I’m sure many of my readers might be fans of TED talks just as I am. One of my all-time favorites is Bill Clinton on rebuilding Rwanda:

Click here if you cannot see the video

Did you know that there is a TED conference dedicated to medicine? It’s called TEDMED and has just happened for the third time.

Next year’s conference will take place in Washington, D.C. (the first three meetings were all on the west coast). The conferences theme will be The Grand Challenges of Health and Medicine. As per the program prototype, they are

1. Non-compliance
2. Childhood obesity
3. The standard American diet
4. Stress
5. Fit vs. sedentary
6. The caregiver crisis
7. Isolation & loneliness
8. Medical communication
9. The role of the patient
10. Stopping hospital infections
11. Medical errors
12. Malpractice
13. Data reliability and validity
14. Wellness programs
15. Alzheimer’s
16. Informed consent
17. Best practices
18. Food and technology
19. The invisible woman
20. Terrorism and natural disasters

I love the concept of The Grand Challenges, that they are addressing the biggest issues in health and health care today. Most interesting to me is that almost half of them have something to do with the delivery of health care. Others address primary prevention, which is, of course, ideal but in many cases not feasible. I think that health care delivery is nevertheless the area where the most realizable (at least in the short term) potential lies for improvement of patients’ lifes.

There are, however, a two problems complexes that are not addressed by any of The Great Challenges, and in my opinion they might be more important than any single one of the ones currently proposed:

One of them involves access to medical care. This is a controversial topic, but there are 30 million or so uninsured people in the U.S., and many more are “under-insured”. Medical bills is the #1 reason for personal bankruptcy.
Being uninsured leads to two other problems. First, people will only go to the doctor when they feel they have to, and this will often be too late resulting in less than optimal outcomes and higher cost. The second is that many seek care in emergency rooms which are clogged and where no continuity of care can be provided.
Hopefully, many of the currently uninsured will gain access to insurance through health reform. In turn, this will lead to higher upfront costs and possible shortages in clinicians.

The second big complex that I am missing is variation in practice as well as the cost of care. More or less aggressive treatments will often not be based on the patients’ preferences but the medical “culture” in a particular area or be determined by where doctors trained. It is obvious that not just outcomes will differ but also costs. More is not always better, and there are some areas in the U.S. where both quality is higher and costs are lower. Of course, not all excess health care costs in the U.S. (compared to other countries’ health care systems) can be linked to over-/under-/mis-use.

Cross-posted at KevinMD.com

Did you know that 41% of all pregancies/births are paid for by Medicaid?

Today marks the 66^th birthday of Medicaid, the jointly funded but State-run program that supports health care for the poor. Medicaid plays a huge role in selected populations: 70% of nursing home inhabitants, 56% of low-income children and 42% of adults, and 44% of people living with HIV/AIDS all have their health care paid for by Medicaid (Kaiser Family Foundation 2009). Medicaid is much talked about these days, and the reason is of course the state of the Federal finances. For a few years now, one keeps seeing threatening graphs such as this one (where Social Security, Medicare and Medicaid eventually eat up the entire revenue, source: Heritage Foundation based on CBO data):

From this chart it’s clear that Medicare and Medicaid (14 and 11% of the current health care costs, as calculated by CMS and me) in its current form are unsustainable programs if there will be no cost control or even not cuts or if the tax rate will not change (or, of course, a combination). See the chart of the GDP versus Federal government spending; while the GDP curve is almost exponential, with a little bump around 2008-9, the government spending curve is relatively flat:

One should bear in mind that the entire U.S. health care system has been outpacing the general inflation, with the medical consumer price index being around 2% higher than general inflation averaged out over the past 30 years. So relevant question in the debate would be: What parts of Medicare and Medicaid are providing good value, compared to the rest of the U.S. health care system?

