While digital radiology can greatly improve an imaging department's capabilities, it also brings a wealth of opportunities for failure -- as measured in downtime, lost or misidentified images, poor image quality, and other problems. A scientific paper presented at this year's Symposium for Computer Applications in Radiology conference in Philadelphia showed how sound quality assurance can go a long way toward preventing disasters.
Presenting a report authored by Charles E. Willis, Ph.D. and colleagues from the Texas Children's Hospital in Houston, Douglas Tucker, Ph.D., of Agfa Medical of Ridgefield Park, NJ, said that quality control management must take into consideration the entire imaging chain.
"Automation minimizes some errors, but aggravates others," Tucker said. "Going from analog to digital, you may be able to eliminate some errors that routinely occur, but you also have to look at new errors, and new problems you've created. In quality assurance you must consider the entire imaging chain, from acquisition to display."
A misidentified image, for example, can yield to a fairly quick fix in a film-based department. With a piece of tape placed over the wrong name or date, and the correct information scribbled on it, a film can be quickly fixed and sent back through the system.
But in the digital environment, incorrect patient data can be a bigger headache. That's because in digital imaging, several copies of the picture are routinely created, archived, and transmitted electronically within the department and beyond, making it very difficult to track down and correct all of the copies. In automation, errors can become self-perpetuating.
The reverse situation is also true: some problems that are traditionally difficult to correct yield easily to digital fixes, Tucker said.
"CR is one of those areas, where a slight over- or underexposure of the cassette can be corrected, reducing the need for QA in that area."
Electronic images are imperfect records, he said. They're subject to deletion, misassociation and misinterpretation, and can be lost just as hard-copy images can. But while losing a film image may affect a single patient, losing digital media such as a disk or tape can affect many more records. Moreover, lost film jackets sometimes leave a trail by which the disappearance can be investigated -- say, by knowing how many films were in the box at the beginning of the shift, he said.
The rate of human error increases exponentially with the complexity of the system and the operator interface, Tucker said. The complexity results in part from the inability of machines, by and large, to think intuitively or do more than just follow instructions.
Moreover, there are many problems that automation can't be expected to correct, such as patient motion, poor inspiration, bad positioning, improper collimation settings, or incorrect alignment of x-ray beam and grid, to name a few. And it certainly can't compensate for choosing the wrong imaging technique -- or imaging the wrong patient.
"The point is that technology still requires you to do a good job, and do all the mental work required in the imaging process," Tucker said. Part of that mental work involves planning for normal maintenance and problems, he said.
"Quality assurance is making sure devices are properly operated, and are operating properly," Tucker said.
For example, maintaining and managing display technology has long been a problem in digital imaging. CRT monitors degrade over time, even though the newer ones are beginning to last longer. Using the wrong display look-up table can ruin an otherwise-good electronic image. But the use of test patterns, such as SMPTE, can help identify display problems, he said.
Service interruptions for scheduled maintenance should be planned for times when the equipment is not in heavy use, or optimally, not in use at all. In addition, Tucker said, keeping the right mix of spare parts and equipment on hand can eliminate downtime.
"If something goes down, you don't want it to bring your whole department down," he said. That means putting mechanisms in place to keep running until the broken item is replaced. Replacements for parts that break down routinely or have a particular life span should be documented and kept on hand, he said.
In addition, calibration should performed on a regular schedule, he said, and QA operators, who are responsible for cleaning, inspection, and documentation, must know exactly what's expected of them.
"When a vendor brings out a new version of your software and loads it onto your devices, make sure they're operating the way they did before. Make sure all of your configuration data is reloaded."
Vendors are an important part of the quality-assurance picture, and are becoming increasingly competent at defining their part of overall quality assurance, Tucker said. While vendor application training programs and documentation cannot substitute for a solid department-wide QA plan, they often serve as a good basis for developing a program that tracks both equipment and individual studies.
"Processes must be in place to verify that all images performed, and all images acquired, reach their intended destinations," Tucker said. Once errors are detected, there have to be people who are responsible for fixing them.
Radiologists play a crucial role in defining quality-assurance programs, by demanding accountability in image quality and availability, and by enthusiastically supporting the development of QA systems, Tucker said. But they can't do it alone.
"Ultimately, the institution is responsible for making sure that quality assurance is routinely performed," he said.
By Eric BarnesAuntMinnie.com staff writer
September 19, 2000
Copyright © 2000 AuntMinnie.com
