The Austrian Social Security Institutions are legally obligated to send data about outpatient services provided by contract partners to the Ministry of Health regularly. In order to do that, it is necessary to use secondary data containing the respective information about the procedures provided. The procedures are listed in different fee structures and contracts. Every sickness fund has its own fee structure and individual contracts with providers. Moreover, in every fee structure and contract, every procedure has a different code and name. In total, there are over 60.000 position numbers with partly same, more often similar and sometimes very different codes and names for the procedures in all the fee structures and contracts. For legal analysis purposes, the name and code for the procedures has to be the same, no matter who is the provider or payer of the service. Therefore, this magnitude of numbers has to be aggregated to a catalogue of ambulatory procedures (CAP) that is issued by the Ministry of health. This is done by a 2 step aggregation (mapping) process.
The mapping is serviced by the sickness funds and the Main Association of the Austrian Social Security Institutions. The maintenance of the mapping to the first aggregation level called “meta fee structure” is conducted by the individual sickness funds. The maintenance of the mapping from the first aggregation level to CAP is periodically conducted by the Main Association of the Austrian Social Security Institutions. In every level, changes of mappings can occur during the maintenance period.
The goal of the project is thus twofold - on the one hand enabling an easy mechanism to detect these changes. On the other hand we want to develop a useful format to display these changes to detect mapping errors (e.g. retrospective change of mapping) and to be able to tell how these changes influence the compilation of the secondary data.
For detecting changes in the assignment of services to CAP positions, we will develop integrity checks as well as queries on the database layer, to trigger notifications of changes in the data. These checks will be performed by comparing current and previous data, and will be executed before the data is imported in the current system.
Further, we will investigate the feasibility of making the changes in the data explicit with the use of temporal databases, which augments the current system with capabilities for data historization. For such an approach, we will rely on solutions developed for data subset identification and citation in dynamic databases, which have been under development in the course of the DEXHELPP project.