Researchers Develop Abstraction Form to Streamline ACO Savings Analysis
Researchers from Johns Hopkins Medicine are formalizing the evaluation of how Medicare Shared Savings Program Accountable Care Organizations report savings distribution and whether the evaluation has any effect on the success of that ACO.
In a study recently published in the American Journal of Managed Care, the researchers discovered that much of the information infrastructure around MSSP ACOs was still in a nascent stage, lacking even a consistent data abstraction form for quantifying savings plan information.
Researchers were interested in using publicly available resources to evaluate what results ACOs of various sizes and organizational schemes realized in distributing shared savings among participants.
MSSP ACOs are required to maintain a website on which they report their shared savings distribution plans. Additionally, MSSP ACOs also should describe the percentage of total shared savings that will be allocated to infrastructure, primary care providers, specialists and hospitals in the ACO, said Scott Berkowitz, M.D., MBA, senior medical director for accountable care at Johns Hopkins and one of the study's authors. He co-authored the study with Matthew DeCamp, M.D., Ph.D., and lead author John Schulz, a third-year medical student.
ACOs "are expected to maintain a web site which shares some information about that ACO, and it is generally consistent among the ACOs," Berkowitz said. "It includes information such as who the entities are that participate and a host of other bits of information about the structure of the organization, the leadership, and so on. The part we were particularly focused on was whether or not they had a shared savings distribution plan listed and some of the characteristics of that plan."
However, the researchers discovered that a data abstraction form covering those elements of ACO organization was not publicly available. Schulz developed one to capture desired data in a cogent way, and the authors have made it publicly available in the appendix of the PDF version of the study, available on the AJMC website.
"There is a template…for the type of information that is supposed to be included on the web sites," Schulz said. "What's different about this is we thought it was important to home in on particular aspects of what would have been included in that template, and we also needed to do certain things so the data could be analyzed.
“For example, we had one category which was the number of participating entities in an ACO. On the ACO web site, they may have had 15 entities listed, they may have had five entities listed, they may have had one, and what we did is try to come up with a grouping of one entity, 2-5 entities, 6-10 entities, and so on,” he added. “That way, we could use that to actually do some of the analysis and see if we could learn anything about ACOs of a particular size. Without doing that, there would not necessarily have been a consistent way to do some of those comparisons."
The study focused on MSSP ACOs that began operations in 2012 and 2013. The authors found 54 percent of the ACOs reported lower expenditures in their first year, generating $383 million in savings for Medicare. Some 52 of those ACOs were eligible to receive shared savings from Medicare.
In addition, the study found ACOs planning to pay more than half their shared savings to physicians — whether primary or specialty care — were more likely to have generated savings than those that did not (39 percent to 22 percent). Further, ACOs that planned to pay more than 60 percent of shared savings to primary care providers were more than twice as likely to have generated savings as those that did not (53 percent to 24 percent). Additionally, they found that ACOs with more than 10 participating entities were more likely to have generated savings than ACOs with 10 or fewer.
However, Berkowitz was careful to caution that the association of organizational size or distribution method and results did not necessarily imply causation, especially in the early stages of ACO formation, function and reporting.
"We did believe that since it's listed as a reporting requirement, most people would be meeting that in some form," he said. "Some met it in a more detailed form than others. That's not to say one is necessarily right and one is necessarily wrong. The other piece that is really important to note about this is, there is a natural evolution within ACOs, and these types of organizations. The web site may represent a snapshot in time, but there are efforts ongoing to improve engagement and performance, and that may link to changes in the way that some of these areas may be characterized. So, some people may have been a little less detailed, knowing that the details had to be still elaborated on over time."
As the ACO concept expands beyond Medicare and into broader markets, Berkowitz said tools such as the one Schulz devised, and related analyses, will contribute a greater sense of transparency and perhaps yield some insights on which approaches work well.
"We're helping to inform of people about what we can learn about these new organizations, but we're also trying to learn about whether there are certain characteristics, trends or other areas that may be more likely to predict success," he said. "Success can be judged many ways – but are there certain factors that may be more likely to predispose to shared savings or investment in engagement within a particular process."