Why the verdict comes first
Designing clinical panels that lead with the conclusion instead of making the reader assemble it
The first version of almost every clinical panel shows the data. Criteria, values, a checkmark or an x beside each one. It feels responsible. It is also the wrong thing to lead with.
A score is not an answer
Consider a panel that says “2 of 3 criteria met.” It is accurate and it is useless. The reader has to work out what the missing one was, whether it is the kind that blocks a referral or the kind that just needs a note, and what to do next. That is three inferences, made during a visit, by someone who has other things to hold in their head.
Worse, “2 of 3” reads as a near miss. It invites the reader to treat a hard exclusion and a missing phone number as the same size of problem.
Lead with the conclusion
The version that works states the conclusion first, in a sentence. “Refer to the program.” “Coverage could not be confirmed.” “This program is for adults.” Then it shows the criteria underneath as the evidence for that sentence, with the failing ones sorted to the top.
The reader gets the answer in one glance and can audit it in three more if they want to. Most of the time they do not want to, and that is fine. The evidence is there for the times it matters.
Colour is not a verdict
A related trap is letting colour carry the meaning. Red for bad, green for good, and nothing else to distinguish them. This fails for the obvious accessibility reasons, and it fails for a subtler one: colour cannot express the difference between “this patient is not eligible” and “we could not check this patient.”
Those need different shapes, different words, and different consequences. One hides the referral form. The other keeps it and asks for a note. Pairing every state with its own icon and its own sentence makes that difference legible without the reader having to learn a colour code.
Refuse to answer when you should
The state worth building that nobody asks for is the one where the panel declines to give a verdict. If the record it holds disagrees with the chart in view, the honest output is “cannot check this patient,” not a confident answer about the wrong person.
It feels like a failure state. It is actually the feature that makes the other three trustworthy.