What Does the Dartmouth Atlas Have to Say About the Politics of...

What Does the Dartmouth Atlas Have to Say About the Politics of the ACA?

60
SHARE

Healthcare reform was a frontline topic during the recent presidential elections. The political warfare and misleading information around the Patient Protection and Affordable Care Act (PPACA), also known as Obamacare, has prevented the public from understanding its intended purpose, and has left many skeptical about its benefits. It is safe to say the general public has little to no idea about the quality of healthcare delivery in their respective regions.

In fact, it is not a far cry to claim that even healthcare professionals might not truly understand the issues facing American healthcare. Thus, most of the public is generally uninformed or misinformed about the population level problems facing the healthcare system. Therefore, it is quite simple for political parties to misguide the public and capitalize on their uninformed perceptions. If the public knew more about the flaws present in the healthcare system, perhaps they would better realize the PPACA is a reasonable start at addressing the failings of our system.

The Dartmouth Atlas Project is an online database which collects Medicare spending and utilization data from around the country. Information gathered from the database has shown immense variation in the way medical resources are utilized by even similar regions, communities, and health care organization. Evidence has repeatedly shown that, from a population perspective, areas that spend more on medical care do not consistently benefit from increased quality of care or patient wellbeing. Variation in the type of care delivered can be attributed to diverse incidence and prevalence of disease severity or the type of care a well- informed patient chooses. Variation in health care delivery is thus omnipresent and expected, because every patient is unique and medical innovation presents a growing number of care options to choose from.

However, much of the variation in healthcare practice is “unwarranted” because it cannot be explained by the degree of illness or patient preference. In fact, the two main drivers of unwarranted variation are the capacity of the local health care system to provide growing number of expensive services that must be utilized, and the physician’s practice habits that may not be evidenced based or patient preferred. The current healthcare reimbursement model propagates variation in care delivery due to the financial incentive of providing more service even when little benefit exists for the patient.

Analyzing Medicare data in a political context, the table below displays spending and service utilization data for the top ten Republican and Democratic states, based on the elections polls that tracked the popularity of each presidential candidate during the race (election polls).

These data tells a clear and simple story. The top ten Republican states have higher Medicare spending than the top ten Democratic states. The rate of hospitalization and surgical procedures are also higher for Republican states. If we investigate a procedure like percutaneous coronary interventions (PCI), the Republican states are performing more PCI procedures with equal mortality benefit compared to Democratic states. The evidence of variation in cost and utilization is a strong indication of inconsistency and inefficiency in the care delivery process. Are the Republican states providing better care by providing more care? We cannot find evidence of for such an assertion. Nor do we find evidence of harm occurring from a lack of utilization to individuals residing in democratic states.

Six of the ten Republican states sued the federal government over the individual mandate and Medicaid expansion earlier this year (Utah, Alabama, Louisiana, Texas, Georgia, and Nebraska), compared to only one democratic majority state (Maine). Yet the Republican states have a higher average of uninsured people, thus inhibiting a greater percentage of their citizens from accessing preventive healthcare. It is possible to draw many conclusions from these data, however it is intriguing that the states that have higher spending and resource utilization supported a Republican candidate who was not a supporter of the PPACA. The intrinsic values of the reform act are to cut down on waste, ensure access to preventive care, pay providers for quality rather than quantity, and reduce unwarranted practice variation and disparities by promoting accountable models of care delivery. Regardless of political stance, minimizing unwarranted variation is an ethical priority and a solution to decelerating the growing of healthcare spending in the United States. Hopefully, the continuing implementation of Obamacare over the next four years will prove to be beneficial.

Dr. Anubhav Kaul is a recent medical graduate from Ross University School of Medicine, and he is pursuing a Masters in Public Health at The Dartmouth Institute of Health Policy and Clinical Practice. Thom earned his PhD from The Dartmouth Institute for Health Policy and Clinical Practice where his dissertation focused on understanding variation in the costs and utilization of care within and between hospitals.

Leave a Reply

60 Comments on "What Does the Dartmouth Atlas Have to Say About the Politics of the ACA?"


Member
Dec 27, 2012

interesting idea.

would like to see data for all states.

where does Florida rate? North Carolina? New Jersey? Arizona?

