Periodic reviews have been conducted to develop information and/or consensus on the adequacy of available information. Federal data, more than state data, have been subject to review, although there is some information on state data. Reports assessing data collected by private-sector groups are less likely to be publicly available. Most assessments preceded the current marketplace changes and addressed specific kinds of data for which agencies had defined accountability. This limits the comprehensiveness of the scope of the efforts. The most relevant assessments we found are from the periodic reports issued by the National Committee for Vital and Health Statistics (NCVHS), the typically more broad-based and topical studies funded through the Institute of Medicine (IOM), contracted studies of health expenditure accounts, and foundation-funded studies of state health data.
-
NCVHS Reports
-
Standing committees that oversee the National Center for Health Statistics’ (NCHS) data collection activities have issued periodic reports, which tend to be technical in focus. The most relevant is a recent report (US DHHS 1996) recommending a standardized set of health data elements for capturing person- and encounter-level information. The set includes 42 elements, 12 of which are person level. Twenty-six elements were viewed as being ready for immediate implementation, 10 as being substantially ready, and the rest as requiring much work. The impetus for NCVHS’s work is a concern about the administrative burden of inconsistency in data collection and reporting activities and about the confusion generated by lack of standardized health data definitions. Current and potential uses for the data cited in US DHHS (1996) include:
- Clinical uses (clinical management and continuous quality improvement)
- Payment- and cost-related uses (payment, cost containment, purchasing)
- Management (planning and budgeting, assessing staffing needs, resource allocation, system reform and change)
- Oversight (performance measurement, ensuring access, utilization review, profiling of physician practice patterns, assessing quality and outcomes)
- Technology assessment (need for technology and its effectiveness)
- Public health (surveillance, appraisal of practice)
- Analysis (risk selection and adjustment, building episodes, sample frame development, health services research and epidemiology)
- Consumer information (plan, provider and treatment choice, education)
NCVHS efforts to develop standardized person- and encounter-level data do not appear to have focused on the impact of the changing marketplace on individual data elements. For example, the payer and charge information included is relatively traditional and fee-for-service based.
-
-
Institute of Medicine
-
The IOM has convened panels to study a number of issues related to health data. With funding from NCHS, the IOM issued a report in 1992 that evaluated NCHS’s planned National Health Survey, which would integrate the four health provider surveys with the National Health Interview Survey. The IOM concluded that current systems are uncoordinated and although duplicative at times, they also suffer from important gaps. The IOM cited four areas for priority attention: (1) better insurance claims data, especially for the under-65 population, for the feefor- service and the prepaid capitated sector; (2) more information on clinical services and physiologic outcomes from medical records; (3) more information from patients (or proxies) on quality of life, health status, and satisfaction with care; and (4) better information on how much is spent (directly and indirectly, including out-of-pocket expenses) on treating particular types of patients. The IOM included the user surveys in its 1992 report and considerable external input, but the emphasis was on the needs of current users, typically from the government or research communities. In our view, the focus of the IOM recommendations is the need for person-level information to support public policy rather than the need for institutional-level analysis or the data needs of specific health interests. That is, there is more focus on the inputs and outputs of care than on the organization of resource inputs and processes of care that potentially influence outputs and outcome. In addition, the focus is more on how well and efficiently the system works for people and less on how it is structured or how entities in it perform or are influenced by change, an area of considerable interest to many of those we spoke with in the private sector. The IOM report includes specific recommendations for reformulating current surveys that are beyond the scope of this paper.
The IOM also has convened other workgroups to focus on issues important to the development of policy-relevant data. In 1995, the IOM issued a report on integrating federal health statistics on children. The summary in the report highlights relevant cross-cutting themes, including the need to link resources and child health outcomes, to assess the effects of state variations in resources and outcomes, and to coordinate efforts across agencies and the public and private sectors. Newachek and Starfield (IOM 1995) developed a paper assessing data needs using the population, the health plan and provider, and health system each as the units of analysis for monitoring health care reform. In their construct, health provider domains include health care services and effectiveness of care. Health system domains include health care resources and health care expenditures. Data issues they flagged for consideration include the comprehensiveness and timeliness of information, meeting descriptive and analytical needs, capacity to assess change, capacity to measure short- and long-term effects of change, provision of adequate geographical detail in measures, capacity to assess outcomes for vulnerable populations, creating flexibility needed to address emerging issues, and integrating efforts of different data developers. The paper concludes by urging that data collection and analysis strategies be jointly considered.
