To put the CMS’ latest voluntary program promoting primary care in its proper perspective, one should think in biblical terms.
In the beginning, the 2006 Tax Relief and Health Care Act called on the CMS to set up “medical home” experiments at willing primary-care practices. In 2011, the newborn Center for Medicare and Medicaid Innovation begat the Federally Qualified Health Center Advanced Primary Care Practice demonstration and the Multi-Payer Advanced Primary Care Practice demonstration.
They begat the 2012 Comprehensive Primary Care initiative, which in turn begat the 2017 Comprehensive Primary Care Plus model.
And, last week, HHS Secretary Alex Azar announced that starting next January the government will launch a new Primary Care First model—the latest voluntary initiative in a long line of such programs. He claimed this one would eventually reach a quarter of Medicare beneficiaries.
What have we learned in the decade since the CMS began experimenting with turning primary-care practices into patient-centered medical homes? Have they delivered on their promise of higher salaries for primary-care physicians? Have there been fewer referrals to high-cost specialists, reduced hospitalizations and lower overall spending?
Most evaluations have concluded: No.
For instance, an RTI analysis of the use of Medicaid medical home programs in three states found they “did not have a clear pattern of impact on utilization, expenditures or quality.” An evaluation of the CMS’ most recent CPC+ model found “few, very small differences in service use and quality-of-care outcomes or total Medicare expenditures.” In fact, CPC+ practices cost the CMS 2% to 3% more than practices in traditional fee-for-service Medicare.
But those downbeat assessments are misleading. What the analysts also found was that the practices that achieved program goals have been at it for years. In other words, it takes time, as well as proper incentives, to create a successful medical home program.
The same has proven true among accountable care organizations. An analysis of physician-led ACOs in the Medicare Shared Savings Program, published last September in the New England Journal of Medicine, showed they collectively saved taxpayers just under $500 per patient even though hospital-led ACOs as a group lost money. The savings at physician-led ACOs grew larger every year they were in the program.
The organizations that do best in these programs invest heavily in care coordination. They focus on preventing hospitalizations among the 5% of patients with multiple chronic conditions who generate half of all spending.
They address unmet medical needs like untreated hypertension, pre-diabetes and behavioral health. They deploy social workers and home health aides to help patients address their nonmedical social needs like poor nutrition and inadequate housing.
Like previous pilot projects, Primary Care First has tailored its incentives to foster these approaches. It will provide practices with a single monthly payment, which “allows clinicians to focus on caring for patients rather than their revenue cycle.”
The two-sided risk model is tilted toward the upside. If practices fail to save the agency money, they can only lose up to 10% of their payments—about what the CMS calculates would be their revenue cycle costs under traditional fee-for-service Medicare. On the other hand, they can collect up to 50% of savings.
There are dangers here that bear watching. The downside only impacts the primary-care payment. The upside applies to total costs. That could lead to gaming—like stinting on necessary hospital care.
More significantly, Azar, like all previous secretaries, opted for a voluntary approach. Don’t be surprised when far fewer than projected sign up. Primary-care docs, wildly underpaid compared with specialists, tend to be risk-averse, no matter how small the downside.
It’s past time for the CMS to adopt mandatory value-based reimbursement, with appropriate ramp-up periods to accommodate the learning curve. Primary-care medical homes would have been a great place to start.
