Aligning Rewards With Large-Scale Improvement

Intensified health care reform efforts increase the urgency to achieve significant improvements in quality and substantial cost savings at the national level. Yet the pace of executing and spreading effective improvement interventions makes it unlikely these outcomes will be achieved in the foreseeable future. Although frameworks for large-scale improvement have been described by the Institute of Medicine, Commonwealth Fund, and others, there is limited evidence about how to design and implement system changes that improve population-based quality measures while also overcoming challenges inherent to large-scale change. Even though there is increasing evidence from innovation networks, improvement collaboratives, and national improvement campaigns about what works, the likelihood of achieving regional, state, or national-level improvement goals is limited without disruptive strategies that accelerate large-scale diffusion of effective interventions.^{1 - 3} Shifting the focus to rewarding sites (eg, primary care practices, hospitals) based on first achieving population-based improvement goals represents a disruptive innovation with significant potential to accelerate the spread of evidence-based interventions, thus maximizing the effects on quality and costs at a national level.

One example of a disruptive innovation linking rewards to population-based improvement across sites was designed and implemented by the independent practice association (Ohio Valley Primary Care Associates) affiliated with the physician-hospital organization (PHO) at Cincinnati Children's Hospital Medical Center and it represents the only published evidence for this approach.⁴ Specifically, an aggregate-level incentive (a reward linked to performance measures aggregated across sites) was linked to an asthma improvement initiative, affecting approximately 35% of the regional pediatric asthma population across 44 primary care practices. Overall objectives were that the aggregate-level incentive would promote shared accountability across sites for improving care across a large population, enhance leadership focus on large-scale change, and change how practices interact and learn from each other, thereby accelerating diffusion of effective interventions at a population level.

Critical to successfully implementing the aggregate-level incentive was that the board representing primary care practices committed to defining overall success as achieving improvement goals for network-level process and outcome measures. Board discussions of the aggregate-level incentive triggered an intense focus on overall design of the improvement initiative, because committing to network-level improvement required successful execution of strategies for developing, testing, and spreading interventions. An example of this effect was board approval of sharing transparent comparative practice data on process and outcome measures within 6 months of project inception. The aggregate-level incentive also promoted learning across practices that accelerated the spread of successful interventions; pushed early adopter practices to even higher performance levels to increase the likelihood of achieving aggregate-level performance thresholds; accelerated engagement of practices in the improvement initiative; and helped sustain focus on improvement relative to the all-payer population denominator within and across practices.

Based on the Cincinnati experience, aggregate-level incentives could have significant implications for achieving population-based improvement goals at the regional, state, and national levels. First, leadership groups for large-scale improvement initiatives would have to commit to defining success based on aggregate performance across sites. Second, shared accountability for improvement would be reinforced through transparency of comparative site data. Third, spread of successful improvement interventions would be accelerated because high performers would have a vested interest in promoting adoption by other sites.

Aggregate-level incentives should be tested with initiatives for which success is predicated on achieving large-scale improvement at the regional, state, or national level (eg, 5 Million Lives⁵ and Triple Aim⁶ initiatives coordinated by the Institute for Healthcare Improvement; and Aligning Forces for Quality⁷ and Improving Performance in Practice⁸ initiatives funded by the Robert Wood Johnson Foundation). A key design characteristic of the proposed framework (Figure) is that site-level rewards are contingent on first achieving designated thresholds for aggregate-level performance and meeting site-level eligibility criteria linked to measurable behaviors that contribute to achieving aggregate-level performance thresholds (eg, highly reliable and accurate data reporting, committing to public transparency of process and outcome measures, sharing successful interventions).

Although initial testing should focus on large-scale improvement initiatives involving single provider types (eg, primary care practices, hospitals), the proposed framework may have relevance to rewarding population-based improvement among providers across the continuum—especially in light of the increasing focus on accountable care systems⁹ and bundled payments. Challenges to implementing this framework include addressing resistance from those already earning rewards under existing payor-driven incentive programs, overcoming the competitive culture between sites, deciding on the balance of process and outcome measures, achieving consensus on reward thresholds, and deciding how high to set the bar on site-level eligibility criteria. Although successful across a network of primary care practices rewarded on process measures, whether this framework can be applied across other settings, as well as outcome measures, requires further study.

As Porter and Teisberg state¹⁰ : “The dysfunctional competition in health care results from misaligned incentives and a series of understandable but unfortunate strategic, organizational, and regulatory choices by each participant in the system that feed on and exacerbate each other. All the actors in the system share responsibility for the problem.”. . . “The only way to truly reform health care is to reform the nature of competition itself.” Aggregate-level incentives shift the focus to population-based performance and shared learning across sites, thus accelerating the spread of what works. Health care organizations, such as primary care practices and hospitals, can still compete on quality and cost; however, promoting shared accountability for improving population-based measures through aggregate-level incentives changes the rules. While not a panacea for what ails health care, the magnitude of potential benefit to patients and families, clinicians and health care organizations, purchasers, payers, and policy makers from accelerating large-scale, population-based improvement and reducing variation within and across regions makes aggregate-level incentives worthy of discussion and testing at the regional, state, and national levels.

While physicians, policy makers, and others can continue to debate whether rewards make a difference in improving quality, an important consideration is how to more strategically leverage current and future investments to accelerate large-scale, population-based improvement. If success is truly defined as achieving population-based improvement, why not align at least a portion of rewards with achieving this overall aim? Isn't this a more rational approach?