The two UK studies worth knowing #
| Study | What it measured | What it found |
|---|---|---|
| Reekie and Devlin, British Dental Journal, 14 November 1998 | 2,500 appointments in one dental practice, 130 failures | Failed attendance fell from 9.4 per cent with no reminder to 3 per cent with one, and no method beat another |
| NHS Fife paediatric physiotherapy, CSP Frontline, 5 July 2017 | A musculoskeletal service that had lost staff while referrals rose | Missed appointments fell from 22 to 13 per cent, and wasted hours by 53 per cent |
Both are single services describing their own results, one of them nearly thirty years old and the other from an NHS setting rather than a private one. Neither is a forecast for your practice. What they are good for is the direction and the rough size of the effect, and for settling an argument that is often had on instinct: reminding people does change whether they turn up.
The second detail in the older study is the one the marketing never mentions. Reekie and Devlin compared reminder methods and found that none outperformed another.
Why reminders stop happening #
The clearest published account of this comes from a dental practice manager, Lisa Bainham, speaking to Dentistry.co.uk on 20 September 2021. Her practice used to ring patients a few days before an appointment as a courtesy. With things being so busy, she said, that had not happened, and cancellations had gone up.
That is the whole mechanism in one sentence, and it is worth sitting with because it is not a technology problem. Reminding is a discretionary task performed by the person who is also answering the phone, greeting patients and taking payments. When the day gets heavy, the discretionary task is the one that goes, and the consequence arrives two days later as an empty chair.
One word, five separate jobs #
"Reminders" sounds like one task, which is why practices buy one product for it and are surprised when the problem persists. It is five, and they fail independently:
| The job | How it fails | Rule based? |
|---|---|---|
| Sending the reminder | Nobody had time, or the appointment was booked after the send | Yes. It is a date and a message |
| Reaching the patient | The mobile number on file is old, or was never captured | Partly. Spotting a missing or malformed number is a check |
| Getting a reply | The reminder is one way, so silence means nothing | Yes. Whether a reply arrived is a fact |
| Reading the reply | It lands in an inbox nobody owns | Partly. Routing is a rule, reading is not |
| Acting on the reply | A cancellation is noted and the gap is left | No. Filling a gap is a decision and a conversation |
Look down the third column. The parts that run on rules are the bookkeeping parts, and the bookkeeping is what a busy front desk drops first. The part that needs a person, acting on what the patient said back, is the part practices are usually happy to do and rarely get the chance to.
Which checks are genuinely mechanical #
- An appointment inside the reminder window with no reminder recorded against it.
- A patient record with no mobile number, or one that has failed to deliver before.
- A reminder sent with no acknowledgement of any kind.
- A reply that arrived and has not been opened by anybody.
- An appointment booked so late that the normal reminder run has already gone.
- A cancellation reply that has not yet produced an attempt to refill the slot.
Every one of those is a question with a yes or no answer, asked of data the practice already holds. None of them requires judgement, and none of them touches a clinical decision. That is the test this whole site applies, and it is set out on AI automation, explained.
Which parts are not candidates #
- What to say to somebody who has missed three appointments.
- Whether a patient who keeps cancelling is unreliable or unwell.
- Whether a reminder is appropriate at all, which is a real question for some kinds of appointment and some kinds of household.
- Whether to charge.
Those are judgements about a person, and speeding them up would not improve them.
The data protection line #
A reminder message is a small thing that carries a large category of data. The fact that a named person has an appointment at your clinic is information about their health, which makes it special category data under UK GDPR, so a lawful basis and an Article 9 condition are both required. Your practice remains the controller whoever sends the message on your behalf. The patient data page points at the ICO's own guidance rather than interpreting it here.
What to look at first #
Before changing anything, count two numbers from your own system: how many of last month's appointments had a reminder recorded against them, and how many of those reminders could be replied to.
If the first number is below your intention, you do not have a reminder problem, you have a capacity problem showing up as one, and the no show figures page explains why the published percentages will not help you size it. The free tools will turn your own count into hours without asking you for anything.