When Clinic Workarounds Become Infrastructure

One MOA built it because referral follow-up had become too easy to lose track of. Some specialist offices confirmed receipt. Some did not. Some needed a second fax. One office reliably answered the phone only before 10:30 a.m., and another seemed to process referrals faster if the patient’s imaging report was attached twice. None of this belonged neatly in the EMR, so she made a tracker.

It worked.

The spreadsheet had columns for referral date, specialist, fax confirmation, patient contact, and next follow-up. Later, colour codes were added. Yellow meant “watch.” Orange meant “call again.” Red meant “bring to physician.” Eventually, only one person fully understood the difference between orange and red when the specialist office had replied verbally but nothing had arrived in writing.

Still, the work moved. Referrals were chased. Patients were updated. Physicians stopped asking about certain items because they trusted that the list was being watched.

That is how a workaround becomes infrastructure.

Not dramatically. Not through a meeting. Not because anyone decides, “This spreadsheet is now a critical part of how our clinic manages referrals.” It happens quietly, through usefulness. The workaround solves a problem well enough that people stop seeing it as a workaround. It becomes part of the clinic’s operating system, even though it was never designed, documented, shared, or protected like one.

A workaround is not risky because it exists. It becomes risky when the clinic starts depending on it as infrastructure while it remains informal, person-dependent, and invisible to the wider team.

At that point, the question is no longer whether the workaround is clever.

The question is what happens once the clinic depends on it.

In a small clinic with stable staff, informal systems can appear stronger than they are. Everyone knows who checks the spreadsheet. Everyone knows which paper folder matters. Everyone knows that unread results do not always mean “unreviewed”; sometimes they mean “patient still needs a callback.” Everyone knows not to move the stack beside the scanner because that stack is not clutter. It is the unofficial holding area for forms that need physician review before submission.

But “everyone knows” is not infrastructure. It is memory distributed across a particular group of people at a particular moment in time.

That distinction matters most when the clinic changes.

The MOA who owns the referral spreadsheet goes on leave. The person covering her has access to the file, but not to the logic inside it. She can see the colours, the dates, the names, and the notes. What she cannot see is the history behind them: which specialist office needs persistence, which referral is waiting because the patient asked for a pause, which item looks routine but has already been escalated twice.

The spreadsheet is still there. The process is not.

This is where hidden infrastructure becomes visible, usually because something stops working. A referral goes quiet longer than it should. A patient calls for an update and nobody can tell whether the clinic is waiting, watching, or already supposed to have acted. The covering staff member hesitates before changing a colour because changing the colour feels like making a decision she was never trained to make.

The clinic has not become disorganized. It has discovered that a critical process was living inside one person’s working knowledge.

The same thing happens with inbox habits. A physician leaves certain results unread because unread means “callback pending.” It is a practical signal, and for that physician it may work reliably. Then a new staff member, trying to be helpful, clears unread items as part of inbox cleanup. From their perspective, they are reducing clutter. From the physician’s perspective, they have erased the signal that something still needed to happen.

Nobody acted carelessly. The process was simply invisible.

When workarounds become infrastructure, the clinic often depends on signals that are not self-explanatory. A paper folder becomes an escalation queue. Appointment notes flag forms that are not appointment-related. A naming convention tells experienced staff which scanned documents need action and which can be filed. These signals work because people interpret them correctly.

That is also why they break.

They break during absences, because the person who knows the routine is away. They break during onboarding, because new staff are taught the formal system first and the real system later, often by overhearing how others handle exceptions. They break during growth, because what three people can hold through shared context becomes harder when six providers, two part-time MOAs, a locum, and a clinic manager all touch the same work. They break when volume increases, because a side list that was manageable at 20 items becomes unsafe at 80.

The clinic may still look functional from the outside. Patients are being booked. Results are being reviewed. Referrals are being sent. Forms are being completed. But under the surface, certain responsibilities depend on local knowledge that no one has fully named.

That creates a specific kind of mental load: not just doing the work, but remembering how the hidden system works.

Someone has to remember that the blue folder is routine but the green folder needs same-day review. Someone has to remember that the physician who leaves results unread is not behind; they are using unread as a safety mechanism.

The more the clinic depends on these signals, the more fragile it becomes when they are misunderstood.

This is why the life cycle of a workaround matters. It begins as a small adaptation to a specific problem. Over time, if the formal system never absorbs that responsibility, the workaround becomes routine. Then it becomes expected. Eventually, the clinic cannot easily remove it because too much work depends on it.

By then, the workaround is no longer a side system.

It is infrastructure without the safeguards of infrastructure.

Real infrastructure is shared, visible, teachable, and able to survive absence. Hidden infrastructure concentrates knowledge in people’s heads. It makes ownership harder to see. It turns onboarding into translation. It allows important work to live in places that look ordinary until you understand the private meaning attached to them.

A clinic can run this way for a long time, especially when the team is experienced and committed. In fact, many clinics do. The danger is that experience can hide fragility. A system held together by excellent people may look reliable because those people have become very good at compensating for it.

Compensation is not the same as reliability.

This matters for clinic owners and managers because the risk often appears when the clinic is already under strain. A staff member leaves. A provider joins. A physician takes vacation. Referral volume increases. Suddenly, the side system that once felt lightweight has to carry more people, more exceptions, and more ambiguity than it was ever built to handle.

The failure is rarely obvious at first. It shows up as hesitation. Rechecking. Duplicate calls. Staff asking, “Do you know what this colour means?” Physicians keeping separate lists because they are no longer sure the shared one is current. Managers discovering that a process they thought was stable was mostly stable because one person had been quietly maintaining it.

That is the moment worth paying attention to.

Not because the workaround was wrong. Because the clinic had come to depend on it without giving it the visibility, structure, and shared ownership that dependency requires.

At Aeon, this is the standard we keep returning to: when critical clinic work is being carried by a workaround, the system should help turn that work into something shared, visible, and reliable enough to survive absence.

That does not mean every side note, tracker, or habit needs to disappear. Clinics will always adapt. Primary care is too variable, too human, and too full of exceptions for any system to remove the need for judgment. The goal is to notice when a workaround has crossed the line from helpful adaptation to hidden infrastructure.

If a referral tracker is the only place the clinic can see what is waiting, that work needs a more reliable home. If unread results are being used to represent unresolved patient callbacks, the system needs a better way to distinguish review from resolution. If a paper folder has become the escalation queue, the clinic should not have to rely on the folder staying on the right desk.

Workarounds often survive because they are useful. That is precisely why they deserve scrutiny once the clinic starts depending on them.

The question is not, “Why are we using this?” The better question is, “What would break if the person who understands this was away tomorrow?”

That question makes the invisible system visible. It shifts the conversation from blame to dependency. It helps a clinic see which routines are merely convenient and which ones are quietly carrying patient care.

Because the risk is not that clinics invent workarounds.

The risk is that some of the most important work in the clinic may already depend on them.