While both Medicare’s and Medicaid’s enrollment (47 and 50-58 mio., respectively, in 2010) and costs are projected to increase, Medicare’s cost have historically always increased a little faster. This might change in the future, according to a new Health Affairs article by Keehan et al. (exhibits 5 and 6, see also blog post), where absolute Medicaid growth outpaced Medicare, but absolute growth of the entire system was projected 1.1% greater than general inflation. Nevertheless, per-capita costs might roughly grow at the same rate or even lower, and at their baseline, Medicaid beneficiaries’ cost might be closer to privately insured, with Medicare’s per-beneficiary higher due to age and other factors. Keep in mind that these absolute figures due to enrollment will always be highly volatile especially in these uncertain financial times (since the begin of the last recession Medicaid added approximately 15% of enrollees).

I’ve written about the Dartmouth Atlas’ analyses on Medicare variation (here and here). The July issue of Health Affairs contained a study by Gilmer and Krocknick on the geographic variation of Medicaid costs (they varied more than three-fold, : $5,000 to almost $17,000, between States for Medicaid-only beneficiary). The authors also found that these difference can largely explained by volume of services per capita.

Back to the question where the best value might be: there will definitely be high-quality low-cost regions, as in the Medicare system. I personally expect that differences between regions will be even larger, both in terms of quality and cost, than in the Medicare data. One reason for this might be, that States run their Medicaid programs slightly differently. Another reason are the different enrollment rates (Sommers and Epstein in a New England Journal article from last November).

From the providers’ perspective, Medicaid patients do not provide value. While they might help shoulder some of the fixed cost, many hospitals claim that they are losing money on Medicaid patients (Medicaid payments are mostly lower than what Medicare pays, while private plans usually pay higher rates). From the governments’ and even the Societal perspective, Medicaid is hence a bargain compared to insuring all Medicaid beneficiaries privately. However, there is a current discussion that Medicaid beneficiaries face decreased access, watch this exchange between Representative Bill Cassidy (R-LA), he himself a doctor, and Secretary of Health and Human Services Kathleen Sibelius (here’s the New England Journal article that they mention):

If you cannot see the video, click here.

Finally, tackling the small proportion of “high cost” beneficiaries in Medicaid, Medicare and those eligible for both programs might be the best shot we have. In 2002, according to Kaiser Family Foundation’s Medicare Chart Book, just 7% of Medicare beneficiaries were responsible for more than half (53%) of Medicare’s expenditures. If we give these folks a medical home with better care coordination between the many providers that these multimorbid patients have, we might able to even save some money.

The new issue of the health economic journal “Value in Health” features a very interesting article by Backhouse, Wonder, Hornby, Kilburg, Drummond, and Mayer on whether or not an early dialogue between a manufacturers and health technology assessment (HTA) Agencies is worthwhile an feasible. HTAs are conducted to give decision makers information about long-term clinical and economic effectiveness of a medical technology.

I cannot write to much about it here since I’ve also written an editorial on this study. However, since many pharmaceuticals, biotechs, and device companies re-position themselves to be better prepared for the “fourth”, the reimbursement hurdle, I wanted to make sure that my readers are aware of this pilot study. Notably, the other three hurdles safety, efficacy, and quality are either global or regional, while reimbursement is regional or even local. This article compares the presumed requirements (although they’re all very tentative) of seven major HTA agencies around the global; unfortunately, the HTA agencies are not named in the article. The manufacturer in this case study was Novartis and the presumed compound was a monoclonal interleukin-17A antibody for psoriasis.

If you’re an ISPOR member, login here. You might see the fulltext through your instutition’s subscription on the Elsevier or the PubMed website. Or email me for the editorial.

However you do it — Go ahead and read the Backhouse et al. article!

[ Twitter users, click here to retweet this post. You can follow me here:@ben_geisler ]

