The ACO Failure Hypothesis: Likely But Not Inevitable

We recently participated in a program at Columbia Business School’s Healthcare Program on whether ACOs (Accountable Care Organizations) will fail. For those of you that don’t know, ACOs are one of the structures promulgated by PPACA (aka Obamacare) designed to encourage better cost control and quality improvement in the healthcare system.

The current zeitgeist among the commentariat is that ACOs will fail (examples: here and here). We think the reason for the one-sided nature of the question is that those of us who lived through the healthcare upheaval in the early and mid “90s” saw first hand the failure of PHOs, PPMs and IDNs (and all of the other acronyms now relegated to the dustbin of history). When ACOs are touted as a saving grace for the system, you can almost hear the collective groan of the industry veterans.

Ever the contrarian, however, we took the side of the debate that said ACOs will NOT fail. The premise of our argument was that since we already have a good idea of why the structure will fail, we can, a priori, fix the shortcomings, and though likely, ACO failure is not inevitable.

There is an extensive list of why list of why ACOs will fail. We put them into four general buckets.

Infrastructure: The system has mis-allocated resources so we have too many of some things and not enough of others leading to inefficiencies.

Technological/telecommunication: For a number reasons the healthcare system has not adopted technology as fast as other industries.

Cultural: Providers are habituated to fee-for-service payment mechanisms and patients aren’t likely to change their own healthcare behaviors.

Inertia: The well known system problems (e.g. asymmetry of information, the Pareto nature of patient demand, unexplained variation of care, counterproductive incentives) have been around forever and are difficult to overcome.

Because we spend most of our time identifying private healthcare companies with investment potential, we often get a view into what is happening in the entrepreneurial space under the punditry radar.

One approach that we think has a lot of merit in terms of correcting the healthcare system’s deficiencies, is re-purposing the existing healthcare infrastructure. We are working with companies to raise capital to take older facilities/systems and update them to more accurately reflect demand in the market place. More efficient utilization of the existing asset base demands far less capital than building entirely new operations. And, for those who have optimally configured operations we are providing a growth capital option to expand. Interestingly, we are also observing the repurposing of therapeutics as well. It turns out that if we apply new technology to older drugs/devices it can represent a less expensive, but equally efficacious alternative.

In terms of technology, we are seeing a number of innovative solutions utilizing inexpensive mobile devices to transmit information to appropriate locations. Maybe because of the ubiquity of devices in almost all other walks of life, it appears that there is greater receptivity (than we would have expected) on the part of virtually all the stakeholders to increased technology interaction. For providers we have encountered companies providing superior analytics to prior generations of IT, which allow for better real-time financial and strategic decision making leading to a better handle on utilization and costs.

Culturally, providers, for lifestyle reasons, seem to want to move away from small group settings and team up with larger employers. In terms of patients we are seeing companies use techniques borrowed from behavioral finance to move patients to more health behaviors. On the business side, systems are becoming better at identifying “frequent fliers” allowing for more intense interventions which in the long run saves money and also provides better care.

In terms of inertia we are a bit stuck. We hope that some of the carrots and sticks in PPACA will be effective. However, in our judgment what will eventually move the needle is transparency of data. However, we think measurement (we know about the shortcomings) is a good thing and it is the only way to hold everyone “accountable.”

We are not suggesting there is a magic bullet. And, in fact we would not be surprised to see abundant failure in ACOs (don’t even get us started on insurance exchanges). However, until someone proposes a better mousetrap it looks like we are going to have to make do with this one. So, we might as well attempt to fix failure before it becomes inevitable since we have tools to make it work.

Les Funtleyder is the managing director of Poliwogg, where this post originally appeared.

