Methods to Facilitate the Capture, Use, and Reuse of Structured and Unstructured Clinical Data.

Abstract

Electronic health records (EHRs) have great potential to improve quality of care and to support clinical and translational research. While EHRs are being increasingly implemented in U.S. hospitals and clinics, their anticipated benefits have been largely unachieved or underachieved. Among many factors, tedious documentation requirements and the lack of effective information retrieval tools to access and reuse data are two key reasons accounting for this deficiency. In this dissertation, I describe my research on developing novel methods to facilitate the capture, use, and reuse of both structured and unstructured clinical data.

Specifically, I develop a framework to investigate potential issues in this research topic, with a focus on three significant challenges. The first challenge is structured data entry (SDE), which can be facilitated by four effective strategies based on my systematic review. I further propose a multi-strategy model to guide the development of future SDE applications. In the follow-up study, I focus on workflow integration and evaluate the feasibility of using EHR audit trail logs for clinical workflow analysis. The second challenge is the use of clinical narratives, which can be supported by my innovative information retrieval (IR) technique called “semantically-based query recommendation (SBQR)”. My user experiment shows that SBQR can help improve the perceived performance of a medical IR system, and may work better on search tasks with average difficulty. The third challenge involves reusing EHR data as a reference standard to benchmark the quality of other health-related information. My study assesses the readability of trial descriptions on ClinicalTrials.gov and found that trial descriptions are very hard to read, even harder than clinical notes.

My dissertation has several contributions. First, it conducts pioneer studies with innovative methods to improve the capture, use, and reuse of clinical data. Second, my dissertation provides successful examples for investigators who would like to conduct interdisciplinary research in the field of health informatics. Third, the framework of my research can be a great tool to generate future research agenda in clinical documentation and EHRs. I will continue exploring innovative and effective methods to maximize the value of EHRs.