Another cost implication is the maintenance of the database being developed by the NPS. (We discuss this cost implication in more detail in section IV. Data but believe the general discussion should be repeated here in the impact analysis as well.) That database, known as the National Provider File (NPF), is currently being designed to contain the data elements shown in the table entitled, “National Provider File Data Elements” in section IV. Data, A. Data Elements, earlier in this preamble. The majority of the information is used to uniquely identify a health care provider; other information is used for administrative purposes. A few of the data elements are collected at the request of potential users that have been working with HCFA in designing the database prior to the passage of HIPAA. All of these data elements represent only a fraction of the information that would comprise a provider enrollment file. The data elements shown in the “National Provider File Data Elements” table earlier in the preamble, plus cease/effective/termination dates, switches (yes/no), indicators, and history, are being considered as those that would form the NPF. The table includes appropriate comments. The table does not display systems maintenance or similar fields, or health care provider cease/effective/termination dates.
We need to consider the benefits of retaining all of the data elements shown in the table versus lowering the cost of maintaining the database by keeping only the minimum number of data elements needed for unique provider identification. We solicit input on the composition of the minimum set of data elements needed to uniquely identify each type of health care provider. In order to consider the inclusion or exclusion of data elements, we need to assess their purpose and use.
The data elements in the table with a purpose of “I” are being proposed to identify a health care provider, either in the search process (which is electronic) or in the investigation of health care providers designated as possible matches by the search process. These data elements are critical because unique identification is the keystone of the NPS.
The data elements in the table with a purpose of “A” are not essential to the identification processes mentioned above, but they nonetheless are valuable. Certain “A” data elements can be used to contact a health care provider for clarification of information or resolution of issues encountered in the enumeration process and for sending written communications; other “A” data elements (e.g., Provider Enumerate Date, Provider Update Date, Establishing Enumerator/Agent Number) are used to organize and manage the data.
The data elements in the table with a purpose of “U” are collected at the request of potential users of the information in the system. While not used by the system’s search process to uniquely identify a health care provider, Race (with a purpose of “U”) is nevertheless valuable in the investigation of health care providers designated as possible matches as a result of that process. In addition, Race is important to the utility of the NPS as a statistical sampling frame. Race is collected “as reported”; that is, it is not validated. It is not maintained, only stored. The cost of keeping this data element is virtually nil. Other data elements (Resident/Intern Code, Provider Certification Code and Number, and Organization Type Control Code) with a purpose of “U”, while not used for enumeration of a health care provider, have been requested to be included by some members of the health care industry for reports and statistics. These data elements are optional and do not require validation; many remain constant by their nature; and the cost to store them is negligible.
The data elements that we judge will be expensive to either validate or maintain (or both) are the license information, provider practice location addresses, and membership in groups. We solicit comments on whether these data elements are necessary for the unique enumeration of health care providers and whether validation or maintenance is required for that purpose.
Licenses may be critical in determining uniqueness of a health care provider (particularly in resolving identifies involving compound surnames) and are, therefore, considered to be essential by some. License information is expensive to validate initially, but it is not expensive to maintain because it does not change frequently.
The practice location addresses can be used to aid in investigating possible provider matches, in converting existing provider numbers to NPIs, and in research involving fraud or epidemiology. Location codes, which are discussed in detail in section B. Practice Addresses and Group/Organization Options of this preamble, could be assigned by the NPS to point to and identify practice locations of individuals and groups. Some potential users felt that practice addresses changed too frequently to be maintained efficiently at the national level. The average Medicare physician has two to three addresses at which he or she practices. Group providers may have many more practice locations. We estimate that 5 percent of health care providers require updates annually and that addresses are one of the most frequently changing attributes. As a result, maintaining more than one practice address for an individual provider on a national scale could be burdensome and time consuming. Many potential users believe that practice addresses could more adequately be maintained at local, health-plan specific levels.
Some potential users felt that membership in groups was useful in identifying health care providers. Many others, however, felt that these data are highly volatile and costly to maintain. These users felt it was unlikely that membership in groups could be satisfactorily maintained at the national level.
We welcome comments on the data elements proposed for the NPF and input as to the potential usefulness and tradeoffs for these elements such as those discussed above.