I’ve received a request from Nanyun to blog more about Russia. Unfortunately, I haven’t experienced much of the health care system there personally; otherwise I would write about that. On a personal note, I’ve just left Russia and will move back to the U.S. shortly. However, I have just been at the 16^th Annual International International Society for Pharmacoeconomics and Outcomes Research (ISPOR) Meeting, and ISPOR’s Russian chapter seems very active. As stated by ISPOR’s outgoing president Dr. Scott Ramsey, “ISPOR is recognized globally as the authority for outcomes research and its use in health care decisions towards improved [individual or population] health”. ISPOR’s new mission statement reads: “to increase the efficiency, effectiveness, and fairness with which available health care resources are used to improve health. There was a presentation by the Russian chapter on issues on prospects of the Russia’s health care system. Unfortunately, the presentation was in Russian – so I didn’t end up going (I regretted not having learned Russian in the past academic year). There was also a systematic review poster presentation on “The State of Health Economics and Pharmacoeconomics”. The paper seems also to have been published as a full publication in the Journal of Pharmaceutical Health Services Research in September 2010 (huh?). The paper concluded that (1) few health economics research studies on Russia were published in English; (2) their quality was related to intervention type (pharmaceuticals>non-pharmaceuticals) and [academic] training of the first author; and (3) that “measures are needed to promote the commissioning of more and better-quality health economics and pharmacoeconomics studies in English”. More info on pharmacoeconomics and outcomes research in Russia is available here.

The New York Times of today (as this post goes up, it might already be yesterday’s as I am a mile high writing this) contains an op-ed by Paul Krugman with the intriguing title “Patients Are Not Consumers.” Here’s an open letter to him:

Dear Dr. Krugman,

As a comparative effectiveness researcher, I agree with many of the arguments that you are making in your most recent article. However, I disagree with your main point that patients are not consumers.

I understand your sentiment and agree that patients are more than consumers just as doctors have more obligations than their own (financial) good.
However, it is not just House Republicans, as you seem to suggest, that favor “consumer-based” medicine. Business economists Michael Porter and Elizabeth Teisberg base their entire theory to shift competition from health plans (you mention the failed Medicare Advantage idea) to individual clinicians and facilities based on the patients’ informed decisions in the market.
Patient-driven medicine is also more than just vouchers for seniors. Just look at e-patients like Dave or the Society of Participatory Medicine or former hospital CEO Paul Levy who rightly point out that patients are the most under-utilized resource in medicine and that health care process should not just be patient-centered but patient-driven.

Clinicians’ help to patients to navigate the health care system is essential. Patients will often be guided, for instance, in referrals to specialists by their primary care physician, if they happen to have one. America’s health care system, unfortunately, is so fragmented that in reality patients often need to find their information on their own, at least at some point along the way, or through families and friends.
I feel strongly that patients should essentially be engaged and empowered and have access to a wealth of information, from medical conditions to outcomes of individual medical providers or facilities.
Comparing and selecting providers of clinical services (yes, I would call doctors and others that) even if they might not pay the entire bill is an essential feature of engaged and empowered consumers.

Nevertheless, health care will never be a perfect market. Patients need to put difficult-to-digest information into context to be able to make use of them. A discussion with the clinician of your choice is most often invaluable and often cannot replace other ways to interpret information. Since you, Dr. Krugman, are an economist, I am wondering why you missed out on naming the imperfect condition in question: information asymmetry.

Yours sincerely,

Ben Geisler

[ Twitter users, click here to retweet this post. You can follow me here: @ben_geisler ]

While flying Delta, I spotted this ad in their corporate magazine:

“Survival rates. Quality of life results. These are things you need to know. That’s why we publish our [...] results and statistics at [our website].”

On their website Cancer Treatment Centers of America, a for-profit cancer hospital and outpatient clinic chain operating in six states, also gives you data on patient experience and on speed of care. However, their survival data compares to SEER data, and they do not adjusted; so basically, this is meaningless. I also had not heard of their quality of life measures, physical and emotional health scores. They seemed to be some kind of visual analogue scale. The speed of care did only show their own improvements but no benchmarks.

Have you ever seen an ad like this? The wording is a bit clumsy, but I think they’re right that that’s important. If I had colorectal (bowel) cancer, I’d like to know.

However, Cancer Treatment Centers of America has a history of making false advertising claims.

Nevertheless, I think more competition and transparency (at least to some degree) will be more likely in the future than in the past. Stay tuned for more on this soon!

Meanwhile, please subscribe to the RSS feed via a reader or email.

