Previous month:
January 2013
Next month:
March 2013

February 2013

Emerging models for payment reform, Part 2

Today’s post will continue exploring emerging models for payment reform through an excellent summary paper [scroll to p.9] by M. Rosenthal in the New England Journal of Medicine (“Beyond Pay-for-Performance – Emerging Models of Provider-Payment Reform”, September 2008).

She lists the following models at various stages of development (as of her writing in 2008):

  • Incremental reform: non-payment for avoidable complication (surgery on the wrong body part or catheter-associated urinary tract infections, for instance) – currently in use.
  • Primary care payment reform
    1. Tiered case-management fees in addition to fees for service, paid per patient per month to practices demonstrating “medical homes” characteristics (a setup with more low-level interactions with patients designed for the primary care physician to take more responsibility in integrating the care of the patient) and including performance fees – pilot under development
    2. Primary care capitation with performance incentives – pilot under development.
  • Episode-based payment reform
    1. Episode-based payment model that defines global case rates for given clinical conditions, with a comprehensive score card, risk stratification and complication allowance – Prometheus, pilot under development.
    2. 90-day global fee paired with high-reliability process improvements to achieve 40 best-practice standards – in use now.
  • Shared savings
    1. Bonus for demonstrating slower growth in spending for patient care relative to peers, quality of performance affects share of savings – in use now
    2. Primary care physicians eligible to share in savings according to their performance on use of generics, emergency room visits and other criteria – in use now.

Common themes to these models include:

  • Value-based payments, driven by cost control as well as adherence to clinical guidelines and quality measures
  • The need to distinguish random variations in outcomes and patient mix from variation in practices and avoidable complications.
  • Connections between payment models and specific care delivery and organizational models (e.g., shared-savings models and large, integrated healthcare systems.)
Rosenthal also points out the importance of political factors in advancing or hindering reform, since the goals of the new payments models are to slow spending growth and move care from intensive (expensive) settings to less intensive (less expensive) ones.

Emerging models for payment reform, Part 1

By now it is well-accepted that the current fee-for-service (FFS) payment model in healthcare has severe flaws, including financial incentives for providers to administer unnecessary procedures. The question becomes: what should this system be replaced by?

This issue has been the focus of many studies and journal papers. A 2009 article in the New England Journal of Medicine provides an overview of the problem and describes global payments as a possible answer. Furthermore, in a January 2009 paper in Health Affairs (“Payment Reform Options: Episode-Based Payment is a Good Place to Start” by Mechanic and Altman), the authors examine four policy options:

  1. recalibrating FFS to make payments more aligned with community needs,
  2. instituting pay-for-performance – in fact Medicare already has implemented a policy of withholding hospital payments for eight preventable complications,
  3. creating episode-based payments, e.g. in the ProvenCare coronary artery bypass surgery program by Geisinger Health Systems (dubbed “surgery with a warranty” by the New York Times in May 2007),
  4. adopting global payments while addressing the concern – central to the backlash against capitation when that early global payment system was introduced in the 1990s – that it could encourage withholding care. A promising example is the Alternative Quality Contract in Massachusetts. In addition, “multispecialty groups such as Kaiser Permanente already operate under global payment”; however, “most US physicians are in solo practices or small groups.”

The authors use four criteria to drive their analysis: (a) potential for reducing unnecessary utilization, (b) potential for encouraging high-quality care, (c) support for provider integration, (d) operational feasibility. They point out that successful payment reform will probably involve “blended models” and require the core involvement of Medicare as a major player to drive meaningful change. Indeed, they advocate the development of an episode-based payment system for Medicare to address the 18% readmission rate of Medicare patients within 30 days of discharge and motivate reform in the private sector. This could lead to savings of up to $12 billion a year.

Mechanic and Altman ultimately argue that “episode payments are the most immediately viable approach” although global payments exhibit the highest potential for positive change (unfortunately, they are also the most difficult to implement on a large scale.) The authors further recommend that “payment reforms precede any payment reductions so that new delivery models can gain traction.”

Value-Based Insurance Design

Value-Based Insurance Design relies on “the idea that consumers' out-of-pocket medical costs should be based on the value of a service to their health rather than its price” (Washington Post, November 2010). This can be approached in two different ways: (1) reducing consumer costs for high-value services such as preventive tests and medications for chronic diseases and (2) increasing consumer costs for medical services that lack evidence of effectiveness.

The landmark paper on the topic is “Value-Based Insurance Design” by Chernew et al, published in Health Affairs in January 2007, which advocates for an approach that “acknowledges and responds to patient heterogeneity,” i.e., the fact that the benefits of many healthcare services depend on patient characteristics. It distinguishes between two implementations of VBID: (a) those that lower copayments of valuable clinical services without attempting to identify the patients who will most benefit from them, (b) those that target patients with specific clinical diagnoses.

The paper also lists the following barriers to VBID implementation (as of January 2007): concern over costs of increased use, cost of implementation, data issues, insufficient research, human resource concerns, fraud, legal barriers, privacy concerns, unintended incentives, adverse selection; however, if these issues can be addressed, VBID has the potential to significantly encourage patients to stick with high-value medical services, for instance taking drugs for chronic conditions that are now provided at lower costs. The paper further describes the case of the University of Michigan, which has tested VBID concepts for its employees with diabetes. Incidentally, the advent of VBID has motivated the creation of a research center dedicated to its study at the University of Michigan. Chernew further discusses new payment models in this YouTube video.

