Developing a self-service query interface for re-using de-identified electronic health record data

The US National Institutes of Health has developed a repository of clinical research data drawn in part from electronic health records. A new de-identified data query tool under development has been developed to support re-use of these data. We used a collection of 30 human-mediated user queries to determine whether features of the tool will be sufficient to allow users to carry out the queries themselves. The results show that the tool implemented in February 2013 will carry out a small percentage of user queries but the planned extensions will be sufficient for carrying out the majority of such queries. Future development of the tool will include extensions that correspond to the features found in human-mediated queries.