14 replies »

  1. Dan, to your question, how can ACOs be tweaked, I think the real question is how ACA can be tweaked that will effect ACOs? In a nutshell, the challenges ACOs face under ACA are:

    1. Regulations are too immense – see centrally planned economies for results. 2. ACO administrative expenses may exceed 50% of total health care dollars – it is already estimated at 40% today. 3. There will be a number of exchanges set up, but not all exchanges will be equal or the same. State and Federal exchanges will focus on mechanical issues for the first couple of years. They will likely struggle with a number problems, driving up the cost of their administration and premiums. Private exchanges run by insurance companies will do their best to comply with the law and will not be able to discriminate individually. But they will be able to make legal, subtle shifts on a group basis that may allow them, of all things, to reduce their administrative expenses, but still maintain a 85% MLR, even as their premiums increase rapidly. And then there is the concept of independent private exchanges that meet the Reform specs, but operate a little differently than public or insurance company exchanges. We have noted, as have others, that there are a number of gray areas within the law. We have not explored all the ways these private exchanges might work, but it appears that they are capable to under cut the high cost of complex ACOs as currently designed within the ACA.

  2. Ah, yes, I have been following this Geisinger model since 2007, which if you measure it in technology time, is the first year the iPhones came out. I think the pilot is interesting, whether it has “market legs” is another thing. My comments were more focused on the broader health market and whether ACOs, in general, will succeed or fail. Over the last two decades, we have seen hundreds of pilots, operating in “closed systems.” I liken these to technology “closed systems” similar to the IBM mainframe bundled software prior to 1969, mini-computers, pre-DOS microcomputers, or even Epic’s EMR systems. For whatever reason, (lack of viability, regulations, self-interest, etc.), these health care pilots struggle to get out into the open market, or at least what passes for the open market in health care today.

    For these pilots, ideas, and concepts to flourish, health care needs to have its micro-computer “DOS moment,” which may or may not be possible within health care reform. Whether it was a stroke of genius or dumb luck, Microsoft was able to set an open industry standard for something few of us cared about, but it delivered the capabilities for the rest of the PC revolution. Technologically, DOS was a pretty simple piece of software bought from a very small third party that originally called it the Quick & Dirty Operating System or (QDOS). Microsoft wisely rename it Diagnostic Operating System, which diagnosed nothing, but sounded very sophisticated. Apple kept their proprietary operating system (which BTW was technologically superior to DOS) and continued to be a minor player until it opened up to the PC world. While Apple’s i-series is built on proprietary systems, people could freely adopt their system, or not. But even now Apple is threatened by another open system – Android.

    What we need to do is figure out a way to get these pilots, ideas, and concepts out of their “closed systems” into the open market.

  3. This is an interesting post. And I’d love to hear more – whether from the author, Jeff Goldsmith, or others – on how the ACO concept can be tweaked to ultimately succeed.

    If care transformation was easy, wouldn’t we have done it already?

  4. Geisinger has built and tested both model- ProvenCare Navigator, which is the relationship-based part and ProvenCare Acute, which is the complete clinical solution/bundled payment part. They’ve gotten extensive physician buy in, including physicians that are not part of their group, and rave reviews from patients. They are working to commercialize the innovation, and consulting widely, both in the ACO universe and outside, to teach other health systems how to do it. They are good business inside and outside a population health model.

  5. Jeff, I like your model. The challenge is, like so many models before, has been getting the idea from the drawing board into real life. As your report mentions, there is so much distrust, not just between physicians and hospitals, but also between every party involved – consumers, insurers, government, consultants, and health technology companies. One company I am working with is trying to implement a network business model with some of the concepts you have outlined. The providers who are against health reform distrust the model because the network, in part, works within ACA and tries to make it work better. The ACA-believers dismiss the network because it isn’t ‘true blue’ enough. It is often hard for decision makers who have other priorities, to want to adopt your models or any other model that might work better. But a good sustainable model should work within or without reform. It seems the only way to prove that a model works is to make it work out in the real world, which is what Trucaris is setting out to do. I would like to connect offline to see if we could incorporate some of your ideas.

  6. The answer to the complex, high intensity Medicare patients is that will be key players in both. They are, along with the urban poor, the prime candidates for a medical home model, but they also will be the major patient populations in the “complete clinical solution” business when intervention is required.

    To offer a complete clinical solution for an acute problem of patients with multiple chronic risks, where the patient doesn’t come bouncing right back to the hospital, will require protocols and pathways for managing the comorbid conditions that most threaten them.

    And for some, the dominant, life changing chronic disease risk will require a specialty oriented medical home where the core specialist is the main doc in the center of the care model.

