Filling the Information Needs for Healthcare Reform

Expert Meeting Summary and Identification of Next Steps

A Meeting Summary of an AHRQ event held on March 19, 2009 with the purpose of developing a strategy to optimize the availability of information and data for enactment and implementation of health care reform.

To help meet the information needs of health care reform, and to inform our own work, the Agency for Healthcare Research and Quality (AHRQ) convened a small group of policymakers, researchers, and producers of health care data. The purpose of the meeting was to begin developing a strategy to optimize the availability of information and data for enactment and implementation of health care reform. This document provides a summary of the objectives and major conclusions from this meeting, as well as an update on early steps taken to date to put the strategy into effect.

Contents

Introduction and Context
Objectives
Key Summary Points
Potential AHRQ Roles in Filling the Gaps
Seven AHRQ Steps To Fill the Gaps

Introduction and Context

Midway through 2009, several factors are converging to create a growing national push for health care reform. Costs are increasing, health insurance coverage is declining, major disparities persist, and quality and safety issues continue. A major public policy preoccupation in the coming months, therefore, will be identifying and weighing alternative strategies to reform the health care system. This scenario creates a formidable opportunity—and challenge—for research. As health care proposals are debated and specific reforms are implemented, there will be a vital need for solid, credible, actionable information about the impact of current policies. In addition, there will be a need for reliable projections about how particular policy changes might affect health care cost, coverage, expansions, and accessibility, quality, equity, and efficiency of the delivery system. Once decisions are made, moreover, the system will need the capacity to track the impact of enacted policies, see if the effects accord with projections, and make any midcourse changes to fine-tune the policy.

To meet the needs of health care reform, research will need to be significantly more responsive to demands for findings that are timely and actionable. The health care research enterprise typically has been rich and innovative, but also fragmented and supply-driven, much like health care itself. Meeting the pressing information needs of health care reform, on the other hand, will require a clear strategy and close collaboration among the government agencies, private data organizations, and researchers producing needed data and information. It also will require strong coordination between the producers and users of that information so that researchers and analysts can to the extent possible anticipate and then meet policymakers' need for the right information at the right time and presented in a way that makes it quickly understandable and usable.

As the Department's lead health services research agency, AHRQ historically has been a major producer of research and data for policymakers; we also have a strong role as a funder and convenor of private-sector efforts. To help meet the information needs of health care reform, and to inform our own work, in March this year AHRQ convened a small group of policymakers, researchers, and producers of health care data. The purpose of the meeting was to begin developing a strategy to optimize the availability of information and data for enactment and implementation of health care reform. This document provides a summary of the objectives and major conclusions from this meeting, as well as an update on early steps taken to date to put the strategy into effect.

Return to Contents

Objectives

The AHRQ-sponsored meeting, "Healthcare Information Needs and Healthcare Reform," was held March 19, 2009, at the John M. Eisenberg Building in Rockville, Maryland. Attendees included representatives from Federal agencies, health research and policy institutions, employer groups, other such organizations, and researchers. The goals of the meeting were to:

  • Identify the major information needs for national, State, and other public health policy leaders, for current policymaking and subsequent fine-tuning.
  • Understand the areas in which existing simulation models and research can provide answers and those where it cannot.
  • Discover areas in which additional capacity is needed.
  • Identify short- and longer term strategies for exchanging information and filling gaps and ways that AHRQ can be helpful to this effort.

Return to Contents

Key Summary Points

What Policymakers Need

While the specific information needs obviously depend on the nature of the policy being considered, policymakers indicated several major cross-cutting requirements that affect all questions. They need information on the likely concrete effects of unique, specific policy proposals. They need the information quickly and customized in a way that addresses the specific characteristics of the proposals and that take into account the specific levers available to policymakers. Assumptions behind the estimates and the methodological approach must be clear and transparent. In essence, they will need to know not just the policy impact but also the political impact of any given proposal: who benefits, who loses, what it costs, who pays, what the effects are on consumers/patients, and what the trade-offs are among alternative proposals. The impact on the public sector and consumer and not just the "system" is critical. In short, they need stand-ready capacity to provide answers with all of these attributes.

