Declaring that “computers are no more intelligent than a wooden pencil,” a keynote speaker at the Society for Imaging Informatics and Medicine meeting outlined how medicine can draw intelligence out of large, increasingly complex digital data sets.
Declaring that "computers are no more intelligent than a wooden pencil," a keynote speaker at the Society for Imaging Informatics and Medicine meeting outlined how medicine can draw intelligence out of large, increasingly complex digital data sets.
Systems that better manage data will allow medicine to make more intelligent use of the information it has, according to Ben Shneiderman, Ph.D., author of Leonardo's Laptop: Human Needs and the New Computing Technologies.
Usually, better management can take the form of visual displays that allow users to see at a glance large collections of data. They can then organize it to identify clusters, patterns, and outliers that give users a clearer and broader understanding of the meaning.
A key objective of these systems should be to give "beginners the performance of experts, and experts the ability to do what has never been done before," said Shneiderman, a professor of computer science at the Human-Computer Interaction Laboratory at the University of Maryland.
He demonstrated systems that present data using shapes and colors to indicate their values. The initial presentation provides an overview. From there, the users should be able to zoom and filter, then obtain details on demand.
Uses for these systems span all types of applications, from financial planning and inventory management to military uses. In the field of medicine, Shneiderman showed a "treemap" analysis of hundreds of visits to the Washington (DC) Hospital Center emergency room that captured the events using a mix of variables.
The treemap captured events and used color, size, and groupings to separate them by gender. One finding, for example, showed that women were more likely to leave the ER without treatment than men. It was also possible to screen out elements of the data to identify particular trends.
Shneiderman showed how similar concepts could be applied to an individual patient record, which could be arranged to show individual entries by time and category with links to individual items - a medical image, for example - for more detail.
Applying the concept to imaging, a CT scan might be annotated and stored. As debatases grow, it would become possible to search across the image database for similar annotations and findings, Shneiderman said.
In a work setting, this will require larger and perhaps multiple monitors, he said.