Cancellation Backfill When the Slot Is Worth the Most

In ophthalmology the slots are not interchangeable, and cancellation backfill fails hardest on the ones worth the most. A cancelled routine follow-up is a small loss. A cancelled procedure slot with dedicated room and equipment time is a hole nobody can fill by lunchtime, which is roughly how long the practice has.

The recovery window is the whole problem. A cancellation that arrives at 8am for a slot at 2pm is recoverable in principle and almost never recovered in practice, because working the waitlist requires a person with time, and the people who could do it are answering the phone.

So the practice recovers the easy cancellations, the ones with three days of notice, and loses the late ones. That is exactly backwards from where the value sits, because a late cancellation is the one where the room, the staff, and the equipment are already committed and paid for.

The industry response has been deterrence, not recovery

It is worth noticing where practices have put their attention, because it explains why backfill stayed manual.

A Jan. 7, 2025 MGMA Stat poll found that 42% of medical group leaders report their practices using a no-show fee, against 58% that do not. The poll had 622 applicable responses.

A fee is a deterrent and a partial recovery of administrative cost. It does not put a patient in the chair, and for a cancellation with notice it usually does not apply at all, because the patient did what the practice asked by calling ahead. Most recoverable volume in ophthalmology is exactly that: patients who cancel properly and leave a slot behind.

That leaves backfill as the only mechanism that converts the cancellation into a completed visit. It is also the mechanism most practices have never built, which means the attention has gone to penalizing the small share who fail to show while the larger recoverable share goes unworked.

Speed is a data problem before it is a staffing problem

Practices assume backfill is slow because nobody has time. The deeper reason is that the list cannot be queried.

A waitlist stored as names and phone numbers requires a human to adjudicate every candidate, which is why it only gets worked when someone has a free afternoon. To match automatically, an entry has to carry the appointment type the patient needs, the provider or provider group they can see, the location they can reach, how much notice they require, and the windows they are genuinely available.

With those attributes present, backfill becomes a query. A cancellation frees a slot with known properties, GET /appointments/waitlist returns the entries compatible with it, GET /appointments/open confirms the opening is still there, and PUT /appointments/{appointmentid} commits the booking once a patient accepts.

The speed follows from the structure. A query resolves in seconds and can run the moment the cancellation is recorded, at 6am or during clinic, without waiting for a person to be free. That is what makes the six-hour window reachable, and it is why storing the waitlist as structured entries against the schedule is the precondition for everything else here.

The complication: an ophthalmology slot is rarely one appointment

This is where generic backfill logic produces bookings that cannot be delivered, and ophthalmology gets hit harder than most specialties.

A large share of visits are a testing appointment plus a provider visit, booked as a pair with specific spacing. Imaging or field testing happens first, then the physician sees the patient with the results in hand. Nothing in the scheduling template enforces that pairing. It exists as convention, and the patient experiences it as one visit.

So when the provider half cancels and backfill fills only that half, the incoming patient arrives without the testing that makes the visit useful. The reverse is worse: filling a testing slot for a patient with no provider appointment behind it produces a study nobody is scheduled to review. Both look like successful recoveries in the schedule.

Dilation compounds it. A patient dilated for one appointment carries that constraint into the rest of their day, and the spacing between the two legs is not arbitrary. Equipment adds a third constraint, because procedure blocks have hard cutoffs driven by room prep and turnover rather than clinic hours, so the last genuinely bookable slot is earlier than the schedule suggests.

The handoff is the pairing. When a freed slot is one leg of a linked visit, the automation offers only candidates who can take both legs, and when no candidate can, it routes the slot to a coordinator with the constraint stated rather than filling half of it. A half-filled linked visit is not a recovery. It is a second problem discovered on the day.

Offer to several at once, and record the declines

Sequential offers are the other reason late cancellations go unrecovered, and the fix is uncomfortable until you see the arithmetic.

Calling one waitlisted patient, waiting for a callback, then calling the next spends the entire recovery window on people who did not answer. Three sequential attempts at twenty minutes each consumes an hour, and the slot was six hours out. Offering to several compatible candidates at once and taking the first acceptance is how the window gets used.

That requires the offer to be revocable in a way patients find acceptable, which is a wording problem more than a technical one. An offer framed as an available opening, first to confirm, is understood. An offer framed as a booked appointment that later evaporates is not, and doing that twice trains patients to ignore the outreach.

Declines are the asset nobody keeps. A patient who has declined three short-notice offers is telling you something the list should remember, and continuing to call them wastes the window on the least likely acceptance. Recording declines against the entry is a rule a list can hold and a person cannot, and it is what keeps offers per fill from climbing quietly over a year.

Measure the late window separately or you will fool yourself

The single most misleading metric here is an aggregate refill rate, because it is dominated by the easy cases.

Split refill rate by notice. Cancellations with several days of notice will refill well under almost any process, including a manual one. Cancellations inside twenty-four hours are the real test, and they are where the value concentrates, because those are the slots with committed room and staff time. A practice reporting a strong overall refill rate is frequently recovering nothing inside the day.

Then watch time to refill, which tells you whether the mechanism is fast enough to catch the late ones or is only working the comfortable ones. And watch offers per fill as the early warning that the list has gone stale, since a rising number means the entries no longer describe who actually wants an earlier appointment.

The honest check is refilled visits that were later cancelled, rescheduled, or denied at billing. Aggressive backfill that ignores the linked-visit pairing or the patient’s plan will look excellent on refill rate and reappear a month later as rework and denials. Counting that is what keeps a backfill program from optimizing itself into a different problem.

Key Takeaways

Measure refill rate by notice window. An aggregate number is dominated by easy multi-day cancellations and hides that same-day recovery is near zero.
Store the waitlist as structured entries carrying appointment type, provider group, location, notice required, and real availability. A name and a phone number cannot be queried.
Run the match the moment the cancellation is recorded, not when a person is free. The window on a late cancellation is hours.
Treat linked testing and provider visits as a pair. Filling one leg produces a study nobody will read or a visit without the testing behind it.
Respect the block cutoffs set by room prep and turnover. The last bookable procedure slot is earlier than clinic hours imply.
Offer to several compatible candidates at once and take the first acceptance. Sequential calling spends the whole window on people who do not answer.
Word the offer as an available opening rather than a confirmed booking, so taking the first acceptance does not read as a cancellation to everyone else.
Record declines against the waitlist entry and stop offering short-notice slots to patients who have repeatedly declined them.

Cancellation backfill is the highest-return scheduling work available to an ophthalmology practice, because both the demand and the capacity already exist and are simply failing to meet inside a window measured in hours. It stays undone because doing it properly means matching on constraints that live in several places, moving within minutes, and handling linked visits correctly every time. An AI team working freed slots against waitlist entries inside athenaOne can run that match the moment a cancellation lands, offer to everyone who genuinely fits, and hand a coordinator the slots where the pairing cannot be satisfied.

Sources