Urinalysis (Dip Stick) Results: eg ++Protein or +Leuks often requires follow up
COVID Test: Positive, Inconclusive, Negative
“Not Detected”: (for a specific marker or substance); “Inconclusive”: (if the test result is uncertain)
As you can see in my use of bolding above, certain results are clinically important.
Currently we do not display these results in any nuanced way - so e.g. a positive pregnancy or HIV test just looks like normal result text, and is not highlighted at all.
Does anyone have ideas on how we can support this?
Perhaps if we had a special tag at the concept-level?
I think this has slightly different answers for actually coded responses (doable in principle, but we need a new class like ConceptCoded to encode the same information as we do for ConceptNumeric.
For generic text-based responses, this would be harder. However, I think we should be careful about highlighting or not highlighting results. Trying to highlight everything that’s clinically significant is what leads down the path to “alert fatigue” even if it isn’t alerts.
Technically it is also possible that the importance changes depending on context. For example, having antibodies positive after immunization is good, but having antibodies to an organism might mean an infection… Also, it is possible that different coded answers are used for the same coded test. For example, some systems might code a test as Negative, another and Non-reactive, etc… Seems like this is a very challenging situation.
Thanks both. Re. Alert Fatigue: @ibacher when it come to Labs, I generally think we should err on the side of alerts/highlights (not popups of course), because of the typical clinician review workflow: It’s usually “visually skim, look at anything that’s bolded/highlighted, ignore everything else/treat as normal”. The only time where a clinician breaks from that skim-pattern is if there’s a lab they specifically wanted to triple-check for some specific reason (like a Troponin for Chest Pain). But in a hectic OPD or Emergency setting, this is not a guarantee. I have even seen some staff say “your labs are fine” when some text-based or tricky-range labs were actually out of normal range, just not highlighted/bolded.
And for better or for worse, this workflow is re-inforced by the pattern that most labs have trustworthy range-based alerts.
That said, if a clinician gets used to a particular lab not having trustworthy ranges (e.g. I’ve used systems that didn’t reliably flag Troponins or pregnancy results), they may specifically double-check those results - if they remember.