Radiation: The Difficulty of Monitoring the Invisible – Post 2 of 2

This post continues the list of articles on HF-related errors in radiation delivering healthcare devices.

As Technology Surges, Radiation Safeguards Lag

But the technology introduces its own risks: it has created new avenues for error in software and operation, and those mistakes can be more difficult to detect. As a result, a single error that becomes embedded in a treatment plan can be repeated in multiple radiation sessions.

A new linear accelerator had been set up incorrectly, and the hospital’s routine checks could not detect the error because they merely confirmed that the output had not changed from the first day.

In another case, an unnamed medical facility told federal officials in 2008 that Philips Healthcare made treatment planning software with an obscure, automatic default setting, causing a patient with tonsil cancer to be mistakenly irradiated 31 times in the optic nerve. “The default occurred without the knowledge of the physician or techs,” the facility said, according to F.D.A. records.

In a statement, Peter Reimer of Philips Healthcare said its software functioned as intended and that operator error caused the mistake.

Radiation Offers New Cures, and Ways to Do Harm

The Times found that while this new technology allows doctors to more accurately attack tumors and reduce certain mistakes, its complexity has created new avenues for error — through software flaws, faulty programming, poor safety procedures or inadequate staffing and training.

X-Rays and Unshielded Infants

Asked about the case, Dr. David Keys, a board member of the American College of Medical Physics, said, “It takes less than 15 seconds to collimate [cover non-scanned portions of the body – AM] a baby,” adding: “It could be that the techs at Downstate were too busy. It could be that they were just sloppy or maybe they forgot their training.”

Other problems, according to Dr. Amodio’s e-mail, included using the wrong setting on a radiological device, which caused some premature babies to be “significantly overirradiated.”

Radiation: The Difficulty of Monitoring the Invisible – Post 1 of 2

Lately, I have noticed a plethora of stories on human factors mistakes with medical equipment that delivers radiation. I have collected them here for those who are interested in this problem. At times a computer bug was at fault, but often radiation overdoses came from:

  • inadequate training (and perhaps a poor display, but those are not available for me to examine)
  • non-transferrable mental models between pieces of equipment
  • team communication issues
  • trust in automation

I provide excerpts from each article calling out the HF-related errors.

A Pinpoint Beam Strays Invisibly, Harming Instead of Healing

In Missouri, for example, 76 patients were overradiated because a medical physicist did not realize that the smaller radiation beam used in radiosurgery had to be calibrated differently than the larger beam used for more traditional radiation therapy.

Linear accelerators can be adapted to perform stereotactic radiosurgery in two ways: with small computer-controlled metal leaves that shape the beam, or with a cone attached to the machine’s opening through which radiation is delivered. That opening is made smaller or larger by moving four heavy metal “jaws” that shape the beam into a square. When a cone attachment is used, the square beam must fit entirely within the circumference of the cone. If the square is slightly larger than the cone, radiation will leak out through the four corners of the jaws and irradiate healthy tissue. In the Evanston accidents, records show, the beam was four times too large. Operators could not see this incorrect setting directly because a metal tray on which the cone is mounted hides the jaws, though the settings should have been displayed on a computer screen, according to people who have worked with this device.

That system is supposed to work this way: A treatment plan is developed on one computer, then transferred into another software system that, among other things, verifies that the treatment plan matches the doctor’s prescription. The data is then sent to a third computer that controls the linear accelerator. Several months after the Evanston accidents, Brainlab reminded customers to verify the correct jaw setting, specifically citing the possibility that treatment information could be altered as it passed “through a chain of devices.”

Evanston Hospital had earlier encountered its own communication glitches after upgrading Varian software in December 2008. As a result, medical personnel had to load patient information onto a USB flash drive and walk it from one computer to another. Then, three months ago, concerned that radiation might leak outside the cone, Varian warned customers that its software did not recognize cone attachments on the type of linear accelerator involved in the Evanston accidents. To work around that problem hospitals needed to, as one medical physicist put it, essentially trick the machine into thinking it was using a different attachment, which it did recognize. To do that, users had to enter additional data into the SRS system.

