If you have "no place to go," come here!

Flawed Dartmouth Atlas study only catalogs dead people, but HR 3200's "efficiency" payments are based on it

Never let it be said that the scientists who publish in dry, staid medical journals lack a sense of humor. That resurrecting dead patients line is the title of an article that appeared in JAMA [Journal of the American Medical Association] a few years ago, and beyond the fact that it provided me with a snappy headline, gives me the chance to post one of my favorite lolcats [again], and is cited in another article in another journal, it has no further bearing on this post.

The another article in another journal, Looking Forward, Looking Back: Assessing Variations in Hospital Resource Use and Outcomes for Elderly Patients With Heart Failure, is monumentally less gripping than, oh, the last installment of Harry Potter, or even the labels on cat food cans, but it's nonetheless an important data point in the present health care deform reform debate.

To back up for just a moment, the Dartmouth Atlas Project is a massive gathering of data gleaned from Medicare spending records over many years. Mapping the data has produced the realization that Medicare spending varies widely throughout the country. Peter Orszag, President Obama, and Tom Harkin, to name just a few personages, are all quite taken with it, and with the Dartmouth researchers' assertions that the patients in higher-spending regions fare no better than those in lower-spending regions.

If only those spendthrifts in Miami and McAllen could be made to behave more like those prudent paragons living in Minnesota, we could save hundreds of billions of $$$$$ in health care spending every year.

Not so fast, corpus breath. The Dartmouth Atlas only catalogs dead people. The researchers looked back over the patients' lives for the 6 months [and for some purposes, 2 years] before they died. Concluding that since they all died anyway [duh!], the ones who got more care [and therefore cost more money], didn't really need all that extra care [and therefore we don't need to be spending that extra money on them].

It's an attractive notion, but one of the things the Dartmouth researchers didn't do so much of was looking forward.

So this is what the California hospitals did, as reported in the slightly-more-readable-than-catfood-labels article on heart failure survival rates. They looked at all patients [not just the dead ones] with a particular medical condition [heart failure] and found that those who received more care [and therefore cost more $$$] were more likely to survive.

This point -- survival -- may or may not be interesting to academics, but you would think that it might have some importance in deciding public policy. You would be wrong on that count, legal-beagle breath. I'm not going to go looking it up in the Senate bills just now, but HR 3200 [from the Energy and Commerce committee website] takes the Dartmouth conclusion [less is more] and runs with it.

Regarding the 'public option':


(1) IN GENERAL.—The Secretary shall establish payment rates for the public health insurance option for services and health care providers consistent with this section and may change such payment rates in accordance with section 224.
(2) SUBSEQUENT PERIODS.—Beginning with Y4 and for subsequent years, the Secretary shall continue to use an administrative process to set such rates in order to promote payment accuracy, to ensure adequate beneficiary access to providers, and to promote affordablility and the efficient delivery of medical care consistent with section 221(a).
(e) CONSTRUCTION.—Nothing in this subtitle shall be construed as affecting the authority of the Secretary to establish payment rates, including payments to provide for the more efficient delivery of services, such as the initiatives provided for under section 224.

(f) LIMITATIONS ON REVIEW.—There shall be no administrative or judicial review of a payment rate or methodology established under this section or under section 224.

So far, we're only talking in general terms of 'efficiency' [which could mean anything] but wait until you see section 224.


(a) IN GENERAL.—For plan years beginning with Y1, the Secretary may utilize innovative payment mechanisms and policies to determine payments for items and services under the public health insurance option. The payment mechanisms and policies under this section may include patient-centered medical home and other care management payments, accountable care organizations, value-based purchasing, bundling of services, differential payment rates, performance or utilization based payments, partial capitation, and direct contracting with providers.

(b) REQUIREMENTS FOR INNOVATIVE PAYMENTS.—The Secretary shall design and implement the payment mechanisms and policies under this section in a manner that—

(1) seeks to—

(A) improve health outcomes;
(B) reduce health disparities (including racial, ethnic, and other disparities);
(C) provide efficent and affordable care;
(D) address geographic variation in the provision of health services; or
(E) prevent or manage chronic illness; and

(2) promotes care that is integrated, patient-centered, quality, and efficient.

(c) ENCOURAGING THE USE OF HIGH VALUE SERVICES.—To the extent allowed by the benefit standards applied to all Exchange-participating health benefits plans, the public health insurance option may modify cost sharing and payment rates to encourage the use of services that promote health and value.

(d) NON-UNIFORMITY PERMITTED.—Nothing in this subtitle shall prevent the Secretary from varying payments based on different payment structure models (such as accountable care organizations and medical homes) under the public health insurance option for different geographic areas.

And not only are we going to experiment on those hapless souls in the public option based on Dartmouth dead people, we're going to experiment on the old folks [and the disabled] in Medicare too.

