Abstract

Spatial health data is becoming increasingly important in research, especially in the field of public health. However, despite their inherent analytical content, the desired information cannot usually be extracted. As this personal data is subject to statutory data protection, access to the data and the possibilities for processing it are severely restricted. Although the publication of health data in the form of aggregates guarantees the anonymity of the data originators, the spatial information content is significantly reduced, provided it is not completely lost. The aim of geographic masks is to modify point data in such a way that as much geoinformation as possible is retained, but the correct linking of geodata with confidential target data by a data user is not possible. In this paper, various adaptations of the so-called donut masking and data aggregation are applied to a partially synthetic data set on sleep disorders of the Cologne population. The focus is on the possibilities of comparing the modified data with the original data. The aim is to be able to describe the extent to which the quality of the information is retained despite alienation. In addition to descriptive evaluations of spatial point shifts in the anonymized data, selected analysis results are to be preserved as well as possible, including the investigation of sleep disorders as a function of age, gender and nocturnal noise exposure.