This month, disappointing evaluation results call for re-thinking of our approach to improving specialty care.
Nancy Keating and 12 of her colleagues from multiple medical schools and consultancies working with the CMS Center for Medicare and Medicaid Innovation (CMMI) published the results of their formal evaluation of the Oncology Care Model (OCM), a three-year large-scale pilot of an episode-based reimbursement arrangement for cancer patients undergoing chemotherapy. The results were published in an Original Investigation paper in JAMA, and also summarized in a Findings at a Glance brief disseminated by CMMI, along with other disappointing recent results from evaluations of other CMMI models.
The authors of the paper and the brief did not try to spin the results, but the appropriate formality of evaluation study publications have the effect of blunting the message, and the authors understandably shied away from too much commentary to explain how and why the OCM so clearly failed. The following is my more irreverent brief, unencumbered by such formalities.
|Fact about CMMI’s Oncology Care Model (OCM)||My Admittedly Irreverent Interpretation|
|Participation was voluntary||CMMI wanted to make it mandatory, but a lobbied Congress rejected it and CMS leaders caved, as they have repeatedly done during the decade long history of CMMI.|
|The OCM dictates that participants must do and document that they do care coordination, patient engagement, patient navigation, use “certified” EHR, provide 24/7 hotline, write care plans with goals, provide out-of-pocket cost estimation, make psychosocial plans, and write survivorship plans.||CMMI government employees want to enforce their favorite list of health care process improvements without regard for lack of evidence of effectiveness in reducing health care cost. They did so because they believed such programs are important for other reasons (as do I), and because they wanted to be in a position to defend themselves against those that might accuse them of only caring about dollars (a reasonable fear in a hyper-polarized cheap-shot environment).|
|Volunteer participants were from larger organizations||Small practices have greater control by physicians, making them more likely to say “no thanks” to the long list of CMS-dictated additional processes and paperwork. Larger practices have professional administrators who seek the “cool kid” status from participating in new programs.|
|100% of participants chose to turn down higher possible upside revenue to avoid bearing any downside risk||Professional administrators lack confidence that their clinicians will succeed in reducing cost because they know their clinicians have limited commitment to making cost-saving changes in their practices, and they know that many of their clinicians may even lack awareness of the existence of the incentives.|
|The OCM was not associated with any changes in chemotherapeutic treatment or survival, and did not reduce futile treatment (as suggested by lack of change in the use of chemo in last 14 days of life or enrollment in hospice).||Oncologists at participating institutions were never committed to making changes to the core of their practices.|
|The only change in utilization that exceeded 2.5% in either direction was a 6.7% reduction in the use of “supportive care nonchemotherapeutic drugs” such as growth factors, antiemetics and bone-modifying drugs.||My guess is that some of the participants provided clinicians with feedback metrics regarding rates of inappropriate use of such drugs, causing some clinicians to remove them from standing orders and stop routinely ordering them when not needed. That’s a good thing, and even though it is a somewhat peripheral component of cancer care, it accounted for half of the economic impact of the OCM.|
|Part D (ambulatory pharmacy) payments grew rapidly during the 3 years period, going from 23% to 30% of total episode payments. But the OCM had no impact on that growth.||The pharma beast is hungry, and CMMI has no slingshot.|
|After 3-years, the OCM achieved a net loss of $315M to Medicare because the 1% reduction in health care cost was lower than the additional 2.5% for program payments for “enhanced oncology services” and performance-based incentives.||“Oops,” or worse: “Duh.”|
In CMMI’s defense
CMMI is correct to focus on cancer care. Cancer is common. It accounts for a big chunk of health care cost. Cost is increasing rapidly. There are great variations in practice that undoubtedly lead to sub-optimal survival rates and patient experiences and wasted economic resources. It is the corner of health care where people are most fearful, most in need of care from someone they can trust to be 100% on their side, and most desperate for hope, even hope that is unsupported by science.
But, these characteristics also make cancer care a tough nut to crack, particular from a distant perch in Woodlawn, Maryland. CMMI staffers face political pressures to avoid upsetting political office-holders and their constituents and “influencers.” They want to have an impact and do good, but they also want to avoid being guilty of overreach and micromanagement. They want to achieve fundamental, long-lasting change, but they have to demonstrate success in a relatively short time period. Three years is nothing if you want to change culture, infrastructure and consequential processes. They want to use rigorous methods to evaluate the effectiveness of their “models,” but to implement programs at sufficient scale, those methods have to steer clear of costly and intrusive efforts to capture and analyze detailed clinical information necessary to really understand how the model did or did not effect actual frontline clinical decision making and care delivery processes. We have to cut CMMI some slack because they have a very tough row to hoe.
Nevertheless, clearly, we need to make fundamental changes in the approach to improving both reimbursement models and health care delivery in oncology and other specialty domains.
The bigger problem of scalable, sustainable improvement in specialty care.
CMMI is facing strategic decisions about how to deal with specialties in general. They have a longstanding emphasis on population-based reimbursement models that necessarily place primary care physicians at the helm. Those models have always had a confusing connection to specialty-oriented models addressing episodes of care and specialty care relationships. For both primary care- and specialty-oriented models, CMMI has primarily contracted with local or regional health care organizations, with some involvement by national organizations conceptualized as having diminutive roles as “conveners” and “enablers.” In recent CMMI listening sessions to obtain stakeholder input on the connection between primary- and specialty-oriented models, the commenters from organizations that were engaged as participants, conveners and enablers of primary care-oriented models (ACO, NextGen, etc.) were, not surprisingly, arguing that the responsibility for driving improvement in specialty care should be left to them. They argued that if CMMI went around them and dealt directly with specialists, they would create confusion and conflicting incentives. They argued that if specialty reimbursement (or at least the incentive component of reimbursement) flowed through them, they would be able to create a more integrated, holistic, patient-centered health care system. They were basically saying “coach, send me in” and “throw me the ball!” and sometimes it sounded more like “I alone can fix it!.”
There is a great deal of truth in those arguments, and I have certainly made some of those same arguments over the years. But, the biggest flaw in those arguments is that they don’t address the fundamental need for (1) deep engagement by frontline specialist clinicians, and (2) critical mass and scale of those specialists to be able to afford the investments required to really re-tool specialty care processes and to achieve scale economies and efficiencies in the ongoing operations that would presumably be required to enable the types of care processes that are needed. Even within large, sophisticated local or regional health care organizations, specialty departments and sections are too small and sometimes too loosely connected to the “mother ship” to be able to do real R&D work, manage real change in the most important clinical processes, and negotiate the contracts that would enable them to capture enough of the resulting economic value not only to motivate the initial change but to sustain that effort over years. CMMI 3-year demonstration projects can’t do that. NIH grants can’t do that. Drug company hand-outs can’t do that either. All these conventional sources of innovation funding in specialty areas are time-limited and come with strings attached. They may result in the publishing of a paper, but they fail to achieve enduring change.
The solution is strong specialty networks.
In my opinion, the only solution to this problem is to establish larger, stronger specialist networks that bring together like-minded specialist physicians from multiple local and regional health care organizations, and to establish value-based specialty contracts to provide a sustainable economic foundation to enable care process innovation and the development of enabling infrastructure and technology over long periods of time. It’s not a perfect model, because it could make cross-disciplinary clinical integration tougher. But, specialty care is far too consequential to allow it to remain in the hands of small specialty departments and sections that lack the scale and economic foundations to accomplish real and sustained improvement or to centralized population health departments within integrated health systems that have their hands full with primary care processes and that are distant from frontline specialists.