I’m not sure looking at how a state was leaning during a presidential election is the best metric to use here, although I get why you’re using these numbers. I’d be more interested in knowing the affiliation of the governor (better) or some sort of comprehensive political scoring system (even better) …

/ j

Guest
Bill
Dec 27, 2012

The reference to the contradiction re states rights is fascinating. “Get the federal government out of our local affairs” gets punted to the federal government by Governors who didn’t like the turnout of the November election.

One of many contradictory messages in today’s political landscape. “Get the government out of my life” becomes “except in bedrooms.”

Guest
bird
Dec 27, 2012

“Evidence has repeatedly shown that, from a population perspective, areas that spend more on medical care do not consistently benefit from increased quality of care or patient wellbeing”

So which comes first, the poor health and wellbeing causing those locations to spend more or does spending more cause poor health?

Guest
Dec 27, 2012

Interesting information. Lots of potential confounding variables, for example what is the correlation to socio-economic status, education, ethnicity? These have all been shown to impact total healthcare spend and also may correlate to political leaning. Without a robust matched cohort and multiple regression analysis, not sure the simple conclusions stated in this post can be supported.

In addition, the observation:

<>

Lacks causality. For example, how do we know that without the higher PCI utilization the mortality would not be higher in GOP states?

Guest
Dec 27, 2012

Oops..meant to insert the following in between marks :

“If we investigate a procedure like percutaneous coronary interventions (PCI), the Republican states are performing more PCI procedures with equal mortality benefit compared to Democratic states”

Guest
Anubhav Kaul
Dec 27, 2012

Thank you for the comment. Your point is absolutely valid. However the data does adjust for age, race and sex at the least.
I would like to clarify that the data is not exact science and the conclusions can only be hypothetical. The main point I would like to get across is that just like how the public knowledge about the ACA is lacking, so is knowledge on how one state utilizes its medical resources compared to another. If we can control this variation, it would answer a lot of problems.
Idealistically, I would prefer the healthcare debate to not be political, but it just is.

Guest
Dec 27, 2012

Interesting way to look at things, but why this consistent assumption that nobody other than “experts” understands the issues facing our health care system? Or is it that those who disagree are by definition not understanding?

The Dartmouth data has all sorts of question marks attached to it and its value has been debated often enough, particularly after Medicare revised its own numbers to pretty much conflict with the policy defining tale of two cities (McAllen & El Paso). We do have variations, but it is not completely clear what those are, and even less clear why we have them, and anybody’s guess whether we should have them.

Following the Dartmouth philosophy on recommending change, I guess this article should recommend that in order to reduce costs, everybody should vote Democrat….

Guest
Whatsen Williams
Dec 27, 2012

Based on the NY Times front page report, the computerization attempt to create vanilla care and eliminate variation has backfired, dramatically running up expense without improvement in quality. Dartmouth Atlas has promoted this cost ineffective system of care, going from bad to worser at quicker speeds.

If the EMRs were any good, outcomes would improve and costs would decline. What do the Atlas folks have to say about that, and about the newly minted safety program for HIT from HHS?

Guest
legacyflyer
Dec 27, 2012

The Dartmouth data is very interesting data and I think we should look at it carefully to see what kind of conclusions we can draw.

First, it is clear that the practice of medicine varies from one part of the country to another – and unless our population varies in parallel – someone is not practicing EBM or the “evidence” is not as clear as we might like.

A simplistic analysis would look at data on; angioplasty, hysterectomy, C-Section, etc. and conclude that places that have high rates are doing too many. Of course the other way to look at it is that places that have low rates are doing too few. And of course nobody would say that regions that have higher immunization rates or better compliance with screening guidelines are doing too much compared to those with lower rates.

A more sophisticated analysis would look at (for example) angioplasty rates and compare them to mortality from coronary artery disease. Again, if an increased usage didn’t produce any improvement in mortality, one might reasonably say that the increased angioplasty was “unnecessary” or not “evidence based”.