Many of the policymakers present were also veterans of past health care reform efforts. A critical point made by these veterans was that the need for solid data is even more critical this time, and the data needs are much broader. While in the past most of the focus was on improving health care coverage, or the "demand" side of health care reform, this time information on the "supply" side (quality, safety, efficiency, effectiveness, and equity of care) is at least as important. Current capacity to model effects on the "demand" side is well-established (as a result of a ramp-up during the last reform effort), but unfortunately there is limited if any capacity to model the effects on the "supply" side. Capacity to model the effects of policy changes on the quality, safety, and cost of hospitals, physicians, and other providers is needed, along with capacity to model (and later track) the impact on whole episodes of care.

Policymakers also recognized a need for better, faster, and smoother communication with producers of the information. They need a ready way to find out what information is knowable, who has it, and how to get it quickly and in a form they can use.

Existing Capacities

Several entities have independent micro-simulation models, most of which are based on AHRQ's Medical Expenditure Panel Survey (MEPS) data. Micro-simulation capacity can examine the effects of Federal policy financing proposals on utilization, uninsured populations, employer-based insurance coverage, and Federal budgets related to changes in tax policies. As noted later, however, there are enhancements that could be made within each of these models, as well as better coordination across models.

Major Gaps

A major purpose of the meeting was to identify major gaps. Policymakers and researchers identified gaps in three categories: the data itself, the evidence base on what factors affect different data points, and the availability of projections and models based on these data.

Data

Informing and tracking health care reform will require solid data on all aspects of the health care system, from insurance coverage to care delivery to outcomes. The consensus of the group was that the adequacy of the data varies considerably from one question to another. MEPS provides information on utilization, insurance coverage, and expenditures by patients and employers. It also provides much of the data for multiple micro-simulation capacities both inside and outside AHRQ. The group had several recommendations for improving these databases and simulations on coverage and expenditures. In particular, attendees suggested, the clinical information and longitudinal design could be improved, and MEPS lacks sufficient sample size for enhanced drill-down capacity at more refined geographic and other levels (e.g., rural or near old).

The group also noted a lack of adequate provider-side data. National Center for Health Statistics provider surveys, while useful for certain purposes, lack the sample size needed for drill-down. The Healthcare Cost and Utilization Project (HCUP) is a rich source of data on inpatient care and, when used with the Quality Indicators, can provide information on quality, safety, cost, and potentially preventable hospital admissions, with excellent drill-down capacity. The database also has extensive data on emergency department care and ambulatory surgery. But physician practice data are sparse, and data that cut across episodes of care are at a very nascent stage. In addition, little modeling has been done using provider-based data of this sort. The group identified a need to accelerate production of provider-side data and to accelerate modeling using these data.

Another gap identified was the need to improve information on vulnerable populations and on disparities, particularly through the improvement of Medicaid data, including more cleaning of the data, faster turnaround on the data, and more linkage of Medicaid with other data. This is important because there is a strong need to ensure that any health reforms enacted improve disparities rather than worsening them. In addition, Medicaid directors have noted a need for better Medicaid data for improving management of their programs. Attendees also suggested greater use of data from the Federal Employee Health Benefit Plan (FEHBP).

While there is recognition that an overall data effort to expand these data sets would be extremely worthwhile and would provide a rich source of information, the long-timeframes to plan, secure funding, collect, and analyze data are impractical for the immediate informational needs of health care reform in 2009. The group felt that in the short term much could be gained by making current data more available, enhancing and linking existing data resources, and in some cases identifying strategies to enhance the timeliness of a subset of high-priority data. Other short-term practical strategies could include statistical matching techniques, imputations, and linkages. But the group also agreed that this is also the moment to jump-start more ambitious and longer term data efforts to bring in new data because the visibility of current gaps creates a political momentum for action. Moreover, it was felt that health reform will not be a one-time fix; therefore, the needs for new data will continue in the future.

A final challenge noted was the need for data to track the impact of existing or new policies. Policymakers need a quick way to assess the impact of any changes in legislation and payment so that they can fine-tune these policies, and managers need this information so that they can make improvements to cost and efficiency as well as quality. One requirement in this regard will be good cost accounting systems that do not exist now. An investment in cost accounting could be a huge impetus to better tracking and management of cost and efficiency.

Knowledge

While data are essential to inform health reform, they are not sufficient. To be able to project the impact of particular policies, we need a solid base of evidence on how providers react to particular kinds of financial and nonfinancial incentives, how patients/consumers react to incentives, and what types of competition or health care markets are most conducive to desired provider behavior. AHRQ and others have funded some research on these economic, financial, social, cultural, and legal-regulatory forces in the past, but the science is still lacking. Thus, it is often unclear how providers will react to particular policy proposals, whether they are incentive-based reimbursement schemes or regulatory approaches. Research and information on delivery, recent policy relevant to organizational change, and underlying market forces are needed. Another requirement is availability of solid qualitative research so that we can identify not only the intended and unintended impact of a new policy, but also when, where, how, and why it is having this impact.

