Core Health Data Elements: Report of the National Committee on Vital and Health Statistics. Compendium of Core Data Elements


To measure the current state of the use of various data sets, the Committee contracted with the Center for Health Policy Studies (CHPS) in Columbia, MD to begin identifying major data sets already in existence, especially in the private field. A range of organizations was contacted including health plans/insurers, trade or professional associations, employers, data standards organizations, and Government. Although 61 requests were made regarding data sets, almost one-third of respondents indicated that they did not have a set of health data items that they collected. Of 18 trade or professional associations contacted, only four submitted data sets. Half of the ten major employers who were asked to participate declined; only four actually sent in a data set. However, in the three remaining areas of health plans/insurers, government, and data standards organizations, the vast majority supplied data sets. A total of 31 responses were received.

Data sets received were assessed for their consistency with other data sets, particularly minimum data sets such as the UHDDS and the UACDS, the HCFA 1500 and the UB 92 data sets, and also with other current and future data sets under development by data standards organizations (ANSI). These comparisons also included consideration of the general availability, reliability, validity, and utility of data elements. A series of matrices were prepared that arrayed individual data elements in use or proposed for use by different organizations with the type of organization.

Producing the compendium was a much more involved effort than was originally envisioned, and probably is representative of problems to be overcome in the future when standardization implementation is planned. Just trying to obtain data from some large organizations was quite difficult; responses were not received in a timely fashion, and when received, the data layouts often were computerized lists rather than lists of data items with their definitions. In some instances, lists of items were received with many basic data items not included. In these cases, it is possible that the data items, such as person characteristics, are part of a more basic file kept by the organization, and the information for that file was not included.

In a number of instances, lists of data items were obtained, but without definitions. Previous experience indicates that at least some, if not many, of these data items have differing definitions. The Committee recognizes the importance of having both data items and identical definitions in order to compare and analyze data elements.

One problem that was encountered was that of requesting what the private organizations consider proprietary information. It was thought that this was one of the reasons why some organizations, especially private employers, declined to participate.

From the respondents, a total of 138 different data elements were obtained. A large number were collected by only a few of the data sets. Also, although different data sets may include the same data element, in most cases it was not possible to verify that the data collection instructions and definitions were the same.

Based on the compendium effort, a working list of 47 data elements frequently collected or proposed for collection regarding eligibility, enrollment, encounters and claims in the United States was prepared (see appendix B).