Dr Clif Gaus Discusses Measuring MSSP ACOs Savings

There are different ways to measure how Medicare Shared Savings Program (MSSP) accountable care organizations (ACOs) have saved money, and the way used can drastically affect the amount of savings reported, explained Clif Gaus, ScD, president and CEO of the National Association of ACOs.

Transcript

NAACOS recently released a study about how much MSSP ACOs are savings, and the findings differ greatly from what CMS has reported but are in line with research out of Harvard. What are the savings NAACOS found and what is the controversy around how to measure savings?

So, there are really are 2 ways of measuring costs and performance. The first way is the MSSP benchmark way. The benchmarks are set based on historical costs, and over a 3-year period, and then trended forward for the performance year. However, those benchmarks are capped in terms of when risk changes and goes up, the benchmark doesn’t go up. So, the patients may become sicker in the subsequent years, and cost more, but the benchmark doesn’t increase. So, ACOs are against a much harder goal or bar of performance, and therefore not always achieving as much savings as much as they truly are.

The other way of measuring is to compare precisely the ACO beneficiaries based on their risk, based on their geography, their age, and match those beneficiaries to a control group of beneficiaries who are not in the ACO, but who are living in the same geographic area, have the same illnesses, etc. And when you compare those 2 groups—the ACO beneficiaries and the like nonbeneficiaries—over time, what you see is that the ACO beneficiaries’ costs are substantially lower than the control group by almost 2-fold.

So, CMS based on benchmarks was showing gross savings of, I think it was in the range of $900 million from 2013, 2014, 2015. But when you compare them to the savings from the control group, those savings were $1.8 billion. So, the savings essentially based upon that match control group are twice what CMS was estimating.