The chief executive and founder of Hummingbirds Medical explains why current recall systems are widely seen as fragmented, admin-led, and clinically unsafe.
Rakeeb Chowdhury is the founder of Hummingbirds Medical, a practising GP and the developer of BookYourGP. A fully automated patient review and recall system, the digital platform is designed to improve patient safety across the UK and has already been recognised by Health Innovation Yorkshire & Humber, Propel Digital Accelerator, the RCGP IMP accelerator, and the NHS Clinical Entrepreneur Programme. Here, he talks to Healthcare Today about patient review and recall systems and how he is addressing patient safety.
What problem are you trying to fix? What is the problem with most fully automated patient review and recall systems?
The simplest way of putting it is that there is a lack of oversight, and many of the systems we use are really unsafe. Too often, we rely on proxies as evidence that care has been provided, rather than checking whether the right care has actually been delivered.
Take blood pressure as an example: we measure it in the arm as a proxy for what is happening in the heart. No one really cares about the blood vessel in your arm – what matters is whether we can prevent heart attacks and strokes. Yet there are plenty of drugs that reduce blood pressure without meaningfully reducing those risks.
That same problem of using the wrong proxy runs right through healthcare. Pay-for-performance schemes like the Quality of Framework (QOF) reward ticking boxes. The dashboard turns green, but a decade later, we discover that outcomes haven’t actually improved.
That’s where our approach is different. We tie everything to human physiology and actual clinical parameters – surfacing whether the work was truly done. It’s embedded in the surgery’s workflow so patients understand why they are contacted, and the team has a clear view of what’s needed.
“What we’re left with now is a kind of Frankenstein system.”
Given the shocking NHS Resolution figures earlier this year and the 10% rise in litigation a year, why has patient safety been neglected?
I think several things have come together here. The core issue is that the systems we use today – electronic health records like EMIS or SystmOne – were built decades ago, in the 1980s and 1990s. That was even before QOF was introduced.
What we’re left with now is a kind of Frankenstein system. In practice, it’s a mishmash of tasks and workarounds: sending tasks, waiting for patients to contact us, seeing alerts when prescribing, maybe getting a flag to do a medication review, running searches in the background, and delegating jobs to different staff. You might have a diabetes champion, an antibiotics lead, a pharmacist doing reviews, and so on. It all seems fine when you walk into a surgery – everyone looks busy and purposeful – but the truth is there is no real clinical oversight.
Doesn’t this require a change in the mindset of staff, rather than just software?
That’s why we describe what we’ve built as clinical oversight infrastructure. The software allows you to embed care pathways directly into the system.
This goes far beyond the traditional idea of recalls. Anyone can create a static patient list and send out bulk text messages. What we’re doing is embedding care: defining the physiological parameters that need to be checked, assigning responsibilities, surfacing what’s missing, and ensuring patients are contacted appropriately. Everything flows into the clinical record, so admin teams aren’t jumping between systems. They can see why a patient has been recalled, who they need to see, and what needs to be done. Nurses know which checks are outstanding. If something is missed, the system shows what wasn’t done, by whom, and why – creating accountability.
How does the system help ensure that people don’t slip through the cracks?
We often show surgeries a demo of how the software works. Traditional systems rely on static filters, which are dangerous because they simply flag that a patient is on a list without showing what’s missing. By contrast, our system tracks everything down to the physiological level. In a patient record, anything overdue is shown clearly in red, with the date it was last done and what is now due.
It also keeps a full audit trail. You can see exactly who has been recalled, when, and by which method. Messages are sent automatically – email, text, WhatsApp or letter – according to the practice’s settings. If a text fails because a number is wrong or the network is down, the system automatically escalates to the next method. If a patient has been contacted multiple times without responding, they appear in a pending tab, so practices only need to follow up manually on a handful rather than hundreds or thousands.
Every action is recorded directly in the patient’s notes: when a recall was sent, why, what is due, and with whom the patient should be booked. That makes mistakes far less likely. Our platform removes guesswork and ensures the process is both transparent and accountable.
Do you have early evidence showing improvements in patient outcomes or safety from practices using the system?
We started in 2020 with our first site in a very deprived surgery where all the partners had left, and there were no GPs. They’ve now sent around 40,000 recalls using our system. Since then, we’ve grown organically and are now working with practices across York, Newcastle, Leeds, Manchester, London, Norwich and elsewhere.
In London, Strong Green Surgery recently started with us. It has already sent 1,000 recalls with only one query – which, when checked, turned out not to be an error but a clinical marker being misread. That shows how the system works silently in the background, surfacing the right needs without confusion.
“The technology to provide genuine oversight hasn’t existed until now.”
It always comes down to money. How are you funded?
We’re self-funded and haven’t raised any external money, which makes us a bit unusual. The reason we started goes back to a case I saw as a trainee. A patient with back pain missed his blood tests, went through physio and other care, and eventually returned years later with prostate cancer that had spread. By then, three of the GP partners were facing legal action. I was astonished – in an age of computers, how could something so critical have been missed? It became clear there was simply no system of oversight.
That experience convinced me that something needed to change. When I first floated the idea, people didn’t believe it would work, and we never fitted the mould for grants or funding. So we just built it ourselves. Within a year or two, it was self-sustaining.
Are you helped by the government’s focus on digital transformation?
We often end up working outside the usual remit of technology vendors. Behind the scenes, we think of ourselves a bit like Stripe in payments. Stripe doesn’t just move money; it detects fraud and keeps the system safe. In the same way, we can’t just automate healthcare recalls and walk away.
The problem is the NHS often sees things differently. It commissions new tools – triage platforms, patches, connectors – and once they’re deployed, it sees it as a job done. That makes the system look fine on paper, but in reality, it isn’t. We’d never accept that level of risk management in another safety-critical industry.
That’s why this issue has been overlooked. The technology to provide genuine oversight hasn’t existed until now. But once you see it, it’s obvious: you need a system that goes beyond static lists, that can show what’s needed, when, why, and with whom – and that can be audited. That’s the only way to make this truly safe.
What are the biggest barriers to adoption in primary care?
The big barrier here is business. In software, you usually have two models: low-cost, mass-market subscriptions like Netflix, where millions of customers cover the costs; or enterprise contracts like Palantir, where one deal is worth hundreds of thousands. General practice falls into a difficult no man’s land. Sales are complex, but GP surgeries have no tech budget, no r&d budget, and very little resource compared to other countries.
If you approach an integrated care board (ICB) as a small company, they’ll ask for evidence. But to generate evidence, you need adoption first – it’s a Catch-22.
The second problem is misalignment in care pathways. The theory is that if innovation reduces pathway costs, it should be adopted. But in practice, it doesn’t work. If you develop an app that helps women detect breast cancer recurrence earlier, the GP won’t pay for it – not their budget. The breast team won’t pay either – they’re funded for surgery, not prevention. So innovations fall through the cracks.
The NHS talks about prevention, but the infrastructure doesn’t allow it.
The bigger challenge is higher up. Many decision-makers are non-clinicians. They don’t understand what’s happening under the hood – the lack of oversight, the safety risks, the missed opportunities for prevention. We’re left with a system where surgeries can’t afford innovation, structures above them can’t enforce it, and everything gets stuck in the middle.