In 1996, the IOM reported on primary care delivery and needs to reorient training. Among its recommendations was a call for better information systems and quality assurance programs for primary care. The IOM also recommended the use of uniform methods and measures to monitor the performance of health care systems and primary care clinicians, careful monitoring by the government of provider supply and federal and state requirements for primary care clinicians, and a national probability sample database with episode- and population-based information. Standards for data collection were also recommended.
The IOM (1995) also has reported on workforce and educational issues for health services research. It highlighted weaknesses in the supply of researchers with “real world” experience who can support research in the following areas: organization and financing of health care (markets, risk selection, and payment rates), access to health care; practitioner, patient, and consumer behavior; quality of care; clinical evaluation and outcomes research; informatics and clinical decision making; and the health professions workforce, including better ways to forecast, plan, and manage. These areas of noted weakness are ones that relate heavily to the “supply side” of the system.
-
-
National Health Accounts
-
HCFA and others have sponsored work on issues relating the needs for information about health accounts based on expenditures. Haber and Newhouse (1991) reported on an effort to revise the national health expenditure accounts in 1988 as well as efforts proposed in 1990. They also made recommendations for future change that are still relevant today. The paper discusses emerging issues and reviews existing revisions to the accounts to better estimate out-ofpocket spending directly, disaggregate expenses, and reduce errors leading to underestimates and double-counting. It also highlights the growing prevalence of vertical integration and managed care, which creates a need for classifications of expenditures that are more responsive to type of service than type of provider, including a more meaningful definition of “professional services.” Self-insurance is flagged as a potential threat to the quality and completeness of data on private insurers. To support predicting growth in home health spending, the paper also identifies a need to better define and distinguish between the types of home health and personal health services and the use of an expanded provider list to capture information on spending.
In addition to this broad review, there are other, more focused efforts relating to health expenditure data. Genuardi, Stiller, and Trapnell (1996) consider expenditure data for the prescription drug sector, pointing out the importance of changes in retail outlets with new emerging pharmaceutical suppliers and actors; rebates and other payment changes; and other industry changes such as the growth of generic drugs and managed care. The authors compare estimates based on manufacturer sales, consumer purchasers, and retail sales, and they develop new techniques for estimating the effects of rebates. Ginsburg and Pickereign (1996) assess the policy utility and quality of data used to track health care costs. The authors focus on three kinds of data: provider data on revenues or costs, claims data from insurers, and premium data from employees. They emphasize how these compare in terms of quality and utility, what they tell you, and how they influence the conclusions one draws about costs and trends.
-
-
State Data
-
Long, Marquis, and Rogers (1995) present insight based on conversations with staff from states on priorities for health expenditure data, including why such information is used, what the concepts and components of health expenditure accounts are, and what issues and priorities exist in enhancing information. The issues include the scope of health expenditures: for example, are expenditures restricted to personal health care or does broader public health spending apply? How are interstate transfers to be assessed? Similarly, what should the categories for measuring spending be and what should the units of measurement (including cost shifting) and sources of data be? The authors believe that priority should be given to developing spending data on hospital, physician, other professional services, and prescription drugs to start, with methods that provide information on the flow of funds from payers to these functional uses. They also recommend that estimates be based on residence, and categories be based on services rather than providers. Intermediate and more long-term priorities include expansion of spending estimates to all health services, substate estimates, subpopulation estimates, capital and research spending accounts, and public health spending.
Gold, Burnbauer, and Chu (1995/1996, 1995) report on a 1994 telephone survey that asked state officials about their perception of data needs and weaknesses. In an analysis of data used to support health care reform, the authors identify major gaps in data on health expenditures (particularly for expenditures in the private sector and outside of institutions) and on health system and health plan performance. The major barriers to improving the data include funding shortages, lack of comparability across datasets, and the unwillingness of providers and insurers to submit needed data. In a related and more general analysis, the authors present information on the low levels of policymakers’ confidence in the ability to address emerging, but priority, health issues with existing data. Also discussed is the shakiness of policymakers’ confidence in specific kinds of public health data and provider data. Particular weaknesses occur in the following areas: the ability to identify alternative sources of care for clinic users, the effects of clinics on outcomes, and the ability to link data across patients, clinics, or jurisdictions. Also, inpatient data was found to be much better than data on ambulatory care, and data on structure (e.g., counts of providers) was more likely than process data to be available.
-
View full report

"infoneed.pdf" (pdf, 363.38Kb)
Note: Documents in PDF format require the Adobe Acrobat Reader®. If you experience problems with PDF documents, please download the latest version of the Reader®