Visualization of multivariate time-series data in a neonatal ICU (original) (raw)

2012, IBM Journal of Research and Development

Visualization of electronic medical data in the Neonatal Intensive Care Unit (NICU) is mainly tabular or in the form of stacked univariate plots of variables over time. In the NICU, norm values differ significantly from adult values, which determine scales and alarm limits in current clinical displays. Thus, the value of information displayed in traditional interfaces is limited by standard visualizations. Providers have difficulties identifying pertinent changes in the patient's condition resulting in delayed diagnosis and harm. We developed a novel interface that allows clinicians to visualize variables critical in the detection of a patent ductus arteriosus (PDA) in a neonate. The interface was designed to allow users to quickly determine changes in variables and the direction of the change. By providing a personalized view that normalizes data points to the patient's state over the total time period reviewed, minor changes in the patient's condition are more easily detected and may allow for earlier diagnosis and treatment of a PDA. By allowing providers to experience the changes in multiple variables simultaneously, we hope to identify patterns that can be recognized by providers as changes in patient status (no PDA vs. PDA). We present the design of a multivariate time series visualization that is interactive and animated, and personalized to an individual patient, such that medical personnel can quickly and efficiently recognize significant changes in the patient's condition.

Sign up for access to the world's latest research.

checkGet notified about relevant papers

checkSave papers to use in your research

checkJoin the discussion with peers

checkTrack your impact