´óÏóÊÓƵ

The health care reform law (the Patient Protection and Affordable Care Act of 2010) focuses on two areas: 

1. Providing nearly all Americans with insurance.
2. Delivery system reform. 

The former has received the most media attention, but the latter is equally important. The discoveries of basic scientific and clinical research do not help patients unless they are effectively used in the delivery system.

The importance of delivery system research as a form of comparative effectiveness research (or Patient-Centered Outcomes Research, as it is often called) has recently been emphasized by the reports of the Institute of Medicine (IOM) Committee on Comparative Effectiveness Research Prioritization (Appendix A) and by the Federal Coordinating Council for Comparative Effectiveness Research.

This paper addresses two broad questions: What do we need to know about the delivery system to change it in ways that will benefit patients? Where should foundations and funding agencies like the ´óÏóÊÓƵ (AHRQ) focus their efforts? This paper suggests four key areas for delivery system research.

The paper begins by offering a definition of delivery system research and its relationship to comparative effectiveness research. It then presents a simple conceptual model of a generic delivery system organization, which it uses to provide a "long list" of potential delivery system research topics organized into broad topic areas. It suggests criteria for selecting key areas and proposes a short list of key areas for research, including some key questions in each area, providing examples of completed research, and noting areas in which research is particularly lacking. The paper concludes by locating AHRQ's recent American Recovery and Reinvestment Act (ARRA) Comparative Effectiveness Delivery System Research grants within the analytic scheme presented.

What Is Delivery System Research?

Delivery system research may be broadly defined as research that focuses on organizations which provide health care and/or research on inter-relationships among these organizations. Delivery system research may focus on the structure of these organizations; on the processes they use to provide and improve medical care; and on relationships among organizations' structures, the processes used, and the cost and quality of care provided. It may also focus on the incentives given to provider organizations by payors and on how these incentives affect organizations' structure, their care processes, and the outcomes of care generated by these structures and processes. Incentives are based on measurement of performance, so research that focuses on performance measures should also be considered to be delivery system research.

Conceptual Model

Figure 1 provides a simple model that can be used to think about an individual delivery system organization or the interface between organizations. The model is based on the familiar structure-process-outcomes relationships attributed to Donabedian. The model adds the critical factor of the external incentives faced by the organization. These incentives have a very strong influence on the structure adopted by the organization and on the processes it uses to provide and improve care. The model also adds organizational culture and leadership. Culture and leadership also strongly influence the processes used to provide care and probably influence organizational structure as well. The emergence of culture and leadership is not well understood, but both are probably influenced by the external incentives the organization faces and by its structure. Appendix B gives examples of important structures, external incentives, processes, and outcomes.

Figure 1. Generic model of an organization and its external incentives

"Long List" of Delivery System Research Areas and Topics

Appendix C presents a long list of research topics organized by the categories in the conceptual model just discussed and also suggests sample research questions for each topic.

Criteria for Selecting Priority Areas for Delivery System Research

The fundamental criterion for selecting priority areas for delivery system research should be: will this research help patients—either directly or by helping providers to provide better care? The reports of the IOM and the Federal Coordinating Council for Comparative Effectiveness Research stress that the areas studied should have a major impact, either on the population as a whole or on subgroups of patients; that research should include age groups ranging from infancy to the elderly, as well as racial/ethnic minority groups; and that research should seek to fill important gaps in knowledge. This paper suggests three additional, more specific criteria.

First, it will be important to have research that focuses on areas of delivery system reform emphasized by the health care reform law (the Patient Protection and Affordable Care Act).¹ These include new models of organization (Accountable Care Organizations, Patient-Centered Medical Homes, Healthcare Innovation Zones), new models of paying for care (e.g., bundled payments and pay for performance), and public reporting of provider performance.

Second, the paper suggests that the best way to improve the quality and to contain the cost of health care—that is, to increase its value—may be to get physicians, hospitals, and other providers into high-performing organizations and to give them incentives to continually improve care for the population of patients for whom they are responsible. Hence, research should focus on 

1. Identifying the types of organizations that are high performing.
2. Identifying the types of incentives that induce these organizations to continually improve care.
3. Identifying the types of incentives likely to lead to the creation of more high-performing organizations and to physicians and other providers becoming members of high-performing organizations.

Third, research should routinely evaluate both the intended and the unintended consequences of the structure, process, or incentive being studied. It should seek to learn the effects on racial, ethnic and socioeconomic disparities in care, as well as the effects on areas of care that are not directly addressed by the structure, process, or incentive.

Suggested Key Areas for Delivery System Research

The paper suggests four key areas for delivery system research at the present time, working from the premise that it is better to be specific and to be wrong than to be excessively general:

1. Analyses of the demographics of the delivery system—i.e., of each component of the conceptual model—and relationships among the components of the model.
2. Seeking ways to structure incentives so that they are likely to induce desirable change in the demography of delivery system organizations (toward the types of organization that research indicates provide better care) and to induce these organizations to continually try to improve the value of the care they provide.
3. Seeking ways to improve the measurement of provider performance.
4. Analyses of interprovider/interorganizational processes for improving care.

AHRQ's Recent ARRA Comparative Effectiveness Delivery System Initiative

In February 2010, ´óÏóÊÓƵused ARRA funds to issue two Requests for Applications (RFAs) to support expanded delivery system research:

• The Comparative Effectiveness Delivery System Evaluation Grants (R01) sought "rigorous comparative evaluations of alternative system designs, change strategies, and interventions that have already been implemented in healthcare and are likely to improve quality and other outcomes."
• The Comparative Effectiveness Delivery System Demonstration Grants (R18) sought "demonstrations of (1) broad strategies and/or specific interventions for improving care by redesigning care delivery or (2) strategies and interventions for improving care by redesigning payment."

Through these two RFAs, ´óÏóÊÓƵfunded six evaluation grants and four demonstration grants. From the point of view of this paper, the grants selected for funding are encouraging: four of the six evaluation grants and all four of the demonstration grants arguably fall within the list of key areas suggested in this paper.