Thursday, February 5, 2009

Community Care of North Carolina: The Disease Management Care Blog Isn't Going to Quote It Without a Big Caveat

It’s impossible it seems to read anything about the Patient Centered Medical Home (PCMH) and not run into Community Care of North Carolina (CCNC) as the ‘The PCMH Saves Money’ poster child. No power point presentation on the topic is complete without its mention, no Meeting Agenda is full if it’s not there, if you’re going to testify on the PCMH’s benefits before Congress, you should bring it up, the Commonwealth Fund is working hard to replicate it and it’s even embedded in Medical Home Wikipedia.

Does the Disease Management Care Blog mind? Not at all. Its only problem is that it doesn’t understand how the savings calculations for CCNC were done. It’s not alone either. It thinks most persons who quote from, bring up, embrace, salute, disseminate and write about the CCNC have no idea either.

But first, as the DMCB understands it, CCNC is a statewide Medicaid waiver program that operates through multiple, geographically defined, physician, hospital, health and social service organization 'networks.' Each network not only provides the usual fee-for-service care, but the primary care physicians also collect a $2.50 per member per month (PMPM) ‘case management fee,’ while the networks collect $3 PMPM for its local disease management programs (including asthma, diabetes and heart failure). A chronic care management program has been added for the Aged, Blind and Disabled Medicaid population with an additional $2.50 for the primary care providers and an additional $3 PMPM for the network. About 80% of the eligible beneficiaries are in a network.

“Mercer Human Resources Consulting Group” has performed a series of analyses that have been posted on the CCNC web site (click on 'Program Impact'). They conclude that a whopping $60 million, $124 million, $81 million and $161 million were saved in fiscal years ’03, ‘04, ‘05 and ‘06, respectively. You can read one typical report here, including how Mercer ‘excluded non-covered categories of aid,’ ‘adjusted’ data sets ‘when appropriate’ with ‘completion factors’ using an ‘historical benchmark.’

In addition to the web link above, there is this publication in the July/August ’08 Annals of Family Medicine. It’s also been referenced frequently as definitive evidence of CCNC’s economic effectiveness. Close reading of this seven page manuscript, however, reveals that it’s really a narrative description of the program with a one paragraph mention of the ‘outside assessment’ of savings by Mercer.

Contrast this with a publication from the February 2009 issue of Medical Care involving Medicaid also, but in Indiana. It describes the 'Indiana Chronic Disease Management Program' (ICDMP), which involved a regional roll-out of State sponsored nurse care management (coaching), telephonic management (from ‘non-clinical care coordinators’) and an electronic information system. The roll-out was staggered, which meant that while there were persons enrolled in ICDMP in some parts of the State, there was a window of time when patients were receiving usual care in other parts. This gave the statisticians at Indiana University a chance to perform an analysis based on a “repeated cohort design.” This enables a 'refresh' of the study population on a quarterly basis, which blunts the likelihood of ‘regression to the mean’ that is typical of a pre-post study around a single point in time. What’s more, 100% of the eligible Medicaid population was included, negating any bias from a non-representative subgroup analysis.

But did ICDMP save money? What the authors report is a big impact on claims trend, going from consistently positive (about 3% per quarter) before the regional program started and then turning flat (less than 1% per quarter) after the program was launched. Take it from the DCMB, for populations with hundreds if not thousands of dollars in PMPM claims expense, blunting an annual 12% trend adds up to millions of dollars.

So is the purpose of the DMCB to pit the North Carolina Tarheels' ‘medical home’ vs. the Indiana Hoosiers' ‘disease management program?’ Actually, they weren’t all THAT different (b-ball excluded, of course), but that’s not the point.

The point is that CCNC’s ‘proof’ of explicit multi-million dollar savings is based on a non-peer reviewed actuarial analysis posted on a web site, while ICDMP’s results are far more nuanced and spelled out in a journal using health services research-based statistics that were peer reviewed . The former is probably accurate but the DMCB has no way of assessing the veracity of an actuarial exercise. The latter is reasonably accurate and, thanks to the excruciating detail and the methods being referenced, the DMCB finds the assertion of savings far more credible.

As for the DMCB's colleagues that love to quote complicated opaque conclusions of others, it’d like to paraphrase advice from Scarecrow in the Wizard of Oz: 'If you’re not sure just how you are going to get there, any road will get you there.'

No comments: