Out there somewhere?

NHS appointments problems worth owning

Following the second (unofficial) design history post on my time working with a team looking to bring more appointments into the NHS App, I felt there was a lot I didn't tap into, especially where next?

Here's a bit of a brain dump. This pulls together thoughts, information, and documents I've come across over the months: where things stand, what people say goes wrong, and where a bit of effort might pay off, especially given how little money and time local services have right now. Some of it is likely to be limited or flawed. It's here to start some thinking — and give my brain a little closure.

Where I've got to

National targets count how many appointments are in the App. I think that's the wrong measure. What matters is whether people can find, change and turn up to the right appointments, and whether they needed them at all. Most of that sits in admin and plumbing, and in services the App barely reaches yet.

Hospitals have come a long way

Every acute trust shows some appointments in the App, about two thirds of hospital appointments are visible, and views keep going up. Everything else builds on that. I think the next gains are in community care and in the admin around appointments.

Seeing appointments hasn't changed how people feel about NHS communication

App users and non-users rate it about the same. What people talk about is admin: late letters, nobody to call, trouble cancelling. A lot of the fix sits behind the scenes, in contact details, letters and the systems that send them.

The hardest problems sit between organisations

Carers managing someone else's care, home visits with no time, people dealing with several services at once, providers that aren't NHS trusts. None of them has one owner.

Targets count what's available

A 95% target doesn't tell you whether anyone's week got easier, or whether they needed all those appointments in the first place.

Locally, there's very little spare

Whatever gets asked of a trust, community provider or GP practice has to be cheap and pay its way within a year. Most will manage one thing well.

Some of the best bets aren't about patients seeing more

A shared view for whoever's coordinating someone's care, fewer appointments through better coordination, and being straight about what a list does and doesn't show.

Where things stand in October 2026

About two thirds of hospital appointments are visible and around half of trusts send App reminders. Views of hospital appointments and referrals in the App hit 8.7 million in March 2026, up 93% on a year. Booking, changing and cancelling depend on which portal a trust uses, and some changes wait for staff to process them.

Outside hospitals it thins out. Two community providers and 19 mental health trusts are live, against a record community waiting list of 1.21 million. Humber is building appointment and visit management for community and mental health trusts that use SystmOne, TPP's records system. Many community trusts use SystmOne, so this is likely to be the main way their appointments reach the App.

The national target is at least 95% of appointments available after triage, across all care settings. Depending on who's reporting it, that's from April 2026 or by 2028/29, and nobody has published what it's a percentage of. Meanwhile NHS England is moving messaging onto NHS Notify, and moving trusts away from separate patient portals towards direct integration with their records systems.

Note: When the team I was in started on all appointments early this year, the working assumption was every appointment a patient could get access to. Over time we pushed for appropriate appointments instead. Which raises its own question: what's appropriate, for whom, and who decides? Some appointments probably shouldn't be bookable in an app at all, some need a gate before anyone books, and some only make sense as something to see. The target's own wording, available after appropriate triage, hints at the same thing without saying what it means.

What local services have to work with

All of this lands on trusts, community providers and GP practices, at a time when they're being asked to do more with less.

• In March 2025, ICBs were told to halve their running costs by the end of that year, and their number is going from 42 to 26. Digital leadership is one of the functions being looked at for a move to providers.
• 86% of trust and ICB leaders say this year's savings will be harder than last year's, and 64% expect to cut or reduce services.
• In 2025, 81% of trust leaders said there wasn't enough money to invest in digital. They've also said small pots handed out late in the year don't help much.
• This year's national referrals and appointments funding is £18.5 million across 109 projects. About £170,000 each.

Problems worth owning

These are happening now, to a lot of people. Most have no clear owner. They're in rough priority order, by how many people each touches, whether there's money in it, whether a local team can start now, and whether it waits on anything national. Each one says where it stands (solved, part solved, solved in a that'll do way, or yet to be solved), where I'd start, and a question any plan for it should answer. The starters are small on purpose, given the money and time above. Some are a few months' work for one team, others need more. A funded national team could go bigger and faster, but most local teams won't have that. Whatever gets built, someone has to own it once the money runs out.

1. Letters and messages that arrive late, wrong or from everywhere

Nearly 2 in 3 people who used the NHS in a year had an admin or communication problem, and 1 in 5 were invited to an appointment after its date. Healthwatch reckon late letters are behind around two million of the eight million missed hospital appointments a year. Letters, texts, the App and portals arrive for the same care, and people find it confusing. A deaf patient is still phoned after asking for text or email, and 7 in 10 deaf people have never been asked how they'd like to get information.

Where it stands: Part solved. Providers are meant to move patient messages to NHS Notify by 2028/29, but the route for local services isn't settled. App messages are free, text and letter fallbacks aren't, and only 41% of App messages are read within a day. Royal Mail now has an NHS barcode to get letters through on time. Since the end of September, every publicly funded provider has been expected to read, write and share reasonable adjustments through the national record.

Where I'd start: Start with the plumbing, before any screens. Check contact details and communication needs are recorded and actually used when something gets sent. Send early enough, by the route that works for that person. A letter that's right and arrives in time is fine.

Any plan should answer: Who owns admin standards, and who reports against them?

Why first: It affects the most people and most of it is in local hands. See the worked case below.

2. No reliable way of knowing whether any of this helps

App users and non-users rate NHS communication about the same, and repeat polling found no improvement. The statistics regulator had to ask for the method behind the 1.5 million missed appointments claim, and the only national evaluation had to rely on aggregated data.

Where it stands: That'll do. Monthly statistics count registrations, logins and messages.

Where I'd start: A few local admin standards, reported against: every appointment has a way to contact someone, nothing arrives after the date, cancelling is always possible. The King's Fund and partners already want this reported.

Any plan should answer: What does the 95% actually count? Every appointment or only some, which settings, what "after triage" leaves out, and does available mean visible or bookable? And what gets measured instead of views and missed appointments?

Why second: Cheap, and without it nothing else can show it worked.

3. Not knowing when someone is coming

District nursing patients aren't given visit times, and the not knowing disrupts the day. One health board says a visit can be any time from 8am to 8pm.

Where it stands: Yet to be solved. Humber's work for SystmOne trusts will show visits in the App. I haven't found anything that offers an arrival window.

Where I'd start: One district nursing team sends a morning message with a rough window and a way to say not today. Count the wasted journeys before and after.

Any plan should answer: What changes for a housebound person with daily nurse visits, and for the relative who sorts them out?

Why third: Cheap to test, and nobody else is doing it.

4. Carers locked out

81% of carers in Derbyshire would use the App, or already do, to see appointments for the person they look after. In Coventry, proxy access is hit and miss and maternity needs a second app.

Where it stands: Part solved for GP services, yet to be solved elsewhere. A national proxy service becomes mandatory for GP practices from 2027. Hospital and community proxy access is still patchy.

Where I'd start: In community care, design for the carer first. Talk to ten carers of people getting home visits and map their week.

Any plan should answer: Who can act for someone, across which services, and how do they set that up once?

Why fourth: Reaches the people least likely to hold an account themselves.

5. Several conditions, several services, no shared view

75% of people with long-term conditions had an admin problem, against 57% without. Their care is split across clinicians and pathways.

Where it stands: Yet to be solved. I haven't found national work aimed at the people with most contact.

Where I'd start: Give one neighbourhood team a read only view of everything booked for 50 or so people with several conditions, and see what it turns up. UCLPartners did something close and cut outpatient appointments.

Any plan should answer: Does it cut the number of appointments anyone needs, or only make them easier to see?

Why fifth: Could make the biggest difference, but information governance and data make it slower.

6. Changing, cancelling or booking depends on where you're seen

Some patients describe being unable to cancel or rebook. In Derbyshire, 111 people wanted to book physio, blood tests and smears in the App and couldn't.

Where it stands: Solved in places, that'll do in others, yet to be solved in a lot of settings. GP appointments are booked or cancelled in the App over 400,000 times a month. 115 of the 153 trusts in the App let you change or cancel, but you can only book a first appointment, and some changes sit as pending until someone gets to them. Blood tests are often bookable online, just not in the App.

Where I'd start: Link to booking that already exists before building anything new. And say plainly in the App what it covers.

Any plan should answer: Appointments don't all work the same way: a clinic slot, a screening invitation, a course of six physio sessions, a group class, a home visit. Which of these are in scope, and which quietly drop out of the count?

Why sixth: Partly being done already, so link to what exists first.

7. Children and families waiting with no sight of the wait

A quarter of children on community lists have waited more than a year. Hospital waiting times are in the App. Community ones aren't.

Where it stands: Yet to be solved for community services.

Where I'd start: Show community waits the way hospital waits are shown, starting with children's services. I haven't found anyone doing it yet.

Any plan should answer: Can a family see where they are in the wait, and who to ask when it moves?

Why seventh: Needs community waiting list data in a usable shape first.

8. Providers outside NHS trusts with no way in

Social enterprises, councils, the voluntary sector and independents have no clear route into the App's appointments list. The route is for services representing at least one acute trust, and the community integrations in progress are all for NHS trusts.

Where it stands: That'll do at best. Messaging through NHS Notify is the open door.

Where I'd start: Two or three providers try structured appointment messages through NHS Notify, once there's a local route.

Any plan should answer: What's the cheapest way in for a service that isn't an NHS trust?

Why eighth: Waits on a national route.

9. People the App doesn't reach

Digital exclusion is higher in a lot of rural and coastal areas, and digital only can widen inequalities. Only 40% of people in low income households think technology improves care.

Where it stands: That'll do. NHS England says digital first doesn't mean digital only, and letters carry on for people who need them.

Where I'd start: Treat letters and phone as proper routes in their own right, and put help where exclusion is highest.

Any plan should answer: Which patients and carers were heard first, in which settings, and does it reach rural and coastal areas where exclusion is higher?

Why last: This applies to all the others rather than being a project of its own.

Demand isn't the problem. 76% would use the App to book hospital appointments.

The King's Fund, Healthwatch and National Voices ask for admin as a national priority with clear standards, reporting on patient experience of admin, record access by default, and digital tools that work both ways.

One worked case: late letters

A rough sum, with the assumptions shown. Some of the figures are old or second hand, so any real case would need recent, local data.

• Missed hospital appointments. About 5.6% of 146 million outpatient appointments were missed in 2024/25, so around 8 million.
• How many are down to late letters. Healthwatch reckon around two million of the eight million.
• What a missed appointment costs. About £120, based on 2018 NHS figures. That's eight years old and quoted second hand by a supplier, so treat it as a placeholder. Up to date local costs would be needed before anyone acted on this.
• So nationally. Two million times £120 is around £240 million of clinic time a year.
• Per acute trust. Spread across 134 acute and specialist trusts, that's roughly 15,000 late letter misses and £1.8 million each.
• What the fix costs. A text costs about 2p and a letter at least 64p through GOV.UK Notify, and App messages are free. Messaging 15,000 people by text costs about £350.
• If it only partly works. Say a third of those misses are prevented by sending earlier and App first, with text as the fallback. That's around 5,000 appointments a trust, about £600,000 of clinic time, for a few hundred pounds of messages plus the work to change the process.

Caveats. The two million is a Healthwatch estimate reported in the press. The £120 is a 2018 average, and most of it is capacity rather than cash, since a freed slot only saves money if someone else fills it. And trusts already overbook to cover misses. Even so, the cost of trying it is small next to what it might save.

Not shipping the org chart

Right now the App shows appointments the way the NHS is organised. Each trust's appointments come through its own portal, what you see varies by portal, and you follow a link out to do anything, maybe agreeing to the portal's privacy policy on the way. Richard Pope describes the pattern: services built around the structure of government rather than what people need. He points to Estonia, which designs services around life events that cross agency boundaries.

Other ways to show appointments that don't start from who runs them:

By time

One calendar of everything booked. Denmark's Appointment Overview covers hospitals, municipal health and care, and GPs.

By what you need to do next

Confirm, prepare, fill in a questionnaire, rearrange, whoever's asking.

By course of care

A referral, its appointments, tests and results as one thread. Your knee, even if three organisations are involved.

By who's helping

A carer's view across the people they look after.

With the organisation as a label

Who it's with and who to ring stay on every appointment. They just don't decide how the list is built.

Two caveats:

• The organisation can't vanish. When something goes wrong, people need to know who to ring, and consent and accountability sit with providers.
• A view across providers needs a shared appointment data shape and common booking patterns underneath. It's only as good as what each provider sends.

What NHS England could do, and where it could stay out of the way

Some fixes only work nationally. A lot of the local work doesn't need national help, just room to try things.

Could do

Say what the 95% counts

Publish what it's a percentage of, and report patient experience of admin alongside coverage.

Settle the NHS Notify route for local services

It isn't settled yet. And make fallback costs predictable, since only 41% of App messages are read within a day.

Open a way in for providers that aren't trusts

The appointments route is built around acute trusts. Social enterprises, councils and the voluntary sector need one too.

Take proxy beyond GP practices

The national proxy service is a start. Carers need it for hospital and community care as well.

Agree a shared appointment shape, visits included

So a view across providers is possible, and a visit with a rough window sits alongside a clinic slot. Denmark did this nationally.

Fund over years, not in small late pots

Trust leaders have been asking for this.

Show the working on benefits

The statistics regulator had to ask for the method last time.

Where it could stay out of the way

Let local services pick their one thing

A checklist of App features to switch on won't fit a community provider the same way it fits an acute trust.

Keep reporting light

A few admin standards, reported against. Another dashboard won't help a team that's already stretched.

