A practice manager in Yorkshire put it plainly — "We're always firefighting... I'm constantly worried about money... It takes a mental toll, and I know plenty of colleagues who are thinking about leaving the profession."
That quote appeared in a Practice Index report on the £550 million staffing gap GP practices now face. It's one voice, but it speaks for thousands. The IGPM (Institute of General Practice Management — the membership body for practice managers) estimates that practices across England need to find £550 million to cover salary increases in 2025/26. On top of that, the 2024 Autumn Budget increased employer NICs (National Insurance Contributions — the tax employers pay on staff wages), adding an average of £47,000 per practice per year, with some practices facing up to £100,000 in additional costs.
Public sector employers were compensated for the NIC increase. GP practices, as independent contractors, were not.
This is the world practice managers are operating in. Not "hidden inefficiency." Survival.
66p per patient per day
GP practices receive approximately £164 per patient per year from the NHS. That works out to 66p per patient per day the practice is open. Staff costs consume around 60% of that, according to NHS England data.
GP partners aren't NHS employees. They're independent business owners who bear personal financial liability for their practices. If a partner wants to invest in new systems, that money competes directly with staff wages. There are no deficit budgets to fall back on. There are no departmental capital allocations. If the investment doesn't pay for itself within months, it's a personal financial risk.
The King's Fund and Nuffield Trust (both independent health policy think tanks) have both noted that the GP contract is "a poor mechanism for allocating capital investment." There's no ring-fencing for capital spending in what practices receive, meaning any investment in efficiency competes with every other pressure on the budget.
This is why most practices aren't investing in better systems. It isn't ignorance. It's rational caution in a funding model that punishes experimentation.
But the cost of doing nothing is growing faster than the cost of practice automation.
Where 25-40% of GP time goes without healthcare automation
The RCGP (Royal College of General Practitioners — the professional body for GPs) published a landmark study in December 2025 titled "Uncovering the GP Workload Burden." For the first time, it formally defined and quantified what practice staff have felt for years: a third of GP time goes to unnecessary workload and bureaucracy.
That figure aligns with earlier BMA data. Their GP Worklife Survey found that 40% of a GP's working day goes to non-clinical activities: test results, correspondence, referrals, admin, and meetings. A GPonline survey found that 51% of GPs spend between 25% and 49% of their time on paperwork. One in four GPs spends half their time on it.
The BMA recommends a 3-to-1 clinical-to-admin ratio. That means 75% clinical, 25% admin. Most practices are nowhere near this. Clinical admin workload is running 34% higher than pre-pandemic levels, according to GPonline analysis.
The pattern is strikingly similar in dental practices, where admin tasks consume just as much clinical time. Translate that into money. A locum GP (a temporary doctor hired to cover sessions) costs £600 to £800 per day, or £80 to £110 per hour (GPonline, NASGP (National Association of Sessional GPs), 2025 data). If 30% of a salaried GP's day goes to admin that could be handled differently, that's 2.4 hours. At locum rates, that's £192 to £264 per GP, per day, spent on work that isn't seeing patients.
For a 3-GP practice over a year, that's roughly £150,000 to £206,000 in clinical time absorbed by admin.
The 200,000 appointment problem
A Pulse Today survey of practices in October 2025 estimated losses of around 200,000 appointments per week nationally due to online access changes. Since October 2025, patients can submit online requests throughout core hours (8am to 6:30pm). The result: GPs are triaging a constant stream of online requests instead of seeing patients face-to-face. One GP described it as a "free-for-all buffet" of urgent requests buried among less critical queries.
The specific tasks healthcare automation can tackle
Generic talk about "admin burden" isn't useful. Here are the tasks that actually consume the time, based on published data.
QOF coding and data entry
The QOF (Quality and Outcomes Framework — the NHS performance-based payment system for GP practices) directly ties practice income to accurate clinical coding. Miss a smoking status entry, misclassify a condition, or fail to update codes to current QOF indicators, and you lose points. Points are money.
Outsourced coding teams process up to 200 letters per hour (General Practice Solutions). That volume tells you how much coding work exists in a typical practice. One practice with an 8-week backlog cleared it in under 4 days after switching to an outsourced coding service, saving 28% in admin costs while improving CQC (Care Quality Commission — the body that inspects and rates healthcare providers) compliance.
The income at stake is significant. QOF is performance-based pay, and coding accuracy determines how much of it you actually collect.