The problem is that mortality is not the only endpoint. Take the example of differing rates of total knee replacement – let us suppose that they are more commonly performed in the Northeast (generally high usage) than they are in the Upper Midwest (low usage). How do we evaluate who is doing too many and/or who is doing too few? There is no mortality data associated with this (other a small amount of peri-operative mortality), the desired endpoint is mostly symptomatic and subjective. An older patient with bad knees becomes more mobile and is in less pain. So are we doing too many knee replacements on the East Coast or too few in the Upper Midwest? And there are a whole lot of other treatments/diseases that fit in this category.

The other statistic that might be interesting to look at in parallel to the Dartmouth data (and maybe this has been done – I have been to the Dartmouth website, but I can’t say I know it well) – is how the number of doctors affects the rate at which certain procedures/tests are done. For example, the reason that the upper Midwest may spend more is that there are fewer docs. And when there are fewer docs, maybe they continue to do what really needs to be done, but cut back on elective stuff.

Finally, analysis of the Medicare data needs to be matched for a variety of factors – some of which are not obvious. I can easily imagine a primary care doc in the Upper Midwest telling a Minnesota farmer that he doesn’t need to go to a specialist or have a fancy test while a primary care doc in New York tries to tell the same thing to an obnoxious New Yorker (who has a lawyer in the family) with much different results.

Finally, one clear weakness of the post was the relationship between the Dartmouth data and the ACA. Exactly how is the ACA going to correct the problem? Is the magic fairly dust of ACOs going to cure all ills? Tell me more …..

Guest
Anubhav Kaul
Dec 28, 2012

Thank you for such great feedback. I will be the first to admit that our methodology may not be accurate; it could just be one big ecological fallacy. However, as you stated above, there are regional differences, and that has been proven repeatedly. Is the variation in medical practice a contributing cause of the healthcare crisis? One can argue that.
How can the ACA help? Well that is the million dollar question and a work in progress, but organized models of care, along with payment reform that limits supply sensitive care is a potential strategy. Not to the mention the 30% waste in the system that can be minimized. Dr. Jack Wennberg, the founding editor of the Dartmouth Atlas, who has dedicated his entire career to understanding practice variation aptly talks about possible solutions, in his book “Tracking Medicine”.
Additional perspective –
http://www.nejm.org/doi/full/10.1056/NEJMp0910001
http://www.nejm.org/doi/full/10.1056/NEJMp0909327

Regarding your comment of supply of doctors/services – One study does come to mind: http://www.ncbi.nlm.nih.gov/pubmed?term=wennberg%20variation%20workforce
In contrast, a higher supply of primary care physicians is associated with better outcomes in the Medicare population. http://www.ncbi.nlm.nih.gov/pubmed/21610242

Guest
legacyflyer
Dec 28, 2012

Thanks for those references. I will look them up and reply.

I know that you don’t want to stick your neck out (and get it chopped off) on this blog, but what do you think are the most significant factors influencing the varying usage statistics?

Guest
Anubhav Kaul
Dec 28, 2012

Down below, the discussion is looking at valid possibilities…but conceptually speaking….