    The Blueprint referred to above argues that we need the health system to offer two basic “new products”: a relationship based primary care model and a fixed price “general contractor” for complex interventions. You can leave the rest of the care system functioning in fee for service mode if it occurs inside an envelope of physician managed global risk, hopefully more rigorous than the ACO.

  7. Jeff, I took a look at your ACO article. In general I think you have a lot of good ideas, but am trying to figure out where in your model the complex Medicare patients with multimorbidity would fit in. Do they need ” long-term, low-intensity primary care” or the “major clinical intervention” track? What is the right model for managing a cancer crisis or big surgical intervention in an older adult with multiple chronic illnesses?

    My own prediction is that ACOs will fail when it comes to the older complex Medicare patients. Unless they require all those patients to sign up for a well-run integrated system a la Kaiser…

    I would certainly like to see older adults being given more a chance to be involved and decide how healthcare dollars are going to be spent on them. I know many who would prefer to have more services in the home and would give up the annual echos.

  8. While I completely agree with Jeff’s observation that a patient/consumer’s involvement in his or her own health care is essential if there is any hope of getting industry-wide spending levels constrained and quality increased, there is data to suggest that this is a huge hill to climb. While the results are somewhat dated, a number of studies emphasize the magnitude of the challenge: in a Center for Health Systems Change study, in general fewer than half of the patient participate “actively” in their care—even for those with chronic illnesses, other than cancer and asthma which run a little over 40% active participation, none of the other major chronic illnesses even reach the 40% level. In the same study, it was reported that as patients age their level of active participation diminishes (no surprise)—even as they presumably encounter more and greater health issues. An American Diabetes Association study some years ago (cited in a recent Atlantic article) found that only 5% of diabetics monitor their status daily, and 65% of those on drug therapies did so less than once a month. In an AHRQ study from several years ago, 36% of the population surveyed had basic or below basic skills (understanding instructions, able to understand a pamphlet of information) at the low end of the scale, while only 12% actually had the capability to actually participate in the management of their own health care. A recent BusinessWeek article noted that according to a NIH study, “one in five patients fails to fill new prescriptions, and half of those being treated for chronic conditions stop their medications after six months”. All of this data suggests that the challenge of having consumers actively involved in their health care is huge, notwithstanding the point that if we can’t find ways to increase their (our?) involvement, we will be struggling with cost and quality issues for a long time.

  9. Managed care without risk (the risk of cash losses, not just the risk of bonuses not earned) is like a gin and tonic without the gin. It’s the fixed pool of funds available to pay everyone that forces you to make choices. The vast majority of ACO’s are “one sided”, e.g. focused in reducing the rate of increase).
    And the prevailing fantasy among providers is that they can mitigate the effects of lost upside revenue by taking marketshare from others, or reducing post acute expenses. Fool’s gold, I believe this is called . . .

    But a more grievous design flaw is the non-role of the patient. Enthoven and others were right: patients need to be active agents in the managed care equation, not just “lives” attributed statistically. What patients got out of the HMO movement were cash savings: lower premiums and reduced cost sharing. The ACO is fundamentally paternalistic: it’s the care system doing things TO people presumably in their interest, and then keeping the savings that result (if any). ACO’s do things TO people, not for them or with them. It isn’t enough to be “patient centric”. Patients need to be active agents in improving their own health, and to get a meaningful share of the savings.

    This was an idea borne in academia by people with limited experience with the real world of managed care or healthcare markets. It did not deserve the huge commitment of resources it’s getting.

    For for alternatives, please see: http://www.healthfutures.net/pdf/Accountable_Care_Organizations.pdf
    as well as the forthcoming Policy Blueprint I prepared for the Physicians Foundation.

  10. People love to tear down rather than build something. ACO’s may not be sliced bread, but please recall our current ‘system’ – is that not the very definition of failure? As noted in this piece, they may be but a step along the way, but we had to move the needle somehow. The status quo was, and is, unacceptable.

  11. Everyone is saying that ACOs are going to almost universally fail but what exactly is the threshold for measuring success of these groups? Are people going to seperate out hospital-led vs hospital-physician vs physician-led ACOs?

    More interested in teasing out the overall amount of risk that hospital systems/IDNs are going to face (e.g., VBP, HAIs, etc) and how they are going to handle this across the enterprise.