The one dataset that exists #
NHS England Digital publishes Appointments in General Practice every month. Appointments in it are marked as attended or did not attend, down to practice level, and the July 2026 release came out on 27 August 2026. It is the closest thing the UK has to a routine national count of missed appointments.
It is also carefully hedged by the people who publish it. It is labelled official statistics in development. It covers only practices using participating clinical systems. And its own summary says plainly that it does not show the totality of general practice activity, because it contains only what was captured on appointment systems.
The one sector with a published cost attached #
Dentistry has a figure, and it is worth knowing because it is expressed in hours rather than in a percentage. Research carried out by the British Dental Association in 2011, quoted in a letter in the British Dental Journal on 12 September 2014, counted the clinical time that failed attendances cost.
81 hours
Average lost time per full time equivalent dentist per year, in practices deriving half or more of their income from the NHS
British Dental Association research, 2011, quoted in the British Dental Journal, 12 September 2014
69 hours
Per dentist per year in practices with lower NHS commitments
British Dental Association research, 2011, quoted in the British Dental Journal, 12 September 2014
Read that carefully before you reuse it. It is research from 2011 reported in a letter rather than a study published in its own right, and it describes NHS and mixed practices under a contract that does not let dentists charge for failed NHS appointments. It is not a figure for a private aesthetics clinic or a physiotherapy practice.
What it is good for is the unit: hours of clinical time, per clinician, per year. That is the unit your own diary can produce, and it is far more useful in a conversation than a percentage.
Which is why the statistics in your inbox do not add up #
Marketing material in this sector recycles no show percentages endlessly, and the figures disagree with each other because most of them cannot be traced to a named study with a published method and a sample size.
We apply that rule to ourselves: there is no invented number anywhere on this site, which is why this page is shorter on percentages than the ones competing with it.
The number that is actually yours #
Your practice management system already holds every booking and every outcome, so the count is a counting exercise rather than a research project. The splits that tend to change what you would do about it:
- By clinician, because the pattern is rarely even across a team.
- By appointment type, because a first appointment and a review behave differently.
- By lead time, meaning how many days passed between booking and appointment.
- By day and time, because the early and late edges of a session are not the middle.
- By whether a reminder was actually sent, and whether the patient could reply to it.
Those five splits will tell you more about your own diary than any published percentage, and they cost nothing but the extract. If you want to convert the result into hours and money before you speak to anybody, the free tools work from your figures rather than ours.
Which parts of the problem run on rules #
A no show is one event, but the admin around it is a chain of small, dated checks, and every link in that chain is the sort of thing that gets missed when the front desk is busy:
- An appointment two days out that has had no reminder sent.
- A reminder that went out and got no acknowledgement at all.
- A cancellation that leaves a gap nobody has tried to fill.
- A patient who did not attend and has had no contact since.
- A course of treatment that stopped part way through with no next appointment.
- A recall interval that has elapsed with no invitation sent.
Each of those is a rule you could write on an index card, applied to data that already exists. That is the definition of rule based work. Two of them get their own pages here: appointment reminders and filling a gap at short notice.
The first of those matters more than the marketing suggests. Reekie and Devlin tested reminder methods across 2,500 appointments in a single dental practice and published the result in the British Dental Journal on 14 November 1998. No reminder method beat any other.
9.4%
Failed attendance with no reminder
Reekie and Devlin, British Dental Journal, 14 November 1998
3%
Failed attendance with a reminder, at minimum
Reekie and Devlin, British Dental Journal, 14 November 1998
Which parts do not #
- Whether to charge a fee, and to whom.
- Whether a repeated non attender should keep being offered appointments.
- Whether somebody who is missing appointments is struggling rather than careless.
- Whether a missed appointment matters clinically.
All judgement, all for the practice, and none of it improved by being made faster.
It is also worth saying where the boundary sits in data protection terms. The fact that a named person has an appointment at a clinic is information about their health, which makes it special category data under UK GDPR.
Your practice stays the controller for it whichever supplier handles your reminders, so a lawful basis and an Article 9 condition are needed, and a data protection impact assessment is required before processing that is likely to result in a high risk. The patient data page points at the ICO's own guidance on all three.
What to do with all this #
Count your own rate, split it five ways, and look at the chain above for the links your practice never gets round to. That tells you whether you have a no show problem or a follow up problem, and they are not the same thing and do not have the same answer.