Cutting Failed Appointments With Automated Reminders

Most practices that set out to reduce dental FTA rate start by rewriting the message. Someone on the team, usually the practice manager who has been staring at the daily DNA list for eighteen months, opens the recall template in SOE or R4 and starts editing. “Please attend” becomes “Your dental appointment is confirmed.” Someone suggests adding the cost of a missed appointment. Someone else suggests removing it, because it sounds threatening. The word “hygienist” gets swapped for “hygiene visit”. Three weeks later the FTA rate has moved from 9.1% to 8.8% and nobody can tell whether that was the wording or the fact that February half term ended.

The wording was never the variable. Cadence was.

What the timing data actually shows

Here is the pattern that turns up across practice management data with irritating consistency. A single reminder sent 24 hours before the appointment recovers a fraction of what a two-touch sequence recovers, and a three-touch sequence beats both. Not because patients need telling three times, but because each touch catches a different failure mode.

Think about why a patient doesn’t turn up. There are broadly four reasons, and they surface at different points in the countdown:

A 24-hour SMS only addresses the fourth row and part of the third. It arrives far too late for the patient who realised on Tuesday they had a shift clash on Friday. By then the slot is dead, because even if they cancel at 5pm the day before, front desk has twelve hours to fill a 40-minute crown prep. That is why single-reminder practices see FTA rates that hover stubbornly in the 7–12% band regardless of how warm and well-written the message is.

A worked example from a three-surgery mixed practice

Take a practice running roughly 240 booked appointments a week across three surgeries, mixed NHS and private, with a baseline FTA rate of 8.5%. That is around 20 failed appointments a week. On NHS UDA throughput alone, with a Band 2 course at roughly 3 UDAs and a contract value around £30 per UDA, a week of failed Band 2 slots represents real contract risk before you count the private hygiene and cosmetic work sitting in the same diary.

The practice changes nothing about the wording. Same template, same tone, same signature. It changes the schedule:

BEFORE
  T-24h    SMS: "Reminder: dental appt tomorrow 10:20. Reply CANCEL."
  FTA rate: 8.5%  |  ~20 failures/week

AFTER
  T-7d     SMS + email: appointment confirmed, link to reschedule
  T-3d     SMS: "Still OK for Fri 10:20? Reply Y to confirm / N to move"
  T-3h     SMS: short nudge, time + address only
  FTA rate: 4.1%  |  ~10 failures/week

Ten slots a week recovered, roughly 430 a year allowing for closures. The wording did not change. Only the number of touchpoints and their spacing did.

The interesting number is not the headline drop. It is where the recovered slots come from. In practices that instrument this properly, the T-3d touch is the workhorse: it produces the highest volume of cancellations, which sounds like failure and is actually the win. A cancellation 72 hours out is a fillable slot. An FTA is a dead 40 minutes and a clinician doing admin. Practices that measure only FTA and ignore cancellation lead time consistently misread which reminder is earning its keep.

Why the T-3d touch outperforms

Three days is long enough to refill and short enough that the patient’s diary is real. At seven days out, a patient genuinely does not know whether their Thursday will hold. At 24 hours they know, but you cannot act on the information. The 72-hour window is the only point where the patient’s certainty and your ability to fill the gap overlap.

This is also where the reply mechanic matters more than the prose. A T-3d message that asks for a binary response, Y or N, gets a response rate several times higher than one that says “please call the practice if you need to change your appointment.” Front desk phone lines are open roughly 40 hours a week. SMS is open always. Every message that routes a cancellation through a phone call loses the patients who decide at 9pm.

Channel, and the mistake most practices make with it

SMS still does the heavy lifting in UK dental, and the reason is unglamorous: near-universal read rates within minutes, no app, no login, works on a nine-year-old handset. Email open rates in dental patient bases typically run somewhere in the 20–35% range, and a good chunk of those opens happen too late to matter. WhatsApp performs well where practices already have opt-in at scale, but opt-in at scale is the catch.

The error is treating channels as interchangeable and picking one. They are not interchangeable, they are complementary, and the sequence should match channel to job:

• T-7d: email plus SMS. Email carries the detail (what the appointment is for, what to bring, pre-op instructions, the fee estimate for private work). SMS carries the fact that an email exists. Email is the only channel where a patient can usefully read 200 words about a sedation appointment.
• T-3d: SMS only. Short, one question, one-tap reply. Email here dilutes the response rate because it creates two places to answer.
• T-3h: SMS only. Time, address, nothing else. Do not include a cancel link at three hours; you are inviting a cancellation you cannot fill.