Don't push new local front ends

Policy is already moving away from separate portals.

Leave room for small local tests

Visit windows, carer views and shared views can be tried by one team before anything goes national.

Worth stopping

Treating "all appointments" as the goal

It counts how many appointments are visible. It says nothing about what it's like for someone juggling six appointments across four services.

Expecting BaRS to get patients seeing their appointments any time soon

Keep backing BaRS as the plumbing between services — but stop counting on it to get appointments in front of patients soon. The parts that would let the App find someone's appointments aren't widely in place yet.

Buying new patient portals for community services

National policy is moving trusts away from separate portals towards direct integration with their records systems.

Dental in the App, for now

The government has no plans for it.

Missed appointments as the main measure

The headline figures are estimates, and for visits and group sessions the measure doesn't mean much anyway.

Safe to assume

Things there's enough evidence for that you can build on them without testing again.

• People will use the App to manage appointments. 76% would book hospital appointments in it.
• Admin goes wrong most for the people with most contact. 75% of people with long-term conditions, against 57% without.
• Seeing appointments hasn't improved how people rate NHS communication. One survey, and about perceptions.
• Reminders cut missed appointments, by about 29% on average for automated calls and texts, and the wording matters.
• Hospital connection is largely done. Community and non-trust care is the gap.
• The national direction is App first messaging and direct integration with records systems. Whether local providers can and will move to NHS Notify is less certain.
• Carers want to see the appointments they manage. 81% in one local survey.
• Home visits can't be booked like clinics.
• Digital exclusion is uneven, and it follows need.
• A national appointment calendar across sectors can be built. Denmark has one, covering hospitals, municipal care including home nursing, and GPs.

Bets worth testing

Some of the starters rest on assumptions that could be wrong. These are the risky ones. If any of them hold, they'd change what's worth building next.

1. Staff get more from a shared appointment view than patients do

• Could be wrong: information governance, patchy data, and staff not looking.
• Worth it if right: fewer clashes, home visits planned around clinics, fewer wasted journeys. Denmark built its overview for staff and relatives as well as patients.
• Test: the neighbourhood team view in problem 5.

2. Coordination that removes appointments beats tools that manage them

• Could be wrong: savings stay local, and clinicians resist joint review.
• Worth it if right: UCLPartners reviewed 922 people with the most complex mix of conditions and outpatient appointments, and cared for many without those appointments.
• Test: give a multidisciplinary review someone's full appointment list across providers and count how many go.

3. Designing for the carer first gets more use than designing for the patient

• Could be wrong: consent, safeguarding and capacity rules slow it down, and some patients won't want it.
• Worth it if right: it reaches housebound, older and very young patients who'll never hold an account.
• Test: the carer conversations in problem 4, then a clickable prototype of their week.

4. A rough arrival window cuts wasted visits

• Could be wrong: caseloads change through the day, and a window turns into a broken promise.
• Worth it if right: nurse time back, and fewer calls asking when she's coming.
• Test: the morning message in problem 3.

5. A clearly labelled partial list builds more trust than one that looks complete

• Could be wrong: the political promise is "all", and people ignore labels.
• Worth it if right: fewer missed appointments from people assuming the App shows everything.
• Test: usability test the same list with and without "this shows hospital appointments, not community ones". Ask people what they think is missing.

6. Two-way contact cuts calls more than booking features do

• Could be wrong: chasing is driven by waits, not channels.
• Worth it if right: the King's Fund's main ask, and switchboard time back.
• Test: sort a week of calls to one trust or community provider by reason. Estimate how many a "where's my referral" message would answer.

7. Messages get most of the value for small providers at a fraction of the cost

• Could be wrong: messages aren't a list, can't be cancelled in place and add noise, and NHS Notify hasn't settled a local route.
• Worth it if right: social enterprises, councils and voluntary services show up in the App this year, not after an EPR integration.
• Test: the NHS Notify trial in problem 8. Measure attendance and calls.

8. Published admin standards change how people feel where visibility didn't

• Could be wrong: standards without enforcement become another dashboard.
• Worth it if right: the admin problem gets an owner.
• Test: one system publishes the admin standards in problem 2, then asks the King's Fund and Ipsos admin questions again locally after six months.

If I had one team and a year

The national target is for 95% of appointments to be available in the App. You could keep working towards that and still have people getting the letter after the date, or waiting in all day for a nurse who might come at four.

Or you could start with how services contact patients about their appointments: letters, texts and App messages. Get contact details and communication needs right, send early enough and by the route that works for each person, and measure whether fewer people miss appointments they found out about too late. Most of it happens behind the scenes, and I think it would help more people than anything else here.