The three main drivers of unwanted variation can be identified as lack of effective care, discord between provider preference and patient preference, and over-utilization due to abundant supply of medical care services. Solutions to unwarranted variation must address these drivers individually and holistically.
Variation in the use of effective care can be due to differences in clinical knowledge and differential rates at which providers incorporate new guidelines into their practice. Medical science is an evolving field and clinical knowledge must be updated regularly. For certain clinical scenarios, like the use of PSA screening and prostatectomy after age 80 years, the effective care guidelines are not established. Thus, providers and patients must come to a preference based decision based on inadequate evidence. The Congressional Budget Office estimates that up to 30 percent of care delivered in the United States goes toward unnecessary tests, procedures, doctor visits, hospital stays and other services that may not improve people’s health; sometime even causing harm. Medical societies must continue updating physicians and caregivers about best practice guidelines. The ABIM Foundation is spearheading an initiative called “Choosing Wisely”. The campaign is a challenge to various medical specialty societies, which asks them to identify five tests or procedures commonly used in their field whose use is questionable. The goal is to initiate awareness and strategies for effective use of health care resources. Such initiatives promote good practice and accountability from the provider end.
Variation in preference sensitive care is a byproduct of ineffective patient-provider communication. Providers, more often than not, tend to take a paternalistic approach towards their patients and “prescribe” care options without exploring what the patient really wants. The patient, in turn, thinks that “doctor knows best”. This is the flaw in communication, and shared-decision making is the answer. Shared-decision making helps limit provider preference and engages the patient in delineating their plan of care. The process can be enhanced by tools, such as decision aids, which can help educate the patient and allow them to make an informed choice. The goal is to achieve patient centered care and respect patient preferences. For example, the tenets of palliative care firmly adhere to these principles and should be incorporated earlier when managing patient with life threatening illnesses or at the end of life.
Overutilization of medical care is a result of supply availability and market competition, and is compounded by the above problems. Lack of effective care guidelines may lead to flexibility on the provider’s part to perform services that are instigated by a fee-for-service system, leading to non-patient centered care. The solution for this problem lies in policy. The Affordable Care Act promotes organized models of care, such as Accountable Care Organizations, to achieve quality and efficiency. Bundle payments and “shared-savings” based payment strategies will prompt providers to be more mindful of providing effective care at an appropriate cost. Providers will be rewarded for meeting quality criteria and improving the health of their patient population. In such systems, hopefully, supply will be dictated by “true” demand.
The foundational solution to all three drivers of unwarranted variation is the code of medical ethics. Regardless of scenario, providers have the ethical responsibility to respect patient autonomy, act in the best interest of their patients, “do no harm”, and approach all patients with fairness and equality. As the stipulations of the health care reform further materialize, physicians need to ignite their capacity of ethical doctoring and help their patients and the failing system.

Guest
legacyflyer
Dec 29, 2012

Thanks for your response. I was struck by how much of it was couched in “health care speak”.

Your remedies are the standard ones; an appeal to “providers” to be good girls and boys and “choose wisely” about tests and treatments and a faith that ACOs will solve the problem.

We will see how effective ACOs are and how well accepted they are by patients. Since as far as I can tell, an ACO is just an HMO with an EMR, I am skeptical.

Guest
Barry Carol
Dec 27, 2012

I don’t think politics is at all relevant in this data.

Seven of the 10 Republican states are in the south. OK is also in the southern tier but is more western. TX is both southern and western. None of the Democrat states are in the south or the southwest. Obesity rates are well above the national average in the southern states and people are poorer as well based on average income per capita even after factoring in the lower cost of living.

In the Democrat group, Hawaii is more akin to Japan than the U.S. given their population’s heritage and diet. The incidence of heart disease, for example, is far below the national average. At the same time, the large Mormon population in Republican Utah probably drives a healthier lifestyle and lower than average medical spending there. The conservative practice pattern culture at Intermountain Healthcare is also a helpful factor I believe.

The data are probably influenced by everything from differences in physician practice patterns to the amount of fraud to the malpractice litigation environment to the prevalence of academic medical centers and number of rural hospitals that operate with low occupancy rates. Like I said at the outset, politics has nothing to do with it.

Guest
Peter1
Dec 28, 2012

“Obesity rates are well above the national average in the southern states and people are poorer as well based on average income per capita even after factoring in the lower cost of living.”

“Like I said at the outset, politics has nothing to do with it.”

You’re kidding, right Barry? All of the above is politics. Would you argue that segregation in the South was also not politics? Anti union legislation in the South is not politics?

Guest
Dec 28, 2012

The Dartmouth data also suggests that those states which more uninsured working people will have higher Medicare costs as those persons turn 65.

That certainly makes intuitive sense. People who have crummy insurance or no insurance when they are 60 years old will postpone expensive care until Medicare comes in to pay for it.

Also, the Southern and Western states have a greater prevalence of beef-and-bourbon diets, at least judged by restaurant and grocery store offerings. That will create a need for more heart care under Medicare.

Neither of my observations are scientifically grounded, although I suspect they can be proven out.

In the defective American model of state vs state competition, the red states have attracted industry by opposing unions and minimizing Medicaid.
This has the byproduct of creating a less healthy working population.
Time after time, the Red states rely on the very federal spending that they claim to oppose.