After Stroke Scans, Patients Face Serious Health Risks

While in some cases technicians did not know how to properly administer the test, interviews with hospital officials and a review of public records raise new questions about the role of manufacturers, including how well they design their software and equipment and train those who use them… While in some cases technicians did not know how to properly administer the test, interviews with hospital officials and a review of public records raise new questions about the role of manufacturers, including how well they design their software and equipment and train those who use them.

…a feature that technicians thought would lower radiation levels actually raised them. Cedars-Sinai gave a similar explanation. “There was a lot of trust in the manufacturers and trust in the technology that this type of equipment in this day and age would not allow you to get more radiation than was absolutely necessary,” said Robert Marchuck, the Glendale hospital’s vice president of ancillary services.

At Cedars-Sinai and Glendale Adventist, technicians used the automatic feature — rather than a fixed, predetermined radiation level — for their brain perfusion scans. But a surprise awaited them: when used with certain machine settings that govern image clarity, the automatic feature did not reduce the dose — it raised it. … GE says the hospitals should have known how to safely use the automatic feature… GE further faulted hospital technologists for failing to notice dosing levels on their treatment screens.

Of course, this isn’t the first time we have posted on this issue at The Human Factors Blog, and I don’t imagine it will be the last.
Previous posts:
Error Leads to Radiation Overdose
“Set Phasers on Stun” still relevant in healthcare industry


Photo credit microwavedboy on Flickr

Driven to Distraction

This editorial from MSN Autos nicely summarizes a topic we’ve covered many times:  in-car technology interfering with driving.  The central problem appears to be that in-car interfaces are designed in isolation–devoid of the context in which they will actually be used (while driving).  So the designs demand a high amount of attention and concentration.

Expert on human-automation interaction Dr. John D. Lee is quoted in the article.

But most automotive experts agree that screen and voice-control systems are here to stay. There are guidelines for good interactive system design; the Alliance of Automobile Manufacturers published a 90-page document outlining the best practices for the industry in 2006. It’s long-winded and a bit dated, but Lee of the University of Wisconsin-Madison summarizes the basic wisdom of the document in a few points:

  • Complex displays that require the driver to search for information using glances longer than two seconds should be avoided.
  • The interaction should not “time out” or force the driver to attend continually to the task. The driver should be able to interrupt the task easily and return attention to the road.
  • Visual information should be placed near the driver’s line of sight.
  • The display should be easily readable with text and icons that can be seen at a glance.

[MSN Autos; thanks Jeremy!]

Website Usability Success Story – Bethel University

The Chronicle of Higher Education has posted a great “interactive graphic” about Bethel’s re-design of their admissions page. It includes their metrics of success, an important but often difficult to quantify validation of usability.

Special problems they faced:

  • A large number of specialized programs with different application methods
  • Including financial aid information appropriately and early in the process
  • Managing multiple accounts within a single system
  • Tracking time to complete and number of students who started an application but did not complete it

Hopefully such attention to usable university sites will become more common and this cartoon will no longer be funny.


Crowdsourced Usability III: Skype

The new Skype client for Mac has received some very loud complaints about the redesign mostly centered on usability.  Skype has heard the cries:

We want you to create the chat style for an upcoming version of Skype for Mac that will be enjoyed by millions of people around the world. We’ve put a template together containing everything that you’ll need to get started and build something wonderful.

The competition will run over three stages, each lasting three weeks. Two winners will be selected at the end of each stage: a judges’ choice winner that will be selected by our panel and a people’s choice winner, as voted for by Skype users. These six winners will go forward as finalists for the grand prize.

To make it easier, Skype is providing the template files (html, css) and a simulator that you can use:

Why not flex your human factors and usability muscles and give it a try?  They are offering some great prizes to the winners too!