- “Niche” Computer Systems
- Meaningful Use
- “Wrong Patient”
- Cognitive Friction
- Dialog-Box Rooms
- What’s in a word?
- Cost Disease
- Model T
- Signal-to-Noise Ratio
- Anti-Data Pixels
- Fitts’s Law
- Bad Apple
On occasion, an academic paper is published, but one of the following Letters to the Editor or editorial is much more important, with a longer-lasting influence than the original article. An example is an editorial about sore throats/tonsillitis by Dr. Centor, of Centor Criteria fame. Well, now we have similar situation in the field of medical software usability.
An article in the Annals of Emergency Medicine discusses a method to help prevent wrong-patient order entry: introducing a popup with name, age, and sex, chief complaint, bed location, length of stay, and recent medication orders, but also a mandatory 2.5 second pause. If you’re interested in informatics, I don’t recommend the article, as it discusses an inelegant, klunky, duct-tape-type workaround that should never be emulated.
But Dr. Robert Wears, a tireless advocate for sanity and user-centered design for medical software, writes in his accompanying editorial, entitled “Just a Few Seconds of Your Time…” at Least 130 Million Times a Year about experts’ concerns that
…the widespread deployment of health information technology systems that are “less mature than we would like” would exceed developers’, managers’, and clinicians’ abilities to anticipate, understand, and cope with the potential consequences and so would lead to a substantial increase in the harms associated with health information technology until improvements in the quality of design, software, and implementation could catch up.
He goes on to say
In the meantime, I suspect that attempts to mitigate these risks will result in the decade of danger becoming known as the decade of the kludge.
I strongly urge you to read the Wears’ editorial (paywalled, but maybe available free through your local library). I cannot do justice to his incisive thinking and cogent writing style, but let me summarize his arguments here.
- Emergency physicians are interrupted frequently, and adding another interruption is probably bad.
- This present-and-delay approach doesn’t get at the fundamental problem. It doesn’t address other wrong-patient issues: charting, reviewing laboratory results and radiology reports, arranging follow-up, admissions, and transfers, talking with primary care or specialist colleagues, labeling specimens, and preparing discharge instructions.
- This approach is not scalable. If we introduce this popup delay for all possible wrong-patient situations, what kind of productivity impact will this have? And what about the cost of replacing monitors and computers that have been smashed by irate providers? (See Interruptions, Memory and Situational Awareness for an entertaining tale of such computer animus.)
- When the ED is busy, even small decreases in productivity may, through non-linear processes, have disproportionate bad effects on crowding and throughput.
- Kludges are bad engineering. Compared with fixing the underlying problem, in the medium and long term, kludges are more expensive, less effective, and make it harder to make the system safe.
Another thing to question about this kludgy approach. Why a 2.5-second delay? Why not 1 second, or 5 seconds? Jakob Nielsen teaches us that we have known for ~30 years that:
- 0.1 second is about the limit for having the user feel that the system is reacting instantaneously, meaning that no special feedback is necessary except to display the result.
- 1.0 second is about the limit for the user’s flow of thought to stay uninterrupted, even though the user will notice the delay. Normally, no special feedback is necessary during delays of more than 0.1 but less than 1.0 second, but the user does lose the feeling of operating directly on the data.
- 10 seconds is about the limit for keeping the user’s attention focused on the dialogue. For longer delays, users will want to perform other tasks while waiting for the computer to finish, so they should be given feedback indicating when the computer expects to be done. Feedback during the delay is especially important if the response time is likely to be highly variable, since users will then not know what to expect.
It would seem that, based on this, a delay of about 0.9 seconds would be ideal. Given the time-pressure and interruptions in the ED, I would suspect – and this is based partly on my personal experience with slow ED computer systems – that in the ED, any delay more than a second will cause providers to turn away to some other task, losing their original train of thought… and thus inviting new errors.
I am in favor of fixing problems at a fundamental level. But, I am enough of a realist (and have experienced enough different Hospital Information Systems) to know that kludges will be with us for a while. And perhaps, those faced with 130 million 2.5-second delays a year can argue for changing it to 130 million 0.9-second delays next year. And if we adopt the techniques used by marketers, such as those in the classic 1954 book How to Lie with Statistics, we can tell providers that they should be happy: we decreased this delay by 278%!