Visualization methods for data analysis and planning in medical applications

https://doi.org/10.1016/S1386-5056(02)00072-2Get rights and content

Abstract

Time plays an important role in medicine, both the past and the future. The medical history of a patient represents the past, which needs to be understood by the physician to make the right decisions. The past contains two different kinds of information: measured data (such as blood pressure) and incidents (such as seizures). Planning therapies, on the other hand, requires looking into the future to a certain extent. Visual representations exist for both the past and the future, and they are very useful for getting a better understanding of data or a plan. This paper surveys visualization techniques for both data analysis and planning, and compares them based on a number of criteria.

Introduction

Analyzing data and planning therapy steps are two of the central parts of practical work in medicine. Both require an overview of the information: the recorded data is the basis for further planning, and therefore, needs to be understood, but on a more abstract level than the single data points; and the plan must be understood so that it can be followed.

Both tasks can be and are supported with graphical depictions of the data—it is much easier to see a trend on a chart than when reading a row of numbers. But for complex information, more sophisticated techniques are necessary to create effective visual representations that ease understanding of the data.

Information visualization (InfoVis) [22] deals with the effective display of abstract information (in contrast to directly understandable information like magnetic resonance imaging (MRI) or computer tomography (CT) data). Other examples for typical InfoVis data are file systems, data bases (medical and other), and computer networks.

This paper surveys visualization techniques for three different applications: recording measured data (blood pressure, oxygen levels, etc.), recording incidents (seizures, pain attacks, etc.), and planning actions (therapeutic steps). Some techniques can be used for more than one application, and are, therefore, discussed in the different parts of the paper, with different requirements.

Such surveys have been done already [19], but with different requirements and a different focus.

This paper deals with the three different applications of InfoVis for time in three sections. Section 2 discusses methods for visualizing measurements, Section 3 presents techniques for displaying incidents, and Section 4 deals with methods that depict planned actions.

Section snippets

Measured data

Data that is measured either continuously (such as the parameters of intensive care patients) or when that information is needed (when the patient goes to see the doctor) later needs to be analyzed. The focus here is on the quantitative data, not so much on periods with certain symptoms—these are treated in Section 3.

Incidents and symptoms

Apart from quantitative data, incidents (like seizures, pain attacks, etc.) and periods with certain symptoms (pain, fever) are also important information for finding causes and planning therapies. Visualizing information about incidents is less common than data visualization (with its ubiquitous chart), but by no means less interesting.

Planning the future

The most complex time visualization is needed for planning. This is due to the fact that planning, by definition, involves uncertainties and needs to be robust enough to also work when unanticipated things happen.

Conclusions and future work

Visualization supports complex tasks such as the analysis of patient records and treatment planning. More sophisticated techniques provide the user with information that otherwise would not be easily accessible.

A few open problems remain. One is the problem of how to represent cyclical as well as other types of events when planning. This is made complicated by the fact that cyclical events might be specified to be performed until a certain state is reached. In such a case, the temporal extent

Acknowledgements

The Asgaard Project is supported by ‘Fonds zur Förderung der wissenschaftlichen Forschung’ (Austrian Science Fund), grant P12797-INF.

References (23)

  • R. Kosara et al.

    Metaphors of movement: a visualization and user interface for time-oriented, skeletal plans

    Artificial Intelligence in Medicine

    (2001)
  • S.M. Powsner et al.

    Graphical summary of patient status

    The Lancet

    (1994)
  • J. Alien

    Maintaining knowledge about temoral intervals

    Communications of the ACM

    (1983)
  • C.A. Brandt et al.

    Visualizing the logic of a clinical guideline: a case study in childhood immunization

    Methods of Information in Medicine

    (1997)
  • J.V. Carlis et al.

    Interactive visualization of serial periodic data

  • L. Chittaro et al.

    Visual definition of temporal clinical abstractions: a user interface based on novel metaphors

  • M.C. Chuah et al.

    Glyphs for software visualization

  • W.G. Cole et al.

    Metaphor graphics to support integrated decision making with respiratory data

    International Journal of Clinical Monitoring and Computing

    (1993)
  • C. Combi et al.

    Visualizing temporal clinical data on the www

  • C. Daassi, M. Dumas, M.-C. Fauvet, L. Nigay, P.-C. Scholl. Visual exploration of temporal object databases. Proceedings...
  • G.W. Furnas

    Generalized fisheye views

  • Cited by (40)

    View all citing articles on Scopus
    View full text