Towards Safer Human-Computer Interfaces
for Medical Devices and Aids

Background: Most of us have had bad experiences in hospitals, or have friends who did.


So I'm building list of discussions, warnings, suggestions, ...

  1. (added March 27, 2013) Gresham College ( U.K. ) video presentation Avoiding death by computer by "Harold Thimbleby, professor of computer science at Swansea University where he established the Future Interaction Technology Lab. His passion is designing dependable computer systems to accommodate human error."
    • Example, if a busy nurse, trying to calculate your dosage, mistakenly enters "1.2.3 * 4 = " into almost any hand-held calculator, she will not get an error warning, but (depending upon the calculator manufacturer and model) will get at least 4 different, unintended, results.
    • Human interfaces with medical devices, such as IV pumps, can be even more confusing, and fatal. Entering modes, volume rates, times, ... and alarms and alarm cancellations are often complex,
    • The log on the medical device will show what the device interpreted an entry to mean, not what the nurse entered.

  2. (added Nov 30, 2015 ) Mark Moulding < mark@markesystems.com > suggests:
    I’m surprised this page doesn’t include Therac-25 X-ray machine – pretty much the poster child for unsafe software development for a medical device.

    Errors in the operator interface software for this machine caused several patients to be given very large doses of radiation – hundreds of times the desired dosages, Three patients subsequently died as a result of this malfunction.
    https://en.wikipedia.org/wiki/Therac-25

    A happier tale about making paper manufacturing hardware and software at Measurex during the 1970s and 80s -
         but in a non-health related industry.
    Several managers (including Don Robinson and CEO Dave Bossen) would come by during the system integration and checkout on the factory floor. They would "clumsily" operate the customer interface, and seemed very interested to cause unanticipated functions, to be corrected on a high priority basis.


Please e-mail ed@ed-thelen.org with suggestions, Thank You