# west2015innovative

# Innovative information visualization of electronic health record data: A systematic review


Innovative information visualization systems for EHR data play an important role for knowledge discovery. West et al acknowledge that visualization has the potential to discover insights deeply buried in EHR data of patient visits over time, with records of changing conditions, medications, treatments, and responses related to evolving health status.

By conducting an initial literature search of 891 papers, the authors select 18 for final analysis, which include both mature systems such as LifeLines [1], LifeLines2 [2], KNAVE [3], KNAVE II [4] and VISITORS [5] and immature systems or prototypes such as DICON [6], Outflow [7] and the radial starburst by Joshi and Szolovits [8].

Techniques for visualizing EHR data are also reviewed and discussed in this paper, focusing on solving problems brought by the complexity of data and the handling of temporal data.


The authors retrieve literatures from various online sources by using several combinations of search terms. The primary objective of this paper is to discover the prevalence of visualization techniques in EHR data, therefore, literatures are screened and the results are classified based on the following categories:

  1. EHR systems that visualize real-time data, such as LifeLines [1:1] and LifeLines2 [2:1] , which incorporate horizontal axis to represent time with events on the vertical axis. Such systems often support visualizations of a single patient and a group of patients but in a separate view, and users can drill down into details for evaluation.
  2. EHR systems that visualize retrospective data, such as KNAVE [3:1], KNAVE II [4:1] and VISITORS [5:1], have a slightly different approach towards the visualizations of multiple patients records. Data is often summarized and abstracted before being visualized. They enable visual explanation of temporal data from diverse records.
  3. EHR systems that visualize static data.
    1. The radial starburst developed by Joshi and Szolovits [8:1] uses machine learning to find cluster in huge dataset but it doesn't provide interactions such as filtering, selection or brushing.
    2. DICON [6:1] also applies algorithms to original dataset for clustering, the primary visualization rendered as a treemap with glyphs.
    3. Outflow firstly visualizes patient records as a sequence of events, then it uses benchmarks (based on previous patients record or consensus reached by the healthcare industry) to make assumptions based on the patient's disease progression.


The authors identify several future challenges, mostly related to the EHR data. They believe more data being visualized makes it more difficult to spot patterns. This is partially a result of the size and complexity of EHR data. Even with scaling and zooming it's still a challenge to extract insight in a large dataset.

Larger datasets also lead to the increase of missing values, incorrect data and mixed data types. Therefore, rectification is almost impossible.

Too many variables can lead to uncertainty and even distort temporal events. A consensus has to be formed among the researchers.

The authors also believe that the healthcare provider community is reluctant to adopt innovative visualization techniques, primarily due to lack of training in using such systems.

Unsolved Problems

  1. The authors believe the learning curve has to be minimized during the system design phase.
  2. Interactive features are often less satisfactory when the amount of data user analyses increased.
  3. A consensus needs to be reached for understanding/interpretation of data among the users (my understanding of users: researchers, clinicians, patients and even policymakers)

Papers Cited


Years Spanned

1801-2014 (1994-2014 if excluding Playfair W)

Application Domain

Information Visualization of EHR.


  1. Plaisant C, Milash B, Rose A, et al. LifeLines: visualizing personal histories. SIGCHI Conference on Human Factors in Computing Systems Proceedings; 1996:221–227. ↩ī¸Ž ↩ī¸Ž

  2. Plaisant C, Mushlin R, Snyder A, et al. LifeLines: using visualization to enhance navigation and analysis of patient records. AMIA Symposium Proceedings; 1998:76–80. ↩ī¸Ž ↩ī¸Ž

  3. Shahar Y, Cheng C. Model-based visualization of temporal abstractions. ComputIntell. 2000;16:279–306. ↩ī¸Ž ↩ī¸Ž

  4. Martins SB, Shahar Y, Goren-Bar D, et al. Evaluation of an architecture for intelligent query and exploration of time-oriented clinical data. ArtifIntellMed. 2008;43:17–34. ↩ī¸Ž ↩ī¸Ž

  5. Klimov D, Shahar Y, Taieb-Maimon M. Intelligent visualization and exploration of time-oriented data of multiple patients. ArtifIntellMed. 2010;49:11–31. ↩ī¸Ž ↩ī¸Ž

  6. Gotz D, Sun J, Cao N, etal. Visual cluster analysis in support of clinical decision intelligence. AMIA Annu Symp Proc. 2011;2011:481–490. ↩ī¸Ž ↩ī¸Ž

  7. Wongsuphasawat K, Gotz D. Exploring flow, factors, and outcomes of temporal event sequences with the outflow visualization. IEEE Trans Vis Comput Graph. 2012;18: 2659–2668.34. ↩ī¸Ž

  8. Joshi R, Szolovits P. Prognostic physiology: modeling patient severity in intensive care units using radial domain folding. AMIA AnnuSymp Proc. 2012;2012:1276–1283 ↩ī¸Ž ↩ī¸Ž

🔄 Last Updated: 7/8/2020, 9:48:18 PM