Digital medical images help push worldwide data explosion

The steady growth in the number of imaging procedures per year, combined with healthcare's inexorable digital migration, means medicine can expect an avalanche of data in the future. PACS may have to be renamed PACKS.

A 2000 study titled "How Much Information?" produced by the School of Information Management and Systems at the University of California, Berkeley, predicts that the amount of new data being generated worldwide will double every year for the foreseeable future.

Specifically, the world produces between one and two exabytes (a billion gigabytes) of unique information per year, or roughly 250 MB for every man, woman, and child on Earth, the study said. Medical images are responsible for a good portion of this deluge.

Worldwide, approximately 2.16 billion radiographic procedures are performed each year. Some estimate the average chest x-ray consumes about 8 MB of storage, although other estimates run somewhat higher. Dr. H.K. Huang, a professor of radiology at the University of California, San Francisco, estimates that a typical examination generates 10 to 20 MB of data. In either case, a single x-ray is larger than the complete works of Shakespeare, which take less than 5 MB.

The demand for storage is insatiable.

In 1992, about 275 million x-rays were performed in the U.S., and this increased to rougly 320 million in 2000 - a growth rate of about 2% per year, the study said. General Electric Medical Systems estimates that 32.5 million mammograms and 4.5 million cardiac catheterization procedures are performed each year, all of which will eventually require digital storage.

To store all of the world's x-rays on computer files therefore requires a minimum of 17.2 petabytes each year. (By way of comparison, 2 petabytes could contain all the information in U.S. academic research libraries.)

Healthcare facilities must keep medical x-rays for at least seven years, and even longer in pediatric cases. If hospitals keep, say, 10 years' worth of images in their collective archives, that means the stock of medical images in any given year equals 21.6 billion images, or 172.8 petabytes of data.

"The difficulty will be in managing this information effectively," said Hal Varian, a senior author and UC dean of Information Management and Systems. "This is no easy task. Our ability to store and communicate information is quickly outpacing our ability to search, retrieve, and present it. Information management may turn out to be one of the major challenges of the new century."

The study is available online at http://www.sims.berkeley.edu/research/projects/how-much-info/ .