A Conversation with Dana Gelb Safran About Getting the Incentives Right: The Blue Cross Blue Shield of Massachusetts Alternative Quality Contract

The formation and success of accountable care organizations (ACOs) depend on changes in payment methods to reward improved quality and efficiency. In January 2009, Blue Cross Blue Shield of Massachusetts (BCBS-MA) announced that two provider groups (one of which is profiled in the case study) had agreed to a new five-year Alternative Quality Contract designed to achieve this goal. Dana Gelb Safran, Sc.D., senior vice president for performance measurement and improvement at the insurer, helped design the new contracting model. Quality Matters interviewed Safran to get an update on the initiative.

QM: How many provider groups are now participating in the Alternative Quality Contract (AQC)?

Safran: Currently there are 10 groups participating in the model, and some new groups are likely to come on board over the next few months. The largest in the network to date, Atrius Health, has over 400 primary care providers (PCPs) and 100,000 covered lives in the contract. The smallest, Hampden County Physicians Associates, has 100 PCPs and 12,000 covered lives in the contract. Altogether the current 10 AQC groups represent 35 percent of the 1.2 million covered lives [members] in our HMO network.

QM: Why did BCBS-MA create the Alternative Quality Contract?

Safran: In 2006, we launched a 10-year health care transformation Initiative following the precepts of the Institute of Medicine's Crossing the Quality Chasm report. The objective was, by 2016, to achieve a health care system that provides safe, timely, effective, affordable, patient-centered care for everyone in Massachusetts. There was a well-thought-out framework for the societal levers that have to be engaged to accomplish something as profound as that goal. These included things like: payment reform, quality measurement, public engagement, legislation and regulation, organizational governance, and information technology.

We were a year into that initiative when then-CEO, Cleve Killingsworth, urged the need for a new, alternative model of contracting with our provider network that could serve as an accelerant toward that 2016 vision. It was February 2007 and he said the alternative contract model needed to do two things: slow the rate of growth in spending and get more for what we’re buying in terms of quality and outcomes. So that was our charge.

QM: How does the contract work?

Safran: One of the key differences between the AQC and our traditional contracts is the payment model. Each AQC organization is paid a global budget for its patient population. That budget encompasses accountability for all care received by the provider’s population—from preventive care to end-of-life care and everything in between. In addition to the global budget, there are performance-based incentives for a broad set of quality and outcome measures for ambulatory and inpatient care. By meeting defined targets on these measures, AQC organizations can earn up to an additional 10 percent on their global budget. This serves as an important backstop against stinting on care. Finally, another key difference in the AQC contracts is that they are five-year deals, and the rate of inflation over that period is tied to general inflation. Because the consumer price index generally rises by substantially less than medical inflation, this payment structure is expected to play an important role in bending the trend on rising health care costs.

QM: How do you measure performance?

Safran: Every group participating in the AQC has the same set of performance incentives and measures. The measure set draws entirely from well-validated, clinically important, nationally accepted measures of clinical processes, clinical outcomes, and patient experiences and encompasses both ambulatory and inpatient care (see Table). Prior to launching the first contracts, we met with early adopters [of the contract] to get detailed input on the proposed measure set. This led to some minor adjustments, but these provider organizations agreed that the measures we had chosen were fair, important, and clinically appropriate.

For every measure, there are fixed performance targets—"gates"— that do not change over the five-year contract and are the same for every AQC organization. For each measure, Gate 1 represents the score that constitutes the beginning of good performance—performance that we think is worthy of additional financial recognition. Gate 5 is an empirically derived score that represents the outer limit of what can be reliably achieved on the measure. A group that achieves Gate 5 across the whole set of measures, which is a very difficult task, earns an additional 10 percent on the global budget. Those whose performance on the measure set is at Gate 1 earn an additional 2 percent. Between Gates 1 and 5, incentive payments increase in a way that motivates improvement on the early and middle part of that continuum, with less emphasis on getting from Gate 4 to Gate 5. This structure has allowed us to reward both good performance and performance improvement—an important feature of the model that helps to refute a widely held belief that one has to choose between rewarding performance or rewarding improvement.

QM: What kind of results are you seeing?

Safran: In the first year of the AQC, participating groups achieved substantial improvement on a broad set of clinical performance measures. As a group they improved the equivalent of more than one "gate" score on the full set of ambulatory clinical process measures—about twice the rate of improvement achieved by the non-AQC segment of our network.

We don't yet have final results for the first year on the clinical outcome and patient experience measures, but we will have these results by early fall. It's harder to create improvement on outcomes and patient experience, so I will not be surprised if we see more modest improvement on those measures. But I’m confident that these dimensions of care will improve over the years of the contract, because we see the kind of attention and resources the groups are devoting, not just for the sake of happy patients, but because they understand that patient engagement is going to help with all aspects of the contract, as patient adherence to treatment recommendations leads to improved clinical outcomes.

QM: What's driving this greater level of improvement?

Safran: I think a combination of factors is creating the basis for the rapid and substantial improvements we’ve seen these groups make on clinical quality indicators: 1) the measures are accepted as clinically appropriate so the AQC leaders do not experience a lot of questioning or pushback from their physicians about improving performance on these indicators, 2) real dollars are at stake for improvement, 3) performance data are being made available monthly to enable the organizations to track progress and manage their patient population, and 4) the groups have the leadership and organizational infrastructure to act on the data.

One of the exciting things about the AQC is that it’s not prescriptive about the ways that provider groups take to get to the end goals. It’s inspiring to see the wide range of approaches that the groups are taking and that each is succeeding in its own way.

QM: How does the payment model differ from typical pay-for-performance (P4P) programs?

Safran: P4P can be an effective tool for improving quality and outcomes. Of course, there's a fair amount of evidence that where P4P has been used, it hasn't been effective. I think that's largely an artifact of the way those P4P initiatives were structured. So while I think P4P can be helpful, I don't think P4P on its own can accomplish the dual goals of improving quality and outcomes while slowing the rate of growth in spending. To be a careful steward of resources is a really different task.