How a Workaround Quietly Becomes Your Policy

A charge nurse shows a new hire how to get a particular order through. There is a step in the middle that is not in any tip sheet — you have to back out of the screen and re-enter from the other menu, otherwise it will not save. The new hire writes nothing down and does not need to. They will do it that way for the next nine years.

Ask anyone on that unit where that step came from and you will get a shrug. Ask when it started and nobody knows. Ask whether it is approved and you will get a puzzled look, because the question does not parse. It is just how you do it.

This is the mechanism by which a workaround becomes your policy, and it does not require anyone to make a decision. That is what makes it hard to interrupt.

The mechanism, step by step

It starts as a genuine fix for a genuine problem. Something in the build was misconfigured, or an interface was flaky, or a vendor bug ate submissions under a specific condition. Somebody clever worked out a way around it. At that moment the workaround is correct, well-understood, and provisional — the person using it knows exactly why it is necessary and would be happy to stop.

Then the underlying problem gets fixed. Or it does not, and nobody checks. Either way, the workaround does not get retired, because nothing connects the fix to the people who built habits around the breakage.

Then it gets taught. This is the critical transition, and it happens the first time someone who does not know why explains it to someone else. The reason drops out of the transmission — not through carelessness, but because the reason takes thirty seconds to explain and is not needed to perform the task. What passes to the next person is the step, stripped of its justification.

Then it gets defended. A few years in, someone proposes doing it the documented way. The unit pushes back, and they push back with authority, because as far as anyone present knows this is how it has always been done and there must be a reason. There was a reason. It expired.

At no point did anyone decide. Every individual in the chain behaved sensibly. And now the organization has a practice that differs from its policy, that nobody can trace, and that will be discovered by an auditor, a surveyor, or an event.

Why oral transmission is the load-bearing part

I want to be precise about what is doing the damage, because the workaround itself is often harmless and occasionally better than the documented path.

The damage comes from three missing properties, and they are all properties of writing rather than of correctness.

There is no version. A verbal instruction cannot be superseded, because there is no earlier copy to replace. When the official process changes, the verbal one continues in parallel, and nothing in the environment signals a conflict.

There is no date. Nobody can tell whether this practice predates or postdates the upgrade that supposedly fixed the underlying issue. Without a date you cannot even ask the question that would resolve it.

There is no owner. If a documented procedure is wrong, there is somebody to tell. If the practice lives in the collective memory of a unit, there is no one to correct, and a correction issued to everyone reaches no one in particular.

A searchable, dated, owned record of the approved path is the only thing I have seen reliably interrupt this. Not because people read policies — they do not — but because when the question does come up, there is something to check against, and checking takes less effort than arguing from memory.

The uncomfortable part for policy offices

Here is the objection I would raise against my own argument.

Sometimes the workaround is right and the policy is wrong. The documented process was written by a committee that did not have to perform it, and the unit found something faster that produces the same outcome. Treating every deviation as an error means grinding down the people who are compensating for a badly designed process, and it means the organization never learns what is actually wrong with the design.

That is real, and any program that responds to practice variation with enforcement alone will lose the trust it needs to see the variation in the first place. The moment staff learn that surfacing a workaround produces a corrective action, workarounds go further underground and you have made the problem invisible rather than smaller.

So the response has to be diagnostic before it is corrective. When you find one, the first question is not who authorized this. It is what problem was this solving, and does that problem still exist. Three answers are possible: the problem is gone and the practice should stop, the problem persists and should be fixed at the source, or the workaround is better than the policy and the policy should change. Only the first is a compliance matter, and it is the least common of the three.

Where they become visible

The reason these persist is not that they are hidden. It is that nothing in a hospital is instrumented to notice them.

Incident reporting catches them only when they fail. Audits catch them at a point in time, on the units audited, if the auditor happens to observe the step. Rounding catches them if someone volunteers. None of these is a detector; they are all accidents.

What does work, and what I have found genuinely surprising in its usefulness, is watching what people search for. A retrieval layer sees the questions, and the questions carry the variation. When one unit repeatedly asks a question in terms that presuppose a step the documented process does not contain, that is a signal, and it arrives without anyone being observed or reported. Solution Compass keeps the questions as well as the answers for exactly this reason — the search log is a practice-variation instrument that nobody has to staff.

It is also, worth noting, a much gentler instrument than the alternatives. Nobody was watched. Nobody was written up. A pattern appeared in aggregate, and somebody senior got to ask a curious question instead of an accusatory one.

Go and ask

Pick one procedure you believe is performed consistently across your units. Pick a boring one — a form, a routine order, an escalation. Then have somebody walk two units and ask a staff member on each to show them how it is done.

Do not compare either one against the policy yet. Compare them against each other.

If they differ, you have found one, and there is no version of the story in which it got there on purpose.