The Centers for Medicare and Medicaid Services (CMS) cannot find financial savings through disease management, and, as a result, curtailed their Medicare Health Support (MHS) pilots. This lack of savings by itself is not surprising, because by almost all accounts, they implemented the program poorly.
What is disturbing is that CMS's conclusion also mirrors that of the Congressional Budget Office and the RAND Corp. Yet, other organizations and consultants are finding enough value to justify contract extensions and expansions. Paradoxically, one such announcement by Independence Blue Cross was made within hours of the CMS announcement.
So how is it that well-informed people can look at the same data and come up with dramatically different conclusions and action implications?
It turns on whether the analysis is done by biostatisticians looking at utilization data in an academically rigorous way, or by actuaries and benefits consultants looking at overall financial trends in a pre-post manner. The latter group's measurements find far more savings, some of which would disappear if they were to "plausibility test" their results. Instead, benefits consultants rely wholly on pre-post comparisons to generate their findings of substantial financial opportunity or results. A brief plausibility test—actual planwide changes in utilization rates for the events being managed—often reveals major mistakes in the actuarial approach.
Changing a few key assumptions makes a huge difference. One Mercer presentation showed that just changing the outlier filter could make results vary 20-fold.
Vendors and plans seem caught in the crossfire, forced by large benefits consulting firms to promise ROIs in order to get employer contracts, and forced later by their employer customers to come up with that same number in an annual reconciliation. Yet vendors know the savings are modest and that the real impact is on customer loyalty, quality and member satisfaction. They achieve implausibly high ROIs only by taking advantage of benefits consultants' lack of experience in measurement.
MOVING THE NEEDLE
CMS and RAND notwithstanding, if a payer contracts with a qualified vendor, doesn't count asthma in the savings calculations, works together with the vendor to minimize that all-important "time to contact" (currently at about 105 days from diagnosis until first outreach), and gets a good price, the program will move the needle enough to show noticeable improvement in avoided events—as measured by changes in populationwide utilization in key event rates such as heart attacks—and hence, earn a positive ROI. Even CMS, had they avoided some head-scratching impediments to implementation (like not giving the vendors phone numbers), may well have shown significant event reduction.
By contrast, the stand-alone pre-post actuarial approach is discredited in the Disease Management Purchasing Consortium's Ultimate Guide to Outcomes Measurement as not appropriate for measuring DM. Since I was the person who, according to MHE, invented pre-post population-based financial measurement, it's not like I am attacking someone here. I was wrong, and many other people still are.
The only remaining argument in favor of the actuarial approach is that payers need to show the CFOs and their customers a number. That "number" if calculated fallaciously, means nothing. The field would be much better off showing modest returns, rather than cave into demands to show large savings numbers which don't mesh with any of the academic literature.
Al Lewis is president of Wellesley, Mass.-based Disease Management Purchasing Consortium and is an editorial advisor for MHE.