Referral letters
Writing referral letters is slow. Each one requires pulling together patient history, investigations, and a clinical narrative. Systems rarely auto-populate the relevant data.
A North Central London pilot programme tested automated referral protocols that pull in QOF codes, investigations, and problems automatically. The result — 1 to 5 minutes saved per referral, with 78% of users reporting time savings in that range. Users rated the system 4.67 out of 5, and 89% found it easier than their previous process.
Scale that across hundreds of referrals per month and the hours add up.
Clinical correspondence and blood test filing
Every practice knows the never-ending inflow — clinical letters, hospital discharge summaries, lab reports, referral updates. Each one needs reading, coding, and attaching to the correct patient record.
At Elm Tree Medical Centre (18,000 patients), staff process roughly 200 blood test results per day. The manual filing process takes over 85 seconds per result. That's 4.7 hours of pure filing time, every day, for one practice.
A British Journal of General Practice study found that clinical paperwork alone (blood tests, prescriptions, referral letters) consumes 12.8% of GP time. That isn't admin staff time. It's GP time spent on paperwork instead of patients.
Patient registration and repeat prescriptions
New patient registration takes 5+ minutes per patient when done manually. Complex or missing information requires follow-up calls or letters. Repeat prescriptions generate constant volume that needs manual review, processing, and routing to the right clinician for sign-off.
These are high-volume, low-complexity tasks. The kind that practice automation handles well, without requiring a qualified person to do them step by step.
What about patient communication?
Appointment reminders, test result notifications, registration confirmations. Most practices still handle these through phone calls or letters. Automated SMS, WhatsApp, and email reminders reduce no-shows (which typically run around 5-8% across UK practices) and free up reception staff from making dozens of calls per day. The technology exists. Most practices aren't using it. We explored the same problem in veterinary settings, where no-shows cost UK vet practices up to £71,000 a year.
What this costs a real practice
Take a typical 3-GP practice in England — 3 GPs, 1 practice nurse, 1 practice manager, 3 reception/admin staff, 1 medical secretary. Around 5,000 registered patients.
Here's what that costs in practice, using BMA, RCGP, and NHS England data.
| Cost area | Annual cost | Source / basis |
|---|---|---|
| GP time on avoidable admin (3 GPs, 2.4 hrs/day each at £95/hr midpoint) | £178,000 | BMA/RCGP 25-40% admin time, GPonline locum rates |
| Document processing (200 results/day at 85+ sec each) | £18,200 | Elm Tree Medical Centre case study, admin staff rate |
| QOF income leakage from coding gaps | £5,000+ | Estimated from QOF points missed due to incomplete coding |
| NIC increase (not compensated by NHS) | £47,000 | IGPM / LMC average per-practice figure |
| Staff turnover (1-2 admin staff/year, burnout-related) | £9,000 | CIPD (Chartered Institute of Personnel and Development) replacement cost estimate £6,000-£12,000/role |
| Total annual admin cost | £257,000+ | Majority is GP clinical time consumed by non-clinical work |
The largest number in that table is GP time. That's deliberate. Practice managers think in GP hours first, and they should. Every hour a GP spends on paperwork is an hour they aren't seeing patients, aren't generating appointments, and aren't doing the clinical work the practice exists to provide.
Not all of that £257,000 can be recovered. Some admin is unavoidable. But the RCGP themselves describe a third of it as "unnecessary." Even reclaiming a quarter of the avoidable time would free up £44,000+ worth of GP capacity per year, in a single 3-GP practice.
What that £44,000 actually buys you
That's nearly a full-time practice administrator. Or 55 extra locum days. Or 7 months of a part-time clinical pharmacist. In a world where practices receive 66p per patient per day, recovering that capacity isn't a nice-to-have. It's the difference between running a viable practice and considering contract handback (returning the NHS contract to the local commissioner, effectively closing the practice). One in six practices is already weighing that decision, according to LMC (Local Medical Committee — the local body that represents GPs) surveys.
This isn't optional anymore
The 2025/26 GP Contract and PCN DES (Primary Care Network Directed Enhanced Service — the contract that groups of practices sign up to for extra funding and responsibilities) include specific digital requirements tied to funding.
By 31 March 2026, Primary Care Networks (groups of neighbouring GP practices that work together, typically covering 30,000–50,000 patients) must certify the following.
