Reducing Design-Induced Error in Medical Devices
|
If you are the presenter of this abstract (or if you cite this abstract in a talk or on a poster), please show the QR code in your slide or poster (QR code contains this URL). |
Abstract
Background: Preventable error is a recognized problem in healthcare, which causes perhaps 15,000 deaths per year in UK hospitals. This compares badly with about 2,000 road deaths, to say nothing of longer recovery times, higher costs, and impact on carers and clinicians, the second victims.
Errors cannot be prevented unless they are noticed in the first place. Yet every health IT and medical device we have examined ignores significant classes of use errors, and therefore fails to support effective healthcare.
We are effectively at the same stage of development as the car industry was before Ralph Nader’s 1966 book, Unsafe at any speed.
Objective: To find principles and techniques for improving the safety of interactive medical devices and systems, whether embedded, mobile, apps, PC-based, or implants. To provide criteria for more informed procurement and use of such systems. Ultimately, to reduce design-induced error, and to make avoiding design-induced error a key criterion in standards, manufacturing, procurement, training, use, and patient awareness.
Methods: Reverse engineering building on formal models of interaction enables us to build faithful simulations of medical devices. Once this is done, computer simulation can very rapidly evaluate performance. For example, user interfaces have n independent feature choices, which may or may not interact with each other; we simulate all 2^n combinations of choices, and can then place commercial products along a dimension from poor to good choices of design features. As a concrete example, "five key number entry" involves moving a cursor left and right, and adjusting digits; the cursor may wraparound or not; there may be minimum numeric values or not; increasing a digit ...6789... may go to 0 or to 10, or not; and so on. We have also done conventional usability experiments. These techniques can be used to help evaluate any computer-based interactive system, not just number entry as described in this example.
Results: In some areas, such as numeric entry for drug doses, we are able to reduce unnoticed error by factors of at least 2 (numeric entry is used for many purposes and is ubiquitous in healthcare). We have a developing theory on managing unnoticed error that can be used by both designers and as a checklist during procurement, and which can be used to help improve product and app design.
Conclusions: It is possible to improve the safety of almost all healthcare devices and IT by better design. Designing systems to recognize classes of use error and to block or otherwise bring them to the user’s attention allows users to notice and hence manage errors to help avoid them leading to patient harm. In some cases (number entry being one), our research shows that it is possible to do this systematically and gain significant improvements in safety.
Errors cannot be prevented unless they are noticed in the first place. Yet every health IT and medical device we have examined ignores significant classes of use errors, and therefore fails to support effective healthcare.
We are effectively at the same stage of development as the car industry was before Ralph Nader’s 1966 book, Unsafe at any speed.
Objective: To find principles and techniques for improving the safety of interactive medical devices and systems, whether embedded, mobile, apps, PC-based, or implants. To provide criteria for more informed procurement and use of such systems. Ultimately, to reduce design-induced error, and to make avoiding design-induced error a key criterion in standards, manufacturing, procurement, training, use, and patient awareness.
Methods: Reverse engineering building on formal models of interaction enables us to build faithful simulations of medical devices. Once this is done, computer simulation can very rapidly evaluate performance. For example, user interfaces have n independent feature choices, which may or may not interact with each other; we simulate all 2^n combinations of choices, and can then place commercial products along a dimension from poor to good choices of design features. As a concrete example, "five key number entry" involves moving a cursor left and right, and adjusting digits; the cursor may wraparound or not; there may be minimum numeric values or not; increasing a digit ...6789... may go to 0 or to 10, or not; and so on. We have also done conventional usability experiments. These techniques can be used to help evaluate any computer-based interactive system, not just number entry as described in this example.
Results: In some areas, such as numeric entry for drug doses, we are able to reduce unnoticed error by factors of at least 2 (numeric entry is used for many purposes and is ubiquitous in healthcare). We have a developing theory on managing unnoticed error that can be used by both designers and as a checklist during procurement, and which can be used to help improve product and app design.
Conclusions: It is possible to improve the safety of almost all healthcare devices and IT by better design. Designing systems to recognize classes of use error and to block or otherwise bring them to the user’s attention allows users to notice and hence manage errors to help avoid them leading to patient harm. In some cases (number entry being one), our research shows that it is possible to do this systematically and gain significant improvements in safety.
Medicine 2.0® is happy to support and promote other conferences and workshops in this area. Contact us to produce, disseminate and promote your conference or workshop under this label and in this event series. In addition, we are always looking for hosts of future World Congresses. Medicine 2.0® is a registered trademark of JMIR Publications Inc., the leading academic ehealth publisher.
This work is licensed under a Creative Commons Attribution 3.0 License.