsequent increase in insurance coverage.” Doing that, however, would have only widened the gaping hole in the research behind the “45,000” statistic.

Randall Hoven at American Thinker, in parallel with the folks at Harvard, concludes that voting Democrat causes cancer:

By applying the same logic in interpreting both studies, three times as many lives could be saved by no one voting Democrat as by providing universal health insurance.

Michelle Malkin caught the fever:

How did these political doctors come up with the 44,000 figure? They used data from a health survey conducted between 1988 and 1994. The questionnaires asked a sample of 9,000 participants whether they were insured and how they rated their own health. The federal Centers for Disease Control tracked the deaths of people in the sample group through the year 2000. Himmelstein, Woolhandler and company then crunched the numbers and attributed deaths to lack of health insurance for all the participants who initially self-reported that they had no insurance and then died for any reason over the 12-year tracking period.

At no time did the original researchers or the single-payer activists who piggy-backed off their data ever verify whether the supposed casualties of America's callous health care system had insurance or not. In fact, here is what the report actually says:

"Our study has several limitations," the authors concede. The survey data they used "assessed health insurance at a single point in time and did not validate self-reported insurance status. We were unable to measure the effect of gaining or losing coverage after the interview." Himmelstein et al. simply assumed that point-in-time uninsurance translates into perpetual uninsurance—and that any health calamities that result can and must be blamed on being uninsured.

Another caveat you won't see on Grayson's memorial to the dubious dead: The single-payer advocate-authors also conceded in their study limitations section that "earlier population-based surveys that did validate insurance status found that between 7 percent and 11 percent of those initially recorded as being uninsured were misclassified. If present, such misclassification might dilute the true effect of uninsurance in our sample."

Media Matters pretends to rebut Malkin, but completely ignore the study author's own admission of the study's "several limitations".

Someone at FactCheck.org was warily honest:

Now, on to the tough question: Is the 45,000 figure accurate? We can’t say for sure, but […]

… and…

There has been some criticism of this type of research and its ability to find a direct causal link. A 2003 commentary by Richard Kronick in Medical Care Research and Review questioned whether other factors beyond uninsurance would reduce the greater mortality for the uninsured. Kronick recreated the Franks study using more recent data and, after adjusting for various factors, also found a 25 percent greater risk of death for the uninsured. But he said: "It seems likely that if we were able to control for additional factors, such as health-related behaviors (smoking, alcohol consumption, obesity, and risk-taking behaviors more generally), wealth, or value placed on health or health care, the estimated effect of being uninsured would be reduced further. What is uncertain is whether the reduction would bring the estimated hazard ratio all the way down to 1.0 or whether an independent effect of being uninsured would remain." (Other studies, including the Franks study, did adjust for smoking, alcohol consumption, obesity and income.)

Another recent report, written by former Congressional Budget Office Director June O’Neill and her husband, economist Dave O’Neill, said "that lack of health insurance is not likely to be the major factor causing higher mortality rates among the uninsured. The uninsured — particularly the involuntarily uninsured — have multiple disadvantages that in themselves are associated with poor health." Those disadvantages include education level and income. The O’Neills’ study, published by the conservative Employment Policies Institute, separated those it deemed "voluntarily uninsured" (anyone earning 2.5 times the poverty level) from those considered to be "involuntarily uninsured." The study looked at data on persons aged 51 to 61 from the Health and Retirement Survey and determined the "involuntarily uninsured" had an 11 percent higher probability of dying; the number dropped to 3 percent when controlling for smoking as well as education and income. The “voluntarily uninsured” had a 2 percent to 3 percent greater probability of dying. EPI, the publisher of the study, supports business interests and has said that the “living wage campaign” is “an organized effort to force employees to inject a welfare mentality into the workplace.”

Carl at NoOilForPacificsts is not be left out and provides a multitude of links:

[T]he study's methodology obviously is flawed. The paper uses survey data gathered between 1988 and 1994 by the National Center for Health Statistics (NCHS). About 34,000 people were interviewed about various health-related characteristics, including whether they were insured. The authors then excluded those over 64 (virtually of of whom are covered by medicare) and children under 17. They also excluded non-elderly already covered by Medicare, Medicaid or VA health insurance, because "a substantial proportion of those individuals had poor health status as a prerequisite for coverage." The researchers also culled those with missing or incomplete interview data. The resulting relevant population was 9,004 surveyed individuals, of which about 84 percent were listed as insured while the remaining 16 percent were not. In a 2000 follow-up of the 9,004 individuals, 351 (or 3.1 %) had died. Then, for the most part, they assumed that mortality among the uninsured was caused by the lack of health insurance. This is nonsense, for three reasons: 

First, demographics matter:

People with insurance tend to be healthier, wealthier, wiser and more gainfully employed than people without insurance.

So you can't blame their longer lives only on their insurance.

To be sure, the study acknowledges several other factors contributing to mortality, such as "gender, age, race/ethnicity, poverty income ratio, education, unemployment, smoking, regular alcohol use, [sub-standard] self-rated health, [and sub-standard] physician-rated health." But, because they found "no significant interactions" between the lack of insurance and those other variables, the authors downplay those factors.

 Second, there's scant causal connection. The lack of health insurance isn't static–over 80 percent of the uninsured regain coverage within two years. And one can't assume deaths were due to any lack of health care–after all, bad health care also can be fatal. Yet the study presumes lack of health insurance is among the causes. As John Goodman, president of the National Center for Policy Analysis notes:

The subjects were interviewed only once and the study tries to link their insurance status at that time to mortality a decade later. Yet over the period, the authors have no idea whether subjects were insured or uninsured, what kind of medical care they received, or even cause of death.

And the study itself concedes that "between 7% and 11% of those initially recorded as being uninsured were misclassified," raising questions about the sample's accuracy. 

Third, other research is to the contrary. For example, a study by UCSD prof Richard Kronick, Ph.D., found:

Adjusted for demographic, health status, and health behavior characteristics, the risk of subsequent mortality is no different for uninsured respondents than for those covered by employer-sponsored group insurance at baseline.

Further, a June 2009 report by Drs. June and David O’Neill of Baruch College found the differential mortality between insured and uninsured to be of questionable statistical significance, and in any event only about 3 percent when comparing the insured with the voluntarily uninsured. Overall, they conclude:

a lack of health insurance is not likely to be the major factor causing higher mortality rates among the uninsured. The uninsured–particularly the involuntarily uninsured–have multiple disadvantages that are associated with poor health.

Conclusion: I've always supported a tax-funded safety net for the poor, including for those who can't afford health insurance. But rejecting phony scare stories is crucial for assessing the scope of the actual problem and crafting the appropriate policy remedy.

And there you have it.  It took me about five times longer to copy and paste it all than it took to Clusty it all up, so how come a guy whose business it is to know this stuff pretends it all doesn't exist?

The more I read Klein and get to know how he responds to the challenges put to him, the more I have to conclude that Ezra Klein is a Quisling and a liar.  Look at his reaction to Grubergate (I only had time for a short comment there, but others made points, fazing Klein not a whit.)

How some people can live with themselves, I don't know.

MikeSoja - January 19, 2010 -- 10:21 pm    Filed in: Dishonesty, Health care, Propaganda  
« I like Mises
The Essential Vandam »
As Happy As A Clam

Kayak2U.com

msoja RSS 2.0

Other's Pages









Search

Archive

November 2026
M T W T F S S
 1
2345678
9101112131415
16171819202122
23242526272829
30  


  • Books4Sale



    Soja's GoodReads Books4Sale Cheap

  • Category Cloud


  • Tag Cloud