Guest
Dr. Mike
Dec 28, 2012

My simplistic hypothesis: Warm states and the industrialized East have higher utilization – i.e. places where there are retirees have higher utilization. Not sure why this is surprising. Yea you can adjust for age,sex, health status, etc, but obviously there is something about treating LOTS of retirees that makes you more likely to order tests. I don’t think it is political at all.

Guest
Peter1
Dec 28, 2012

“but obviously there is something about treating LOTS of retirees that makes you more likely to order tests.”

Obviously? My take of the data is that it looks at Medicare recipients, which would assume retirees. It does not appear to total the tests, but does it per capita – that should adjust for higher number of retirees in the south/west. If anything, the northern retirees (assumed to be healthier?) would make the southern locals look better.

Anubhav Kaul will have to comment.

Guest
Anubhav Kaul
Dec 28, 2012

You are correct, We are comparing rates rather than absolute numbers.
If we had to compare total number of Medicare beneficiaries by state, you can refer to:
http://www.statehealthfacts.org/comparemaptable.jsp?yr=255&typ=1&ind=290&cat=6&sub=74
You will notice that California has the biggest Medicare population by far.
Regarding the health of individual state populations, that is obviously a questionable issue and could explain some of the variation seen.
However, it is interesting to note that New York, Massachusetts and California are states that have a high supply of medical care, yet they are able to keep the cost of care per Medicare beneficiary below $10,000.
Will the ACA help high utilizer states curb their high costs? That will remain to be seen.
Again, the culture of medical practice is not inherently political but the current solution to the healthcare crisis is politically tormented. States are opposed to the ACA just because of party allegiance. But these very states may stand to benefit the most from continued implementation of the ACA.

Guest
Dr. Mike
Dec 28, 2012

Are you accounting for the ‘snow-bird’ effect? I have heard internists complain about the mega workups their patients get from florida cardiologists – are not those patients counted as residing in NY but the cost to CMS counted as coming from Florida? I don’t think you have disproven the effect that numbers have on rates.

Guest
Anubhav Kaul
Dec 28, 2012

Regarding the “snow-bird” effect – All services used will be reflected under the location of residence (in this care NY), regardless of where the services are obtained.

Guest
southern doc
Dec 28, 2012

I think Dr. Mike makes an anecdotally valid point.

If one has a patiient mix of healthy young people and chronically ill old people, all scheduled every 15 minutes, you’ll steal time from the healthy and give it to the sick.

If all your patients have 10 medical problems to deal with in 15 minutes, it’s going to be very much a move ’em in, move ’em out situation: less time on history and exam, more tests ordered to compensate, more follow-up visits, more referrals. It adds up.

Guest
Barry Carol
Dec 28, 2012

I think the culture of medical practice should not be underestimated and can vary significantly even with in a state. For example, practice patterns are much more conservative at UCSF than at UCLA and utilization in Southern CA is generally higher than in Northern CA where, by the way, HMO’s are well accepted. Northern FL is a different (more conservative) world than Southern FL, especially Miami-Dade County where fraud is a huge issue. Utilization in NYC is much higher than it is upstate and fraud is higher as well. There are lots of academic medical centers in NYC. In MA, the Boston metropolitan area has much higher utilization than the rest of MA in part because of the prevalence of academic medical centers. Interestingly, other data show that Massachusetts has the highest per capita spending in the country across its entire population.

Separately, Bob Hertz’ point about pent up demand among newly eligible Medicare beneficiaries who were previously uninsured has some validity though it usually runs it course after the first couple of years. Medicare Advantage insurers report higher claims on average in the first year of eligibility than in subsequent years though they’re not sure how much of that to attribute to the pent up demand phenomenon

In the upper Midwest, Medicare spending per capita is well below the national average and the primary reason, I believe, is conservative practice patterns. Again, politics has nothing to do with it.

Guest
legacyflyer
Dec 28, 2012

Barry,

I think you are right about practice patterns differing.

The question is: “Why do they differ?”

Possibilities:

1) Number of docs per capita. More docs available means more mouths to feed means more procedures/tests/etc. This may explain Boston, NYC and Miami.

2) Malpractice risk. Although this has not been proven, it is certainly a plausible reason to explain increased usage. Again, NYC and Miami are high risk. Not sure about Boston

3) Intrinsic health of the population. Since many of the Southern States have higher rates of poverty and poverty is associated with poor health, this may help explain some regional variation.