Managed Care Magazine, in its cover story of the October 2011 issue, makes 5 important points about VBID: (1) it shouldn’t be limited to being a “drug giveaway”, (2) “successful incentive programs… start with a careful analysis of the population you are targeting” and aren’t simply giving things for free, because people must take responsibility for their own health, (3) encouraging employers to modify their benefit structure isn’t easy, (4) value needs to be defined in some cases where precise guidelines don’t exist or for palliative treatment with serious side effects (is it delay of disability? Not being hospitalized?), (5) the advent of state-driven health exchanges doesn’t bode well for adoption of VBID since state employees aren’t familiar with implementing such concepts. 

Other Health Affairs papers investigating VBID (all published in the November 2010 issue) include:

The concept of VBID has also received attention in recent years in the mainstream press, in no small part thanks to that November 2010 issue of Health Affairs. I've already mentioned earlier the Washington Post article, which quotes many numbers gleaned from HA and summarizes the results from the real-life implementation at Pitney Bowes and in Oregon. For instance, “[Oregon healthcare plan] members are now being charged an extra $500 if they get services that the state Educators Benefit Board has determined are overused or "preference-sensitive" to patient choice, including spinal surgery, knee and shoulder arthroscopy, hip and knee replacement and upper endoscopy exams. Patients will pay an extra $100 for advanced imaging tests and sleep studies.”

In contrast with car insurance, personalized health premiums remain far in the future, but VBID offers an intriguing step toward a system that builds upon patient characteristics and encourages positive behavior. Defining and assessing value will, of course, make all the difference between an idea that profoundly impacts the healthcare field and one that fades from memory after a few years.

"The insured and the unsure"

Before tomorrow's scheduled post on health care, I thought I'd mention an article I read recently in The Economist, with the excellent title of "The insured and the unsure". It tackles the uncertainty surrounding continued health care coverage by employers in 2014 and beyond, when the main provisions of "Obamacare" come into effect.

The article points out that individuals have long had no or little alternative to receiving health insurance through their employers, but one of the achievements of Obamacare has been to make the individual health-insurance market more attractive. "The catch," the article states, "is that employers may see this as a chance to save money," although they might "pass some of the savings [generated by dropping health coverage] on to workers in the forms of higher wages."

While the new laws "will impose a fine of $2,000 per worker on any employer that does not sponsor health insurance" (with exemptions for companies employing 50 workers or less), a survey by the Kaiser Family Foundation found that "[t]he average cost of covering an employee at a large firm with a family is a hefty $15,745 a year." You do the math.

In fact, "[a] 2011 survey from McKinsey, a consultancy, found that 30% of employers would "definitely or probably" drop insurance in the years after 2014." (The full survey questions and results are available here, and examples of reactions here and here.) The Economist says of the survey that it "sparked uproar... [b]ut surveys from other consultancies were not much cheerier... [and] [e]ven employers that continue to sponsor insurance are changing their health plans," in particular increasing out-of-pocket costs for employees.

The article also touches upon "defined contribution" health plans, which echo the concept of reference pricing being put in place by some healthcare insurers for specific procedures (typically procedures for which there exist wide variations in price but not so much in quality, such as hip or knee replacement). It concludes with a warning: "there is a danger that workers will delay seeking essential treatment for fear of the bill. That could leave companies with a sicker, less productive workforce."

Flexible Partnerships in Accountable Care Organizations

To follow up on my post last week on Accountable Care Organizations and shared-savings programs, today’s post is on another Health Affairs paper, this one by Jeff Goldsmith: “Accountable Care Organizations: The Case for Flexible Partnerships Between Health Plans and Providers” (January 2011).

Goldsmith announces his views of the ACO program as early as the paper abstract: “[T]he dismal history of provider-led attempts to manage costs suggests that this program is unlikely to accomplish its objectives.” Building upon the insight that hospitals “are likely to dominate the ACO contracting process” due to the nature and objectives of the ACO process, he provides the following reasons for this opinion:

  1. “Hospitals wishing to become accountable care organizations will have to make sizable investments… to function according to that model”,
  2. “[H]ospitals are experiencing larger economic losses on the practices they are acquiring now than hospitals did in the wave of acquisitions in the 1990s”, which had already ended in “catastrophic economic losses”.

Goldsmith’s view is that these costs and losses “are likely to exceed the potential gain-sharing opportunities.” He also points out that hospital-physician collaboration has a troubled history, including mistrust, infrastructure constraints, weak incentives and income redistribution (because ACO incentives are added to the traditional fee-for-service structure) and lack of patient involvement, among others.

What he suggests in his paper is “a more flexible payment model… that would divide health care services into three categories: long-term, low-intensity primary care; unscheduled care, including unscheduled emergency services; and major clinical interventions.” The key idea is that “each category of care would be paid for differently” to reflect the different amounts of financial risk.


  • Primary care would be delivered through the concept of medical home (“an enhanced primary care model that incorporates clinical information technology; more continuous, low-intensity contact with patients… and medical management and support services provided by advanced practice nurses and nurse educators.”) Services in this model would be paid through subscriptions, i.e., physicians would be paid a risk-adjusted amount per enrolled patient per month.
  • Unscheduled care “would continue to be paid for essentially as [it is] today: fee-for-service with cost sharing for patients.”
  • Specialty care represents the most expensive component of health costs and should be paid by “a single, severity-adjusted payment when a diagnosis has been made and a clinical approach chosen.” This would require a “precise division of clinical responsibilities”, since specialty care is usually provided by teams of specialists. The author anticipates that “specialty care marts” centered on specific conditions or specific populations of patients could play a significant role in providing effective treatment.
An appealing feature of this approach, in Goldsmith’s opinion, is that it has fewer barriers to adoption than the ACO system. Most importantly, it would replace the fee-for-service system rather than overlaying it and thus has the potential to significantly transform healthcare payment models.