Data Management Practices for Collaborative Research

The success of research in the field of maternal–infant health, or in any scientific field, relies on the adoption of best practices for data and knowledge management. Prior work by our group and others has identified evidence-based solutions to many of the data management challenges that exist, including cost–effective practices for ensuring high-quality data entry and proper construction and maintenance of data standards and ontologies. Quality assurance practices for data entry and processing are necessary to ensure that data are not denigrated during processing, but the use of these practices has not been widely adopted in the fields of psychology and biology. Furthermore, collaborative research is becoming more common. Collaborative research often involves multiple laboratories, different scientific disciplines, numerous data sources, large data sets, and data sets from public and commercial sources. These factors present new challenges for data and knowledge management. Data security and privacy concerns are increased as data may be accessed by investigators affiliated with different institutions. Collaborative groups must address the challenges associated with federating data access between the data-collecting sites and a centralized data management site. The merging of ontologies between different data sets can become formidable, especially in fields with evolving ontologies. The increased use of automated data acquisition can yield more data, but it can also increase the risk of introducing error or systematic biases into data. In addition, the integration of data collected from different assay types often requires the development of new tools to analyze the data. All of these challenges act to increase the costs and time spent on data management for a given project, and they increase the likelihood of decreasing the quality of the data. In this paper, we review these issues and discuss theoretical and practical approaches for addressing these issues.