- Digital telephony data is being used for capacity and demand planning
- A consistent approach to triage exists across online, face-to-face, and telephone requests
- Structured information is collected for walk-in and telephone requests, not just online
This is linked to the CAIP (Capacity and Access Improvement Payment — extra NHS funding practices receive for meeting digital and access targets), worth roughly £45,850 for a typical 50,000-patient PCN. Practices that don't meet these requirements don't just miss out on better systems. They lose income.
The NHS is spending £10 billion on technology and digital transformation by 2028/29. But most of that goes to trust-level infrastructure (hospitals and large NHS organisations) — electronic patient records, shared care records, national platforms. Practice-level automation isn't centrally funded. If you want to automate your own workflows, you pay for it yourself.
The question isn't whether to act. The contract requires it. The question is whether you do it reactively, at the last minute, or whether you build systems now that actually save time.
What practice automation actually looks like
Automation in a GP practice doesn't mean buying a software product and hoping it fits. It means looking at the specific tasks consuming your team's time and building systems that handle them. For a broader look at what this means across healthcare settings, our guide on practice automation covers the concept from the ground up.
In practice, this means things like the following.
- Patient registration. Online forms submitted, data validated, records created in EMIS or SystmOne (the two main clinical record systems used by GP practices) without manual re-entry. Patients notified via SMS, WhatsApp, or email when registration is complete
- Appointment reminders. Automated SMS, WhatsApp, and email reminders sent based on appointment type and lead time, reducing DNAs (Did Not Attend — missed appointments) without reception staff making calls
- Document processing. Clinical letters and lab results sorted, coded, and attached to the correct patient record. Staff review exceptions instead of processing every document
- Referral tracking. Every referral logged and tracked. Automated follow-ups when expected responses are overdue. Nothing sits in a queue unnoticed
- Repeat prescriptions. Requests received, checked against criteria, routed to the right clinician. Patient notified when ready for collection
- QOF coding support. Automatic flagging of coding gaps, missing indicators, and data entry errors that affect QOF income
- Online triage routing. Incoming online requests categorised and routed by type and urgency, so GPs aren't triaging a single unfiltered queue
These aren't theoretical. Practices using automated referral systems report 1-5 minutes saved per referral. Practices using outsourced coding cleared weeks of backlog in days. The North Central London pilot programme showed 89% of GPs found automated referrals easier than their previous process.
You don't need a developer to start
Modern automation platforms connect to existing clinical systems and NHS services. You don't need custom software built from scratch. Many of these automations can be set up in days, not months, and adjusted as your needs change. The barrier to entry is lower than most practice managers think. What you do need is someone who understands both the technology and how GP practices actually work, so the systems fit your real workflow rather than an idealised version of it.
The ROI (return on investment) timeline that matters
Practices facing £47,000 to £100,000 in additional NIC costs can't wait two years for a return. The investment framework for GP practices is different from corporate IT projects.
| Question | What matters |
|---|---|
| How quickly does it pay back? | 3-6 months maximum. Anything longer is too risky given current funding pressures |
| Does it protect existing income? | QOF accuracy and contract compliance are more certain than revenue growth |
| Does it meet a contract requirement? | If the PCN DES mandates it, the only question is how you comply, not whether to |
| What's the alternative? | The choice is often invest in automation or make redundancies. 68% of practices in one regional survey expected to cut staff after the NIC increase |
| Can we prove it works in NHS context? | US case studies are irrelevant. UK practice-level evidence is what partners need to see |
Most practices that invest in practice automation for core admin processes see a return within 3-6 months. Not because automation is cheap, but because manual admin, measured in GP hours and locum-equivalent costs, is very expensive.
Where to start
You don't need to automate everything. Start with the task that wastes the most clinical time.
Ask your reception team to log, for one week, every task that involves entering the same information into more than one system, chasing something that should have happened automatically, or correcting an error from manual entry. The volume will be higher than you expect. If your existing software handles records but not the gaps between systems, that is a common pattern we cover in why off-the-shelf software is not enough.
Then pick the single biggest time sink. Build practice automation for that one process. Measure the hours saved. Use those savings to fund the next one.
This isn't a technology project. It's a capacity project. Practice automation gives your GPs and admin staff back the hours they're spending on work that adds no clinical value, so your practice can do more with the funding it already receives.
With over 6,000 FTE (full-time equivalent) GP partners lost since 2015 (a 25% decline, per NHS England data), and 76% of GPs saying patient safety is compromised by excessive workloads (RCGP "GP Voice" survey), the capacity problem isn't going away. The practices that will still be operating in five years are the ones that find ways to do more with less, starting now.