England recorded 8.1 million outpatient did not attends in 2024-25. Set that number next to the elective waiting list, which sits at roughly 7.4 million pathways, and the comparison does most of the work on its own. The volume of missed appointments is running at close to the same scale as the backlog those appointments are meant to clear. That is the case for treating the work to reduce outpatient DNAs as a capacity strategy, not as a patient behaviour problem.
The framing matters because it changes who owns the fix. A behaviour problem belongs to patients. A capacity problem belongs to operations. And capacity is the thing every transformation lead, operations director, and ICB performance lead is already accountable for.
There is a second number worth holding alongside the first, and it is the one this post is built around. But before getting to it, it is worth being precise about what a DNA actually costs, because the headline figure understates it.
What a missed outpatient appointment actually costs
A DNA is not one lost slot. It is a slot that was protected, staffed, and then left empty while someone else waited for it. The clinic ran. The consultant was present. The room was booked. The administrative time to schedule, confirm, and chase had already been spent. None of that is recoverable.
The GIRFT outpatient work put a usable figure on the upside. A 25% reduction in DNAs would release the equivalent of close to 2 million outpatient appointments. At a national level, that is a larger return than most service redesigns deliver, and it requires no new estate, no new consultant sessions, and no new clinical workforce. It requires the slots that already exist to be used.
This is the relief in the data, and it is worth stating plainly. A trust hitting an 8% DNA rate is not running a badly managed service. The national average sits at roughly 7 to 8%, and it has been broadly stable for years. The rate is a structural feature of how outpatient demand and patient contact currently interact. It is not a verdict on any individual booking team. What it represents is unclaimed capacity sitting inside a system that is short of exactly that.
So the question is not why patients miss appointments. It is which patients are missing them, and why the contact methods currently in use are failing to reach those specific groups in time to do anything about it.
The shadow waiting list
Here is the second number. Every DNA is a slot that could have treated someone already on the list. When a patient does not attend, two things happen at once. One patient loses an appointment they needed. A second patient, further down the list, loses the chance to be moved up into it. The missed slot is not just waste. It is a waiting list entry that never gets cleared, multiplied across millions of appointments a year.
That is why DNAs function as a shadow waiting list. They are a backlog hiding inside the schedule rather than in front of it. The official list counts the people waiting for a slot. The DNA rate counts the slots that were available and went unused. Reduce the second and you have, in operational terms, found capacity to address the first. No procurement of new clinical sessions delivers that conversion as directly.
This is the argument that makes the ROI conversation simple internally. Most capacity interventions require spending to create supply. DNA reduction recovers supply that has already been paid for. The investment case is not “spend to add clinics.” It is “spend to stop losing the clinics you already run.” Those are very different conversations to have with a finance director.
Before getting to the three points where that recovery actually happens, it is worth explaining why the tool most trusts rely on does not reach the patients who drive the rate.
Why SMS-only reminders leave the hardest cohorts uncovered
SMS reminders are the right baseline. They are cheap, they scale, and the evidence shows they reduce non-attendance against no reminder at all. Every provider should be sending them. The problem is not that SMS fails. The problem is what SMS does not do, and which patients fall into that gap.
A text reminder confirms that a message was delivered. It does not confirm that the appointment still works for the patient, and it does not create a moment where the patient can say so. The systematic review evidence is specific on this point: personal phone reminders significantly increase the rate at which patients cancel and rebook, with cancellation or rescheduling rates of 17 to 26% among patients who received a phone call, against 8 to 12% for those who received nothing. A text does not produce that effect, because a text is not a conversation. It carries information one way and then stops.
That matters because a cancelled appointment with notice is not a DNA. It is a recovered slot. The patient who cannot make Tuesday at 10am and says so on the phone has just handed that slot back to the booking team in time for it to be reused. The patient who receives a text, cannot make Tuesday, and does nothing has become a DNA. The difference between those two outcomes is whether a two-way contact happened, not whether a reminder was sent.
Then there is the cohort question, which is where the equity dimension enters. Around 11 million people in the UK, close to 20% of the population, lack basic digital skills or do not use digital technology at all, and that group skews older, in poorer health, and more deprived. These are not edge cases. They are disproportionately the same patients who appear in higher DNA rates. The reminder review found that more intensive contact is specifically warranted for deprived groups, ethnic minority groups, and patients with comorbidities, the exact populations a text-only approach reaches least reliably.
Phone-based contact also outperforms SMS for older patients in particular, where attendance after a personalised phone reminder runs materially higher than after a text. For patients whose first language is not English, a structured voice conversation can flex in a way a templated text cannot, confirming details, answering a question, and resolving confusion in the moment.
So an SMS-only model has a predictable failure shape. It works well for digitally confident patients who were likely to attend anyway, and it works least well for the cohorts who actually drive the DNA rate. The result is a reminder system that looks active on paper while leaving the highest-risk slots exposed. Closing that gap is not about replacing SMS. It is about adding the channel that covers the patients SMS structurally misses.
The three intervention points where DNAs are actually recovered
Reducing outpatient DNAs is not a single action. It is three distinct contact moments, each catching a different failure mode. A serious DNA strategy addresses all three rather than relying on one.
1. Confirmation calls 48 to 72 hours before the appointment
This is the highest-yield contact and the one SMS handles worst. A confirmation call 48 to 72 hours out does three things a text cannot. It confirms the patient still intends to attend. It surfaces the patient who needs to cancel in time for the slot to be rebooked. And it catches the patient who has forgotten, lost the letter, or never received it. The 48 to 72 hour window is deliberate. It is late enough that the patient knows their plans, and early enough that a cancelled slot can still be filled from the list.
The operational value sits in the cancellations, not just the confirmations. Every patient who uses that call to hand back a slot is a DNA prevented and a waiting list entry cleared in the same motion.
2. Rebooking conversations
When a patient does cancel, or when a slot opens up, the next failure point is rebooking. A cancelled appointment that is not promptly rebooked is still a lost slot, just a lost slot the patient cannot be blamed for. The rebooking conversation, ideally in the same contact as the cancellation, captures the recovered capacity before it leaks away. This is where the cancel-and-rebook rate from phone contact turns into real reallocated slots rather than empty gaps in the next clinic.
3. Patient-initiated follow-up safety-net calls
The third point connects directly to national policy and is where the DNA conversation and the PIFU conversation merge. Under the NHS Elective Reform Plan, providers are expected to offer patient-initiated follow-up as standard across all appropriate pathways by March 2026, and to increase PIFU uptake to at least 5% of all outpatient appointments by March 2029. The policy language is explicit that this increase should come in part through the enhanced identification of suitable patients using AI and automation.
PIFU reduces fixed follow-up appointments, which directly reduces the pool of appointments that can be missed. But PIFU has a known weak point: the safety net. A patient on a PIFU pathway is responsible for initiating contact when they need to be seen, which works well until a patient who should make contact does not. A structured safety-net call, checking in on PIFU patients who have gone quiet, catches the people who have fallen through the self-management model before that gap becomes a clinical risk. It is the human reassurance layer that makes PIFU safe to scale toward the 5% target.
It is worth being honest about the limits here. PIFU is appropriate for specific stable pathways, not universally, and the national experience shows the gap between target and delivery is real. The 2022/23 guidance set a 5% PIFU ambition for March 2023, and in practice closer to 2.5% was achieved. The lesson is not that the model fails. It is that the model needs an operational engine behind it, including the proactive contact that keeps PIFU patients safe. That engine is exactly the outbound capability this post is about.
The operating model that covers all three points
Each of the three intervention points is an outbound call. Confirmation, rebooking, and PIFU safety-net contact are all proactive, structured, two-way conversations, mostly with the cohorts that text reminders reach least well. The constraint is obvious to anyone who has run a booking team. There is no spare administrative capacity to make that volume of calls. The same staff who would make confirmation calls are already absorbed handling inbound demand, processing referrals, and managing the queue. Outbound contact is the work that gets dropped first when capacity is tight, which is precisely why DNA rates stay stable year after year.
This is the structural reason the problem persists. It is not that trusts do not know phone contact works. It is that the staffing to do it at scale does not exist alongside everything else the team carries.
The operating model that resolves this is outbound AI voice with human escalation, running alongside the booking and admin team rather than in place of it. The team is the team. What changes is that the routine, high-volume confirmation and reminder calls are handled in parallel as concurrent capacity, while the calls that need judgment, sensitivity, or clinical context route back to a person with the full context attached. The AI layer absorbs the volume that was never going to get made by hand. The human layer handles the conversations that need a human. Neither replaces the other.
This is where Jackie fits. Jackie runs outbound campaigns of exactly this kind, holds a natural two-way conversation, confirms or cancels in the clinical system, and escalates to the booking team when a call needs a person, with structured context already captured. The point is not automation for its own sake. It is that the confirmation calls, rebooking conversations, and PIFU safety-net contact that recover DNAs only happen if there is capacity to make them, and that capacity has not previously existed at the volume the problem requires. For trusts working toward the March 2029 PIFU target with AI-assisted patient identification written into the policy, this is the operating model the policy is already pointing at. You can see how the outbound side works in more detail on the Jackie outbound page, and the inbound equivalent in the Park Street deployment.
What good looks like in numbers
The internal business case for DNA reduction is unusually clean. Take a trust running 200,000 outpatient appointments a year at an 8% DNA rate. That is 16,000 missed appointments. A 25% reduction, in line with the GIRFT figure, recovers 4,000 appointments. Those are 4,000 slots returned to the waiting list with no new clinical capacity purchased.
The cost comparison that matters here is not the AI against a member of staff. The booking team is not the alternative, and holding admin headcount flat on a phone they do not have time to use is not a strategy. The alternative is the status quo: thousands of recoverable slots lost every year because the outbound contact to save them is not being made. The comparison is between the recovered capacity and the cost of the contact that recovers it, and at the scale outpatient services operate, that arithmetic is rarely close.
This is the post that pairs with an outbound ROI model, because the inputs are all knowable: appointment volume, DNA rate, cost per appointment, and target reduction. Plug in a trust’s own numbers and the recovered-capacity figure falls out directly.
Frequently asked questions
What is the average DNA rate for NHS outpatient appointments?
The national outpatient DNA rate sits at roughly 7 to 8% and has been broadly stable in recent years. In 2024-25 there were 8.1 million outpatient DNAs in England. Individual trusts vary widely, with some specialties and some sites running well into double figures, so the national average is a starting benchmark rather than a target.
Why do SMS reminders not fully solve DNAs?
SMS reminders reduce non-attendance compared with no reminder, but they carry information one way and do not create a moment for the patient to cancel and rebook. They also reach digitally excluded cohorts least reliably, and those cohorts, older, more deprived, and non-English-first-language patients, are disproportionately represented in higher DNA rates. Phone contact closes both gaps.
How does reducing DNAs help with the elective waiting list?
Every recovered slot can be used to treat a patient already on the list. Because missed appointments run at close to the same scale as the waiting list itself, reducing DNAs converts directly into usable capacity. A 25% reduction nationally would release the equivalent of close to 2 million appointments without adding any new clinical sessions.
What is the connection between DNAs and PIFU?
Patient-initiated follow-up reduces the number of fixed follow-up appointments, which shrinks the pool of appointments that can be missed. PIFU’s weak point is the safety net for patients who should make contact and do not. Structured safety-net calls catch those patients, which is why proactive outbound contact is part of scaling PIFU safely toward the March 2029 target.
Can outbound calling be automated without losing the human element?
Yes, in a hybrid model. AI voice handles the high-volume routine confirmation and reminder calls as concurrent capacity, and escalates to the booking team for any call that needs human judgment, with the context already captured. The admin team is not replaced. The calls that would never have been made by hand get made, and the team handles the conversations that need a person.
See the outbound model on your own numbers
The DNA problem is quantifiable, the recovered capacity is quantifiable, and the contact that recovers it is now operationally possible at scale. If you want to see what a 25% DNA reduction looks like against your trust’s own appointment volume and DNA rate, our outbound ROI model runs those figures in a few minutes, and a 20-minute demo walks through how the confirmation, rebooking, and PIFU safety-net calls actually run on a live outbound queue. No new telephony, no hardware, and the four-week pilot runs on your existing systems.