That last point trips people up. A reminder is not always asking for the same thing. The early touches are asking “is this still right?” and the late touch is asking “are you on your way?” Practices that send an identical message three times get diminishing returns and a reputation for spamming, which is how the “patients hate reminders” objection gets born.

Tools, and what they actually give you

The UK market here is reasonably mature, and most of what follows integrates with SOE Exact, R4 or Dentally rather than replacing them.

Dentally has reminder sequencing built in with configurable multi-stage timing and two-way SMS, which makes it the lowest-friction option if you are already on it. Cloud-native, so the reminder logic and the diary are the same system and a cancellation refills without a sync step.

Software of Excellence (SOE) Exact with the Patient Portal and reminder modules supports multi-stage reminders, though older on-premise setups are frequently configured with a single 24-hour SMS because that is what was switched on at install and nobody revisited it. If your FTA rate is stuck and you are on SOE, check the reminder schedule before you change anything else. It is very often still the default.

Carestream R4 similarly supports staged reminders via its communications module, again with the caveat that the configuration is where the value lives.

Zesty and Clearstep-style patient engagement layers sit on top of the PMS and add self-service rebooking, which is the piece that converts a T-3d “N” reply into a filled slot rather than a hole. Connect My Patients and similar UK-focused messaging platforms do the same job with a front-desk-first interface.

For the refill side, cancellation-fill automation is where the arithmetic gets good. A practice with a 60-patient short-notice list and automated broadcast to that list fills a meaningful share of 72-hour cancellations without a single phone call. This is the point where reminder cadence and the rest of your front desk stack stop being separate projects; the wider set of choices around recall, waiting lists and inbound triage is covered in our Front Desk and Recall Automation pillar, and the cadence work below sits inside it.

The measurement discipline that makes this work

You cannot improve cadence without measuring two things most practices don’t currently log.

Cancellation lead time. Not “how many cancellations” but “how many hours before the appointment.” Bucket them: over 72 hours, 24–72 hours, under 24 hours, FTA. Your goal is to shift mass leftward from the FTA bucket into the 24–72 bucket. A practice whose total cancellation-plus-FTA volume stays flat while lead time improves is winning, even though the top-line “appointments lost” number looks unchanged.

Per-touch attribution. Which reminder produced the response. Any platform with two-way SMS can tell you this if you look. The practices that get FTA below 4% are the ones who discovered their T-7d email was doing almost nothing and their T-3d SMS was doing everything, then reallocated accordingly.

Here is what a month of properly bucketed data looks like in practice:

Sept 2026 — 1,040 booked appointments
                       Count    % of booked
  Cancelled >72h         74         7.1%     ← 68 refilled
  Cancelled 24-72h       39         3.8%     ← 22 refilled
  Cancelled <24h         21         2.0%     ←  4 refilled
  FTA                    43         4.1%     ←  0 refilled

  Net lost chair time:   83 slots (8.0%)

  Response by touch:  T-7d  11%   T-3d  34%   T-3h  6%

Before the cadence change, the same practice’s cancellations clustered under 24 hours and the FTA row read 88. The total “patients who couldn’t attend” barely moved. What moved was when they told you.

Where the ceiling is

Cadence gets you from roughly 8–9% to roughly 4%. It does not get you to zero, and the remaining failures are a different problem: a small cohort of repeat non-attenders, typically 3–5% of the active list, who account for a disproportionate share of FTAs regardless of how many messages you send. That cohort needs a policy response (deposit for private work, short-notice-only booking, a conversation), not a fourth reminder. Practices that keep adding touches past three see response rates fall and opt-outs rise.

Two practical cautions. First, check your opt-out handling against UK GDPR and PECR before you increase message volume; more touches means more consent surface, and a patient who opted in to “appointment reminders” in 2019 has arguably opted in to three of them, but your privacy notice should say so plainly. Second, watch the T-3h send window against your first appointment of the day. A 07:00 SMS for a 10:00 appointment is fine. An 05:30 automated send for an 08:30 appointment will generate complaints, and most PMS reminder schedulers will happily do exactly that if you set the offset in hours and forget about early starts.

So if you are about to spend a fortnight workshopping reminder copy with the team, open your reminder schedule first and count the touches. If there is one, that is your project. If there are three but they all say the same thing at the wrong intervals, that is your project. The copy can wait.