Metadata Quality for Biobanks

The mission of biobanks is to provide biological material and data for medical research. Reproducible medical studies of high quality require material and data with established quality. Metadata, defined as data that provides information about other data, represents the content of biobank collections, particularly which data accompanies the stored samples and which quality the available data features. The quality of biobank metadata themselves, however, is currently neither properly defined nor investigated in depth. We list the properties of biobanks that are most important for metadata quality management and emphasize both the role of biobanks as data brokers, which are responsible not for the quality of the data itself but for the quality of its representation, and the importance of supporting the search for biobank collections when the sample data is not accessible. Based on an intensive review of metadata definitions and definitions of quality characteristics, we establish clear definitions of metadata quality attributes and their metrics in a design science approach. In particular, we discuss the quality measures accuracy, completeness, coverage, consistency, timeliness, provenance, reliability, accessibility, and conformance to expectations together with their respective metrics. These definitions are intended as a foundation for establishing metadata quality management systems for biobanks.