describes what can and cannot be done now and what the agency would like to tackle in the future—namely, the quality change issue. One of the most difficult issues that arises in health accounting and price indexing work is comorbidity; when patients require medical care for multiple conditions, the task of assigning expenditures to predefined categories accurately becomes much more complicated. This will be a key topic in these papers.
Data needs for advancing health accounting were identified at several points throughout the day. Dale Jorgenson emphasized the need to consider a wide range of possible data sources to underpin the satellite health care account. His suggestion to BEA and BLS was that they work out a way of coordinating their data infrastructures as efficiently as possible. He made the point that a solution will need to be found for combining data on providers’ prices with the information collected from claims and suggested that kind of work be put on the table for BEA. Ana Aizcorbe agreed that exploiting multiple data sources was a key task, but she noted a number of difficulties, among them the barriers to linking provider data on expenditures with data on patients. A lot could be done with claims data for the insured population if, for no other reason, because of their enormous size and coverage. BEA staff agreed that using the Medical Expenditure Panel Survey (MEPS) as the backbone of the data infrastructure, and then claims information in a supplemental role wherever gaps appear, was a reasonable strategy. Many participants noted the importance of anticipating how alternative strategies will play out once the research program is in full swing.
To monitor quality change in medical care for purposes of price measurement, accurate data on outcomes for treatments—defined in parallel with the expenditure categories—would be needed. Mark McClellan spoke about measuring treatment outcomes in this context. Among participants, there seemed to be complete agreement that quality adjustment of price indexes for the satellite health care accounts is extremely important, and also that it is very hard to do.
Gail Wilensky noted two specific types of data needs that will need to be combined: (1) individual observations in a national probability sample (e.g., MEPS), which provides good measures for conditions that occur frequently enough in a relatively small sample and for which a good reliable statistical basis can be formed, and (2) registries or other specialized data for those rarer conditions that would otherwise require very large data sets to provide that information.
In closing comments, Wilensky summarized the role of BEA’s program, noting that many of the key tasks will involve long-term research commitments. Because of this, using a mechanism such as a satellite account—in which experimental strategies can be explored while not interfering with the workings of the regular national accounts—offers a good strategy for balancing the need for additional information with the need to maintain methodologies that have a proven historical record.