Guest
Barry Carol
Dec 28, 2012

legacyflyer –

I agree with you. To your list, I would add the following:

Doctors as entrepreneurs – This is the McAllen vs. El Paso phenomenon that Dr. Atul Gawande described within Texas in on of his New Yorker Magazine articles. Both locations are similar in socioeconomic makeup if I remember correctly. I’m not sure if the litigation environment differs regionally within TX or not outside of the largest cities.

I’ve also read that a very large percentage of physician owned surgery centers are located in Texas. HCA, the largest for profit hospital chain, has half of its hospitals in TX and FL, both high usage states. Finally to the extent that hospitals and large physician practices base bonus compensation in large part of relative value units billed, there are likely to be more relative value units billed by those docs than there would be otherwise.

At a conference I attended a few years back, a Medicare expert put up a slide that showed that Medicare spending per capita was higher in large cities than in less populated areas. For the commercially insured population under 65 years old, it was exactly the opposite. There are a lot of complexities here. That’s for sure.

Guest
legacyflyer
Dec 28, 2012

Barry,

I am very familiar with the concept of doctors as entrepreneurs. Without trying to start an inter-specialty battle, the data on usage of physician owned/self referred imaging equipment vs. referrals to facilities from which the referring doc gets no financial benefit is striking.

Still, I don’t think this explains everything, or even the majority of the variance.

I still believe that most docs are sincere and trying to do the best they can within the framework of malpractice risk and (sometimes unreasonable) patient expectations.

Guest
Dec 28, 2012

Barry,
Earlier this year, Medicare has revised its data and published it:

“The Centers for Medicare & Medicaid Services’ analysis offered a very different assessment than did the Dartmouth Atlas…….
………
Hospitals in McAllen, Texas, another region of the country that Dartmouth judged as so extremely expensive that its name became nearly synonymous with excess health care use, hewed close to the national median in the Medicare analysis. ”
http://www.kaiserhealthnews.org/Stories/2012/May/09/Medicare-Hospitals-Costly-Patients.aspx

CMS still found great variations, but they were different than the Dartmouth data. The point here is that if we built theories (entrepreneurial doctors) based on flawed numbers, we should have the intellectual integrity to revisit them when the numbers are corrected.

Guest
Barry Carol
Dec 28, 2012

Margalit –

After a 40 year career in the investment business, I’ve learned to be very distrustful of statistics that come out of the OMB, the CBO and the Department of Commerce. Often it’s based on flawed methodology or assumptions and turn out to be little better than garbage in garbage out. I’m equally distrustful of a lot of the CMS data as well. I’m more inclined to trust the Dartmouth Atlas data and common sense regarding the issues in this thread. Interpretation of data is a separate issue.

Payment policy that rewards doctors with more money when they do more and bill more, perceived malpractice litigation risk, fraud, self-referrals and the standard of care in the local area along with the internal culture at specific hospitals and physician practices are all going to drive regional practice pattern variation independent of the health status and medical needs of the served patient population.

Guest
Dec 28, 2012

Well, we could talk about the data too, which I also find peculiar.
The dollar amounts are from 2009, the procedures are from 2007, and the rankings of states politics are from a random newspaper projection before the 2012 election, some of which was not quantitative, thus not rankable.
Besides the 2012 election results are in. Would it be too much trouble to adjust the top 10 list? Why aren’t Wyoming (68.6% Romney), and Idaho (64.5% Romney) in the top republican States? Where aren’t DC and Connecticut on the Democrat side?
So Barry, I have issues with both the data and the tortured analysis of it.

Guest
Anubhav Kaul
Dec 29, 2012

Thank you for pointing out the missing states. The polling results used were not quite accurate. We have made the changes and sent them to editor to update. We have also expanded the list from 10 to 20 per side to make the data more inclusive. The stated speculations still hold true.

As for the analysis….this was a quick and dirty investigation to an intriguing hypothesis. It is obviously open to vast criticism. We hope to update this article with more variables and further analysis.

Guest
Dec 29, 2012

The revisions should be interesting to read… Looking forward to the updates. Thanks.