TimeLine: visualizing integrated patient records
Introduction
This article describes the data mapping and data visualization components of TimeLine, a problem list based information management and temporal visualization application.
Background
Clinical information systems have increased the volume of accumulated health data within the electronic medical record (EMR). Storing text and value based data as well as other multimedia modes of communication (images, video, and audio) have increased the necessity to present appropriate information to guide the physician in clinical tasks. To date, the complete longitudinal, patient record has not been fully realized for a variety of reasons including appropriate visualization methods capable of rendering the interplay of multiple data.
TimeLine, originally used to provide a visual record of tumor response to treatment within thoracic oncology, provides unique visualization methods for integrating patient data.
Technology Summary
The authors provide a background of related efforts of medical data visualization and describe the TimeLine architecture and data mapping techniques as well as future development prospects of the TimeLine project.
Based on “evidence that problem-oriented data visualization can enhance clinical cognitive processes”, TimeLine extends the early work of previous temporal and knowledge-based visualization projects. The display of TimeLine demonstrates the ability to integrate distributed clinical data sources and map problem-specific views.
TimeLine utilizes eXtensible Markup Language (XML) representation to facilitate the stream based data reorganization. The TimeLine architecture follows the classic Model-view-controller (MVC) paradigm to render the data mapping, visualization, and presentation.
Discussion
The authors have described a comprehensive framework offered by Timeline for accessing the EMR as a chronological, problem-oriented display that creates mechanisms for automatically mapping, re-organizing, and transforming clinical data.
The article describes in detail the process and architecture required to parse, map, reorganize, and display information. The authors highlight the issues of data mapping and classifications with respect to integrating data structures including multimedia formats (QuickTime, MPEG-4), LOINC, HL7 and DICOM as well as ICD-9 and SNOMED classifications. Additionally, TimeLine functionality allows temporal grouping or clustering that facilitates hierarchical data grouping as input into the visualization engine.
Lastly, the authors note that Clinical Context Object Working (CCOW) Group and other efforts are advocating for standards to link together HIS, RIS, PACS, and other disparate systems, only recently are groups beginning to understand how to visualize medical data by using qualitative human computer interaction (HCI) methods.