[ Twitter users, click here to retweet this post. You can follow me here: @ben_geisler ]

Cross-posted on PLoS Medicine’s Speaking of Medicine blog and KevinMD.com

Everybody is talking about Accountable Care Organizations or ACOs these days. The reason being that the Centers for Medicare and Medicaid Services or CMS has just released the rules for shared savings between ACOs and CMS have just been released last week. The complete proposed rule is 429 pages long and they’re seeking comments on it now until early June.

There is a lot of articles on ACOs in the news these days. Don Berwick, the current CMS Adminstrator (whom Iwrote about earlier), wrote in the New England Journal. Mark McClellanand and  Elliott Fisher (who were instrumental in panning out how exactly ACOs should be constructed) wrote in Health Affairs’ blog.
The best one that I’ve found so far is by the independent health care journalist Merrill Goozer’s GoozNews.

What is an ACO?
ACOs will take care of Medicare patients. An ACO is supposed to be formed by a group of providers, hospitals or physician groups, or both, so there will be no “middle men” that could manage the providers (and siphon off more money!), as was the idea in the Medicare Advantage program.
The ACO model is not just another attempt to curb costs, but ACOs will also be held accountable for the quality. Other than that, CMS will not regulate them much. The idea is that different ACOs can develop different models of organizing and paying for care, as long as they meet the budget and quality goals. Decentralized accountability and leadership with (monetary) sticks and carrots is likely to produce better results for the whole country than central rules with no enforcement: If ACOs incur too many costs or do not meet the quality targets, they may have to forgo payments or even pay CMS money back; if, on the other hand, they meet or beat their goals they can get bonuses.

How much money will this save CMS?
CMS hopes that ACOs could save it $170-960 million over three years, so $60-320 mio. annually. GoozNews writes that that’s a “droplet” given the $1.8 trillion budget for that period (0.01-0.05%, to be precise). However, one should bare in mind that ACOs will not care for Medicaid patients (which, supposedly, are also included in this budget, albeit most is for the Medicare program).
Also, ACOs will only care for 1.5-4 million beneficiaries, so that’s only between $14 and $213 for beneficiary per year.
If start-up costs in the first year are between $132-263 mio., then the savings in years 2 and 3 would already be $126-438 million or $32 to $292 per enrollee. The hope is that ACOs will pay off more years after they’re introduced, I couldn’t find anything on projections. As Elliott Fisher and others wrote last year in Health Affairs, the ACO model relies on quality improvement that will eventually lead to cost-savings system-wide, and on “reliable and progressively more sophisticated performance measurement, to support improvement and provide confidence that savings are achieved through improvements in care”. The performance measurements (or standardized “metrics”) are not yet in place, and most are just process measures (which are though to lead to better outcomes), not actual outcome measures.

What kind of a payment models are we talking? Is this the end of fee-for-service?
The short answer is: No, it’s not the end of fee-for-service!

Some might think the ACO models means capitation for sure, but Medicare will still pay for each service a clinician or a facility provides. This means that still the more services a provider organization renders, the more he/she is paid, and, as it is often recognized, this is the wrong incentive (purely economically, why should the individual clinician provide less services if they’re paid by the number of procedures?). So the ACO payment model only differs from the traditional fee-for-service model in the following ways: it’s all about the bonuses and penalty payments. Estimates are about $800 mio. in bonuses and $40 mio. in penalty payments over three years.

As you can see, the success of a ACO will depend on how they will land within the “sweet spot” of not cutting too many services and getting enough bonuses and if they can influence their decision-makers (the individual clinicians!) to provide only reasonable services and referrals that will increase quality.

Why should an existing provider organization become an ACO and who will apply?
ACOs will have a minimum size of  5,000 “ensured lives”. So this not for a small physician group. Rather, the first systems that will apply and be approved as ACOs will be already integrated health care systems. Physicians in those organizations will likely already be used to working together closely and there will be a focus on primary care. In some, physicians might be salaried and make only as much use of hospital stays, referrals, tests and procedures as medically necessary. Having said that, I realize that medical practice varies heavily and is influence by the local or sub-regional “medical culture” (as proven by the Darmouth Atlas and eloquently described by Atul Gawande). Therefore, if supposedly already low cost/high quality regions like

(as identified by the Institute for Healthcare Improvement) will have the first or the most ACOs, the potential savings for CMS might rather be small while ACOs will or will not be paid slightly or a lot more. The latter depends on if the estimates of CMS will be regional (or even sub-regional) or national estimates. ACOs will be paid 50-60% of the savings, but if there aren’t much savings in a given region, than ACOs are stuck with their costs.