Modeling

Participants agreed that multiple independent micro-simulation models exist that apply the same host data set (e.g., MEPS) to address the same policy initiative. This richness of models presents some enormous opportunities. On the other hand, it was noted that these models are quite sensitive to the assumptions that underlie them, so different simulations often give quite different answers to the same question, even though most rely on the same underlying data. One suggestion was to have greater transparency about the underlying assumptions. In addition, syntheses of current research findings that provide empirical support for model assumptions would be beneficial to modelers.

The largest and probably most important gap in modeling, according to the group, is the lack of modeling on the provider side, projecting how different changes in policy, regulation, payment, and other factors would affect health care quality, safety, cost, and effectiveness. As the group noted, this time the health reform debate is very much about provider cost and efficiency, along with quality and safety, so we are going to need models not just for insurance costs and coverage. In the short term, filling this gap will require applying modeling and simulation techniques refined elsewhere to existing data and evidence; in the longer term, as noted above, it requires substantially enhancing both the data and the evidence base in these areas.

Return to Contents

Potential AHRQ Roles in Filling the Gaps

The group identified several potential roles for AHRQ in filling these gaps. These included:

  • Summarizing the meeting and post it on the Web as a way of informing a broader audience and extending the dialogue.
  • Convening groups of users and researchers to further flesh out individual pieces, such as collaboration among current data modeling efforts using MEPS and developing a modeling capacity on the provider side.
  • Making improvements to AHRQ data to make them more helpful for policymakers.
  • Identifying new data sources that could be useful.
  • Identifying ways to facilitate use of existing information and links to enrich the data.
  • Finding quick ways to track the impact of policies.

Return to Contents

Seven AHRQ Steps To Fill the Gaps

In reviewing notes from the meeting, AHRQ has identified the following next steps. In the meantime, we continue to welcome ideas on all these steps, as well as others that are not included here:

  1. Enhance transparency on AHRQ data used for reform:
    • Post this meeting summary on the AHRQ Web site.
    • Develop special section of the AHRQ Web site to highlight data and information that would be useful to address health care reform issues. This plan is under development.
  2. Provide leadership in convening public-private partnerships:
    • Plan followup meetings and potentially provide infrastructure to facilitate better communication between entities that need information about the effects of policy proposals and entities that produce this information.
  3. Accelerate creation of evidence base on behavioral and provider impact of policy decisions:
    • Solicit grants in these areas.
    • Synthesize existing knowledge.
  4. Enhance MEPS for modeling efforts:
    • Develop linkage of establishment-level data to supplement household data.
    • Develop linkage of medical provider-level data to supplement household survey.
    • Provide summary of set of best practices for modeling efforts.
    • Explore ways to release currently restricted data to the public in support of the modeling efforts, while still adhering to confidentiality provisions. This may include release of State identifiers or more discrete indicators that distinguish areas as urban or rural. Explore data-sharing arrangements to improve data access.
  5. Improve/expand provider-level data:
    • Identify potential new data sources.
    • Facilitate all-claims cross-site data.
    • Identify ways to increase data timeliness.
    • Plan better integration of electronic health records and claims data.
    • Enhance data on race and ethnicity.
  6. Begin to build simulation capacity for provider behavior:
    • Support expert meeting with particular focus on provider side.
    • Explore intramural project to illustrate simulation capacity for hospitals.
    • Enhance HCUP and other provider data to make them more useful for modeling.
  7. Pursue longer term goals:
    • Work with others in the Department of Health and Human Services (HHS) on a strategic approach to data investments that are most responsive to short, intermediate, and long-term HHS and national needs, and explore the feasibility of public/private collaboration.
    • Integrate advances in adoption of health information technology and electronic/personal health records as supplements/replacements to current data collection efforts.
    • Develop and implement a data strategy to support comparative effectiveness research in concert with these efforts.
Current as of July 2009
Internet Citation: Filling the Information Needs for Healthcare Reform: Expert Meeting Summary and Identification of Next Steps. July 2009. Agency for Healthcare Research and Quality, Rockville, MD. http://www.ahrq.gov/research/data/hininfosum/index.html