Why Wait Times Get Long (and the Fixes That Hold)

The strange thing about a long wait is that nobody caused it. The providers were working the whole time. The desk never sat idle. And yet Medimap's Wait Time Index put the average Canadian walk-in clinic wait at 68 minutes in 2023, with Ontario at 59 minutes, more than double the 25 minutes of a year earlier, and Toronto the worst in the province at 72 minutes. Waits that long are not produced by slow people. They are produced by structure, which is good news, because structure can be changed.

This piece is about the why. If you want the cost of the walkaways those waits produce, that is a different article. Here we take the machine apart: where a backlog comes from, why it refuses to clear, and which repairs actually hold once the next rush hits.

Arrivals cluster. Capacity does not.

A clinic that can see six patients an hour and averages five arrivals an hour looks safe on paper. It is not, because that five is an average, and averages arrive in clumps. Nine people walk in between 4:30 and 5:30 when work lets out; nobody comes at two. During the clump, arrivals outrun service and a backlog forms. That much is obvious. The part that surprises people is what happens next: the backlog does not clear when arrivals return to normal, because normal already uses nearly all the capacity. Burning off a queue needs a genuinely quiet stretch, and on a busy day the quiet stretch never comes before close. A wait built in one bad hour is still being paid at eight in the evening.

Capacity is lumpier than the staffing sheet admits, too. Providers take breaks, a procedure runs long, a complex visit eats three ordinary slots. The margin between average demand and real capacity is thinner than it looks, which is why the same clinic can feel fine on Tuesday and drown on Wednesday with identical staffing.

Check-in is serial; arrivals are not

Before anyone waits for a provider, they wait for the desk. Arrivals happen in parallel; a counter works one person at a time. Say the after-work clump delivers five people in four minutes and each registration takes two: the fifth person does not reach the front until minute eight, waiting to be allowed to start waiting. Worse, the desk does double duty. The same person processing new arrivals is fielding how-much-longer questions from everyone who arrived earlier, so the queue's own anxiety slows the queue's intake. And on a paper sign-in sheet the failure compounds quietly: nobody timestamps the walk through the door, so those eight minutes at the counter never appear in any measurement and the clinic cannot see the stage where its wait actually begins.

Batching turns a trickle into lumps

Work done in batches saves the worker and costs the queue. Rooming patients in threes because the walk to the back is annoying. Writing up charts in a block at the top of the hour. Closing intake for a twenty minute break and reopening to a lobby that kept arriving the whole time. Each of these converts smooth flow into a lump, and the arithmetic from the first section says lumps are precisely what create backlogs. The lunchtime closure is the classic: demand does not pause when you do, so the clinic effectively schedules its own rush for 1 p.m. every single day and is then surprised by it.

No-shows and walk-in collisions waste capacity twice

A no-show is capacity wasted in the one way you cannot recover. The reserved slot sits empty while the lobby is full, because by the time the desk is certain the patient is not coming, the window to pull a walk-in forward has mostly closed. Fifteen dead minutes at 3 p.m. is fifteen extra minutes for everyone still waiting at seven.

The collision is the second waste. When appointments live in one book and walk-ins on another list, the front desk becomes the integration layer, reconciling the two by eye. A booked patient arrives on time to a lobby of walk-ins who have waited fifty minutes, and someone has to decide who goes next. Whichever way they decide, somebody feels robbed, and the deciding itself takes time the desk does not have. Two scheduling systems sharing one physical lobby will always fight; the lobby is just where the fight becomes visible.

The wait is also longer in the mind

The clock is only half the wait; the other half happens in the patient's head. David Maister's best-known observation about queues is that an uncertain wait feels longer than a known one, and the traditional lobby manufactures uncertainty by default: no posted position, no estimate, and no way to tell whether the person just called ahead of you jumped the queue or was triaged. Patients sit with nothing to do but audit the room. Forty real minutes are experienced as an hour, and the experienced wait, not the measured one, is what drives the walkout and the review.

The fixes, ordered by cost

Each cause above has a repair, and they conveniently sort by price. Working up from free is also the sensible order in which to reduce wait times without hiring anyone.

Free: measure, then tell people the truth. Time your real service time and your busiest hour instead of guessing at both, and put the current wait where arriving patients can see it. A posted 45-minute estimate is tolerated far better than a mysterious one, and an honest number shifts some arrivals to quieter hours all by itself. Wait-time software produces the measurements as a by-product, because it timestamps every arrival, call and completion.

Cheap: parallelize check-in. A QR code at the door lets ten arrivals register from ten phones at once, so the desk stops being the first queue and handles only exceptions. This kills the serialization problem outright and costs a printed poster.

Moderate: move the wait out of the lobby. A virtual queue with a live position and a text when the turn is near attacks the psychology directly, and it smooths the flow of returns so a provider is not idle while the next patient is found in the parking lot. It does not add capacity; it stops capacity leaking between visits.

Most involved: put appointments and walk-ins in one queue. One ordered list ends the collision, and it converts no-show gaps from dead time into walk-in time, because the next walk-in flows into the space automatically. This is the structural fix for the structural cause, and it is the one that holds when flu season arrives, because it does not depend on anyone at the desk making judgment calls under pressure.

Frequently Asked Questions