Will this work? What are other consequences?
For CMS, it’s almost a win/win. If ACOs won’t incur savings, they will have to pay 7.5% of the costs above the “expected” costs in the standard model (or 10% if you’re shooting for the 60% bonus). So CMS will only lose if the ACOs’ patients will cost more than 107.5% (or 110% for the more risk-seeking ACOs). Of course, these figure do not include the costs of setting up and running this system. The hope is, I guess, to change the entire or a large proportion of the system, and not just <1% of the entire budget.

For the provider organizations, this is a huge chance. Currently, most hospitals just break even with Medicare patients, some lose money on them. Throughout the country you could already see that independent practices and small physician groups were acquired. If you improve your operations in an integrated system, set up quality improvement and potentially reduce the usage of some unnecessary tests and treatments, this will also have repercussions for how your organization treats other patients. There will be some investments, but the pay-off will also come from non-Medicare patients.
Nevertheless, this is not for everybody. If the provider organization is more fragmented, if there is weak physician leadership, if there is a high usage of specialists in the system and they’re independent, it will be difficult to get everybody on board to form a high-performing ACO.

For patients, this is a mixed blessing. You could argue that in some ACOs there will be restrictions for seeing specialists and other services, something that patients usually hate. However, if quality and outcomes improve in the ACO model, this would be highly desirable. The first thing that will happen is that Medicare patients will be sent a notice, and if they don’t want their information to be shared in the ACO, they can actually opt out. I’m just wondering: how will this play out politically?

[ Twitter users, click here to retweet this post. You can follow me here:@ben_geisler ]

I’ve just published an open-access chapter in a book on “Applications of Monte Carlo Method in Science and Engineering”. My chapter is on how to automate a certain kind of simulation in a standard modeling package.

Certainly this book chapter is only relevant to you if you are building your own decision-analytic models in a software called TreeAge Pro.

So for the geeks among you:If you need 1^st-order Monte Carlo simulation for your model (for example, because you use tracker variables), you can’t just run deterministic sensitivity analyses. For each single value across the range you’re looking at you need to run another 1^st-order.

I’ve written a script which does that for you, from an Excel spreadsheet for example. The code is in Visual Basic and it comes a pre-populated application for one-way sensitivity analysis.

This work builds on what I’ve worked on at Mass General with my advisor there, Alex Göhler, and others. We’ve previously published a short note paper on the technique in the journal Value in Health.

Here is the complete reference in the form that the publisher proposes (just in case you need to cite me
Benjamin P. Geisler (2011). Automating First- and Second-order Monte Carlo Simulations for Markov Models in TreeAge Pro, Applications of Monte Carlo Method in Science and Engineering, Shaul Mordechai (Ed.), ISBN: 978-953-307-691-1, InTech, Available from: http://www.intechopen.com/articles/show/title/automating-first-and-second-order-monte-carlo-simulations-for-markov-models-in-treeage-pro

You can also check out the book chapter on my website: Research-Automated Sensitivity Analysis.

I am considering several viewer applications to use on my website and ask for your help in a poll here: what do you think is the best viewer? I will use the one that you think is best!

Cross-posted on KevinMD.com

A recent N.Y. Times article on how to value a life drew almost two-hundred heavy-handed comments. It discussed how different governmental agencies such as the Food and Drug Administration (FDA), the Environmental Protection Agency (EPA) or the Department of Transportation (DoT) place a monetary value on each life saved.

In many public policy areas, Cost-benefit Analysis (CBA) is being used to assess whether an investment in a particular area is worthwhile. CBA uses an “exchange rate” in which the consequences are monetarized.

