Viability of Accountable Care Organizations, Reasonable Rules and Wise Use of HCIT

In recent days and weeks, a flurry of articles and white papers has been written in response to the proposed Accountable Care Organization regulation. In simple terms, there is broad consensus that we have great people and institutions involved in delivering care today. If we use technology wisely, subject to reasonable regulation, we should be able to improve value (access, quality, and cost), through better decision making and coordination of care.

Implementing HCIT that provides the functionality and capacity to meet Meaningful Use criteria now and in the long term will play an equally important role for those healthcare organizations that opt to participate in the ACO initiative. And we should all keep in mind that if the initiative shows even a modicum of success, payers other than Medicare are likely to adopt similar value-based reimbursement models.

In the meantime, questions have arisen as to the real world viability of the ACO initiative. The proposed regulation has come under considerable scrutiny by a number of influential healthcare leaders and organizations. To say the least, it is still a work-in-progress, but a work that’s mandated by Congress to be in place by January 1, 2012.

As pointed out by many over the past several months, including John K. Iglehart in NEJM, December 22, 2010, CMS has previously sponsored a relevant demonstration project that began in April 2005 with 10 large physician group practices participating. The project measured both quality and shared-savings payments over a four-year period.

Of note, only two of the participating PGPs received payments each year. Four received no shared saving payments, and only three achieved double-digit savings greater than $10 million cumulatively. Further, more than half of the savings were accrued by just one of the 10 participating PGPs. Where does the variation come from? Is this predictive of success or failure for your organization under an ACO? Iglehart’s article outlines the analysis and informed conjecture concerning the project as provided to CMS by RTI International. It’s well worth a read.

During the Healthcare Informatics Executive Summit last month in San Francisco, I asked two of the most informed physician leaders in attendance, “Where does this variation come from?”

I posed my question first to the opening keynote speaker, Farzad Mostashari, MD, ScM, National Coordinator for Health Information Technology, who replied:

“It’s about how you choose to play the music. The elements of Meaningful Use are like the keys on a piano. You can do a lot with them. A lot right and a lot wrong. A lot of it is about culture, it’s about workflow, it’s about leadership, and it’s about the intangibles. Meaningful Use is (only) about the tools.

There is always going to be variation, whether it’s a car dealership, or a plumber, or healthcare organizations that implement healthcare information technology.

Our mission is to shift the curve, and not just have a few benchmark institutions that do amazing things. We are trying to move the mass of American healthcare to a different place.”

The closing keynote speaker was Carolyn Clancy, MD, director of the Agency for Healthcare Research and Quality, U.S. Department of Health and Human services. Her style and tone was that of a calm, collected scientist when answering my question:

“The short answer is I don’t know off the top of my head what drove the variation in the Medicare Physician Group Practice Demonstration Project. A number of organizations that participated have been meeting and working together over the last year or two (to try to figure that out). They see the Innovation Center as an opportunity to further their efforts (to understand that).

One of the things they’ve learned is that at almost every level, whether it’s how they implemented their electronic records (even when they are using the same system more or less), to how they do Medication Reconciliation, to how they assess measures for PQRI, to anything else, at almost every step of the way, they have their own tweaks on it.

This would suggest to me that there isn’t a robust evidence base. We can learn a lot from this variation in terms of what seems to work.

Just as a lot of our knowledge of what works under what circumstances, our broad care improvement strategies are at a much more primitive phase of development than for our clinical interventions.

There is an incredible opportunity to learn from this variation."

Unquestionably, reasonable regulation that sets standards for cost-effective, attainable goals for both Meaningful Use and ACOs is required as we move forward. Judging by the responses from the healthcare industry to date, Stage Two MU criteria and the ACO regulation have a long way to go if they are to be considered reasonable.

Of course, we must also address the technology component in this mix. Dr. Mostashari said that variation will always exist in the implementation of HCIT. I believe this statement is true, and it supports what I wrote earlier in this post, that we must use technology wisely.

The demonstration project undertaken by CMS in 2005 produced some interesting data, but could have been more comprehensive. First, it did not feature a concurrent study of ACO configuration. As developed by Devers and Berenson in a 2009 Robert Wood Johnson funded study, there are at least four ways to structure an ACO (see their Figure 1). As we all know, PGPs and hospitals are very different in more ways that I have space here to note.

In addition to conducting a concurrent study, the participants should have been selected not only by like size, but could easily have been paired by the HCIT vendor they implemented – ten hospitals, five vendors. Why? So we could determine which of them was using the technology wisely, thereby truly learning from the variation, as implied by Dr. Mostashari’s and Clancy’s observations above. The analysis could have also made it clearer whether shared savings were in any way proportional to the number of treated or retrospectively enrolled patients. Wouldn’t CMS as well as providers want to come away with those kinds of estimates?