Section 1833 of the Social Security Act (42 U.S.C. 3 1395l) is amended by adding at the end the following new subsection:
‘‘(1) IN GENERAL.—In the case of services furnished under the physician fee schedule under section 1848 on or after January 1, 2011, and before January 1, 2013, by a supplier that is paid under such fee schedule in an efficient area (as identified under paragraph (2)), in addition to the amount of payment that would otherwise be made for such services under this part, there also shall be paid (on a monthly or quarterly basis) an amount equal to 5 percent of the payment amount for the services under this part.
‘‘(A) IN GENERAL.—Based upon available data, the Secretary shall identify those counties or equivalent areas in the United States in the lowest fifth percentile of utilization based on per capita spending under this part and part A for services provided in the most recent year for which data are available as of the date of the enactment of this subsection, as standardized to eliminate the effect of geographic adjustments in payment rates.
‘‘(B) IDENTIFICATION OF COUNTIES WHERE SERVICE IS FURNISHED.—For purposes of paying the additional amount specified in
6 paragraph (1), if the Secretary uses the 5-digit postal ZIP Code where the service is furnished, the dominant county of the postal ZIP Code (as determined by the United States Postal Service, or otherwise) shall be used to determine whether the postal ZIP Code is in a county described in subparagraph (A).
‘‘(C) LIMITATION ON REVIEW.—There shall be no administrative or judicial review under section 1869, 1878, or otherwise, respecting—
‘‘(i) the identification of a county or other area under subparagraph (A); or
‘‘(ii) the assignment of a postal ZIP Code to a county or other area under subparagraph (B).

In the Dartmouth Atlas, the geographic areas in the lowest fifth percentile of utilization based on per capita spending are always deemed the efficient ones, with all the others indulging in wasted care, and gee, if the HHS Secretary should have trouble deciding which zip codes to assign to which 'counties or other areas' the Dartmouth Atlas is here to help.

For someone who waxes positively shrill on the proposed payment innovations [particularly accountable care organizations] to eliminate treatment variations, you should read Don McCanne.

No votes yet


Submitted by hipparchia on

and it's something i want to talk about too. thanks for the pointer.

Submitted by gob on

This is going to save money? Let's see, we need to hire people and give them enough resources to produce the data to identify the "efficient" areas (I wonder what the funding was for the Dartmouth Atlas Project). Then we also spend money on "incentive payments" to providers in those efficient areas, who, I assume, just pocket the payments as pure profit. I wonder how much money that works out to be. Then, by some magical process, this will drive costs down in the areas that are not "efficient". How is that supposed to happen? Wha?

I realize this is just a side point to the main topic here, which is that these so-called efficient areas are not actually likely to be efficient at anything except delivering corpses to the local funeral home.

Submitted by hipparchia on

you've actually identified one of the many points that are connected to this [and that i'd like to say more about later], but i had to stop writing this post at some point. :-)

as for funding, valley girl dug up some stuff on the robert wood johnson foundation, which is one of several funders [think: aetna, wellpoint, unitedhealth] of the dartmouth atlas.

some of the low-spending areas don't need to spend as much on their old people, because the population is relatively healthy and well off and doesn't need as much of the ultra-expensive healthcare once they reach old age. giving them extra 'rewards' is a waste of taxpayer dollars.

otoh, some of those low-spending areas probably need to be spending more and the extra infusion of cash will help them. it's a silly and wasteful method of getting a few extra dollars to a few more people who could use it, and will leave some people in middle-spending areas without any extra help, even though probably some of those need to be spending more money too.

Submitted by lambert on

Readers, please spread linky goodness to this post at other blogs; it really deserves to be widely read.

Looks to me like this is the smoking gun on how Medicare "entitlement reform" is going to, er, work.

NOTE I changed the headline, I think accurately, to tell the story better, and also changed the break from front page to inside.

Submitted by hipparchia on

dude, you can rewrite my headlines anytime.

it's not so much that the atlas is flawed, as it is the interpretations that are being derived from it. the atlas is an excellent pointer to topics where more work is needed, but on its own it still doesn't contain enough information to design major public policy around.

which is a real shame, because jack wennberg's original work if the description in this podcast is accurate] was pretty darn good. the dartmouth atlas is the second generation of that work, and demonstrates what we could do if we had a fully nationalized health data system like canada's or england's or australia's or ... but we don't, and it's folly to think that the care that the people in medicare are getting is representative of the care that the rest of us are getting.

not to mention there's that whole dead people thing. data collection on dead people is way easier and cheaper and faster than data collection on live people, but one would hope that health care policy would be designed to benefit live people.

Submitted by lambert on

... as dead people, then everything falls into place, doesn't it? Ha ha, only serious.

* * *

I think it's fair to call the study "flawed." If only because "tendentious" is too long to fit even one of my headlines.

Submitted by hipparchia on

for looking at where medicare is spending money. it's lacking some significant information if you really want to try to figure out the why though.

Submitted by hipparchia on

podcast, in which several aspects of health care are discussed, including the dartmouth atlas and wennberg's original work.

Bryan's picture
Submitted by Bryan on

So basically, we are structuring a system designed to reward the hospitals that hire the most batshit crazy staff, who go around and kill patients in the first couple of days. rather than waiting to see if they might recover.

Logic and reading comprehension are not highly prized among the political class, so we have to expect these little problems.

Submitted by hipparchia on

you crack me up, dude [made my day to read that, too, though i wish i'd been on the ball enough to think of it myself]. there's a counter to that though -- monetary penalties for unnecessary rehospitalizations -- so you'd have to be sure to use a method that killed them quickly and cleanly.