The article mentioned the following values: The DOT value each life saved at or around $6 million 2010 USD$; $9.1 million 2010 US$ was the corresponding value of the EPA; and the FDA put a figure of $7.9 million 2010 USD$ (increased from $5 million in 2008 USD$) on each life saved from cancer death caused by cigarettes.

I did not know that the FDA considered efficiency measures such as money per life saved at all.

What I find fascinating is how arbitrary the approach of the different agencies can be. They could have just funded certain policies by how cheaply they can save a life up to a certain threshold (e.g. when the budget is exhausted).

Instead, the EPA uses a methodology derived from logging industry (yes, you heard right). $1,000 worth of extra-work for the lumberjacks each year is generally accepted to save 1 in 1,000 lumberjacks. This was apprently developed by a Professor Viscusi who wrote his first paper on CBA as an undergrad at Harvard in the 1970s.

Other governmental agencies seem to survey citizens. In economics, this could actually be considered a valid approach if done right.

There were interesting comments by the readers. Some did not want to put any value of life. It was controversial if the value was too high or too low. One reader mentioned that the Federal Aviation Administration might have had a value of $450,000 per life in the late 1970s.

I think the article misses a few things:

First, it does not mention that other countries’ “exchange rates” are much lower. A life was valued at £1,312,260 ($2.14 million in today’s exchange rates) in 2003 in the United Kingdom (UK Department for Transport (Highways Economics Note No. 1. 2003 Valuation of the Benefits of Prevention of Road Accidents and Casualties. London: UK Department for Transport, 2004.) and at €2 million ($2.76 mio. in today’s exchange rates) in the France of 2000 (Boiteux and Baumstark 2001).

Second, the article does not mention that there are alternative approaches, which exist, for example, in health economics. Unlike CBA, where costs and consequences have a common denominator, Cost-effectiveness Analysis (CEA) expresses efficiency in a ratio (costs are divided by effectiveness, e.g. in life years or quality-adjusted life years). CEA was developed by Milton Weinstein in the late 1970s at Harvard. Policy makers can compare the relative values of each strategy with alternative health care investments available to them in order to make informed decisions.

The other day I was chatting with Jeffrey Levin-Scherz, MD MBA about the health care reform bill, the recent repeal by the Republicans in the House, its implementation, and its costs. Jeff teaches a course on health care costs at Harvard and writes very thoughtful blog, Managing Health Care Costs. As one of my professors, he inspired me to start blogging about my topic, the value of health care.

There’s the question if health reform will turn out to be more expensive than anticipated, if it will still pay for itself (by creating better markets, lower unit costs, less expensive emergency treatment etc.) or cost the taxpayers extra, if a possible repeal will be more expensive etc.

The other side is 30 million or so people who did not previously have coverage but presumably will.

It never occurred to me (until now) that you can not just look at the cost (do a cost or “cost minimization analysis”). You also need to look at the benefits or effects. And you could quantify and compare them.

So what kind of study would that be?
As I said, a Cost-minimization Analysis looks just at costs. It does not take (health) outcomes into account.

Cost-benefit analyses are widely used in public policy. Health or outcomes are monetarized. So for each live saved you would attach a $$$ value to the balance sheet.
How would health reform look then? Will it save lives? I think it definitely will! So you would need to throw those opportunity costs in.

An even better way to look at it would be a Cost-effectiveness Analysis (CEA). CEAs were introduced to medicine by Milton Weinstein (Harvard) in the late 1970s. You take the ratio of incremental costs over incremental effectiveness (incremental cost-effectiveness ratio or ICER). Effectiveness can be expressed in all kinds of ways, e.g. as life years gained, ulcers healed etc.
Cost-utility Analysis (CUA) is a special case of CEA: the effectiveness is expressed in the difference of quality-adjusted life years (QALYs) gained.

For CEAs/CUAs there are willingness-to-pay thresholds, so a $$$ value that society would be willing to pay for each life year (again, these are ICERs). In the UK, this threshold could be at 30,000 £/QALY; in the US it might be 50,000 or 100,000 $/QALY.

So what is the ICER of health care reform? Will it be under the willingness-to-pay thresholds? Will it even be cost-saving?