CareCruiser

In this project we support medical experts in exploring patients' conditions as well as the effects of clinical actions on patients' conditions.

Problem Description: Data - Tasks - Users
Data: 

We are dealing with sequences of various patient parameters in combination with clinical activities defined in a treatment plan.

The multivariate data can be characterized as follows:

Patient parameters

Quantitative, continuously measured, abstract, time-stamped data that refer to states

Treatment plan

Qualitative (nominal), abstract, time-stamped data that refer to events

Tasks: 

The task can be described as exploring the effects of clinical actions on a patient’s condition.
Questions arising can therefore be stated as follows:

  • How did the patient’s condition change after a certain clinical action?
  • What was the effect of a certain clinical action?
  • When applying a certain clinical action – is the effect always the same or can is it possible to identify patterns?
  • Is the effect of a clinical action the same as for other patients?
  • What clinical action improved/downgraded a patient’s condition?
  • When is a patient in a critical condition?
Users: 

The intended users of the system are medical experts and physicians.

Modelling Time
Scale: 
continuous
Scope: 
point-based
Arrangement: 
linear
Viewpoint: 
ordered
Granularity & Calendars: 
single
Time Primitives: 
instant
Visualization Methods
Data Level: 

We deal with linear representation of time and abstract data.

Task Level: 

We can define the following tasks to be accomplished:

  • Visualizing individual values of parameters and activities
  • Identifying effects of clinical activities and the patient’s condition
  • Localization of drops and rises of parameter values, good/bad condition of a patient
  • Comparing effects of a clinical activity
  • Lookup activities that worsen/improved a patient’s condition

 

Presentation Level: 

We map time on a 2-dimensional, static representation. Out of this we were able to define the following visualization methods:

We use line and point plots for visualizing the patient parameters. Furthermore, different color schemas are provided to highlight interesting portions of the development of a parameter:

  • Highlighting the distance of the actual values to the specified intended value to help to localize critical values,
  • Highlighting the process of the actual values relative to the initial value to show to what extent the applied treatment plan has the intended effect, and
  • Highlighting the slope of a value to help to explore the immediate effects of applied clinical actions.
Interaction Methods

For the comparison task described above we provide the ability to change the spatial arrangement of the data (e.g., by aligning clinical treatment plans vertically).

To support the localization and lookup of specific curve events (e.g., critical values or extreme drops or rises) we provide a range slider for the color scheme to filter for such events.

Using a focus window the user is supported in investigating a specific region of the treatment. The focus window grays out the color-information outside its borders and the user can look for noticeable vertical color-patterns.

Analytical Methods

We use classification methods to determine which value is above, equal, below the intended value range.

The CareCruiser Prototype

We combined the visualization, analytical, and interaction methods into the CareCruiser prototype.

CareCruiser screenshot

Publications
Gschwandtner, Theresia, Wolfgang Aigner, Katharina Kaiser, Silvia Miksch, and Andreas Seyfang. "Design and Evaluation of an Interactive Visualization of Therapy Plans and Patient Data." In Proceedings of the 25th BCS Conference on Human Computer Interaction Conference (HCI2011), 421-428., 2011.application/pdf iconpaper.pdf
Gschwandtner, Theresia. Interactive Visualization of Effects of Medical Treatment on a Patient’s Condition In Institute of Software Technology & Interactive Systems. Vienna: Vienna University of Technology, 2012.application/pdf iconpaper