Can the NHS crisis be solved by partnering with tech start-ups? The cautionary tale of Babylon Health

The pressures facing the NHS are constantly in the UK news. If the NHS is in crisis, as headlines frequently proclaim, one suggestion is that partnering with tech start-ups that use digital tools like AI could be the answer.

In a new paper, Dr Alison Powell, Associate Professor in the Department of Media and Communications at LSE, questions whether the values of tech start-ups are compatible with those of public service institutions like the NHS.

Speaking at an LSE Research Showcase talk, Dr Powell tells this story through the cautionary tale of Babylon Health. The company contracted with the NHS in 2015 but went bankrupt in 2023. Yet barely more than five years prior, the company was being described as ‘‘revolutionary’’ by Matt Hancock, the UK’s then Secretary of State for Health, while Babylon Health’s founder, Ali Parsa, was claiming that it was to ‘‘do for healthcare what Google did with information’’.

Babylon Health promised digital health services for many parts of the NHS, all ‘‘powered by AI’’. These included what journalists called a ‘‘robot doctor’’ and a digital health service advertised on billboards that caught Dr Powell’s eye. ‘‘They read ‘Get well now’ and ‘See an NHS doctor in minutes – for free’.’’

These ads were for a service called GP at Hand, which not only benefitted from Hancock’s enthusiastic support, but also gained vast amounts of media coverage for its claims that its online diagnosis chatbot had scored better than human GPs on a major medical test. But how did speedy online appointments, chatbots and a rapidly scaling start-up impact the NHS? As Dr Powell discovered, features of Babylon Health’s ‘‘blitzscaling’’ growth strategy created a mismatch with NHS public service commitments.

The bargain that underpins the funding of GP services is that...the higher cost of serving higher needs patients will be balanced out by the contributions made by lower needs patients.

The promise of a GP on demand

Babylon Heath’s offering of a GP on demand could seem a welcome solution to the much-publicised difficulty of securing a GP appointment in the UK. But there were some key differences between a regular GP service and GP at Hand. GP at Hand’s employees could be based anywhere, provided they were allowed to practise in the UK, they didn’t have access to patient medical records, and typically they might see a patient only once. As Dr Powell observes, ‘‘that doesn’t quite line up with some of the patient-centred aims of the NHS that focus on continuity of care.’’

Indeed, in October 2018, this lack of continuity of care led the Advertising Standards Authority (ASA) to ban the very ads that had caught Dr Powell’s attention. The ASA deemed the marketing misleading – rather than see a GP in minutes, patients had to change their GP surgery and potentially wait several weeks to be re-registered with GP at Hand.

In fact, as Dr Powell explains, GP at Hand was itself a GP practice registered in the West London borough of Hammersmith and Fulham. Despite offering only online appointments, it received the same NHS funding as a typical GP surgery – around £120 per patient at the time. The disproportionately large number of patients registered with GP at Hand – over 50,000 by 2019 – is argued to have affected the budget of the borough, contributing to a shortfall of 22 million pounds.

Alongside this, an Ipsos MORI-backed assessment concluded that GP at Hand primarily served younger, healthier patients. Dr Powell explains the knock-on effect: ‘‘The bargain that underpins the funding of GP services is that, based on a flat rate per patient, the higher cost of serving higher needs patients will be balanced out by the contributions made by lower needs patients.’’ That GP at Hand’s patients skewed younger and healthier meant other GP surgeries had to take on the increased costs of treating those with higher needs, who became less likely to receive timely care.

Chatbots as online doctors

GP at Hand promised more than just speedy access to online appointments with GPs. It also developed a chatbot that was presented as being able to do some diagnostic work, with media coverage explicitly referring to it as an ‘‘online doctor’’.

In response, human doctors began to question the chatbot’s diagnostic abilities. Dr Powell refers to a social media video with over 10,000 views on Twitter (now X). The video was posted in February 2019 by user @DrMurphy11 (consultant oncologist David Watkins), who had used the chatbot to simulate an interaction with a 66-year-old woman, entering a painful breast lump as the only symptom. After asking if the patient was pregnant or breastfeeding, the chatbot concluded that one of the most likely diagnoses was osteoporosis – even though painful breast lumps are a major symptom of breast cancer among this age group.

In an interview on BBC’s Newsnight in February 2020, Dr Watkins correctly pointed out that the symptom checker had never been regulated by the Medicines and Healthcare products Regulatory Agency (MHRA) as a diagnostic tool.

By late 2020, Babylon Health announced that its products were being reclassified by the MHRA, which was also investigating its overall governance. But this didn’t stop the company pitching for larger and larger NHS contracts – including one to develop a digital-first integrated care model to serve Wolverhampton, a city with 300,000 residents.

In 2022, the contract with Royal Wolverhampton NHS Trust was terminated eight years early. Only 5,000 people had registered with the service. “On 5 September 2023, Babylon Health went bankrupt, GP at Hand was sold, and the rest of the services wound down”, Dr Powell notes.

Blitzscaling might be good for Groupon, and it might not be good for the NHS.

Blitzscaling and start-up culture

What went so wrong with the partnership between Babylon Health and the NHS? Dr Powell reflects that ‘‘the way that I read the story is that the aims of Babylon Health were misaligned with the aims of the NHS.’’

For Dr Powell, this mismatch particularly arose through Babylon Health’s adherence to a software-inspired start-up strategy known as ‘‘blitzscaling’’. Made famous by LinkedIn co-founder Reid Hoffman, blitzscaling is defined by three core elements: growth; speed; and efficiency.

Growth is secured through increasing measurable indicators, such as investment, revenue or employees. Speed means moving companies into new markets and developing new products quickly. Efficiency prioritises ensuring that software can be re-used in different settings and sold to different markets. These three elements will ideally generate growing venture capital interest and investment through repeated rounds of fundraising, with the ultimate aim that the company will be purchased or absorbed by another company within several years.

This ‘‘mismatch between making a reliable set of medical technologies and making a good story for the next funding round’’ encapsulates the fundamental gulf between the values of start-ups like Babylon Health and institutions like the NHS. As Dr Powell observes, ‘‘blitzscaling might be good for Groupon, and it might not be good for the NHS.’’

Babylon Health capitalised on the NHS’s legitimacy and manipulated media visibility to stay present in public discussions long enough to move through venture capital funding rounds.

Lessons for public service innovation

So, what lessons should we take from the cautionary tale of Babylon Health? The first, outlines Dr Powell, is how the media and the status of the NHS are powerful tools for start-ups to use when promoting their products and seeking funding: ‘‘Babylon Health capitalised on the NHS’s legitimacy and manipulated media visibility to stay present in public discussions long enough to move through venture capital funding rounds.’’ Dr Powell references Dr David Kampmann’s PhD research, undertaken in the Department of Sociology at LSE, which explored how the company particularly leveraged the promise of AI as part of its coverage.

Yet, these automated systems carry the risk of epistemic injustice. Unlike an experienced clinician, AI tools might overlook crucial data and ignore age, gender and ethnic background considerations that could profoundly affect a patient’s diagnosis.

As a result, Dr Powell argues, innovation needs to be aligned with existing expertise and knowledge: ‘‘Expert knowledge is essential, and so is the tacit and contextual knowledge gathered from working in particular places, like with specific patients in a specific setting.’’

Another lesson concerns the kind of innovation frameworks that fit with the stated values of public services. Dr Powell elaborates on the consequences of an overfocus on growth, speed and efficiency: ‘‘Babylon’s growth model had negative outcomes for clinical labour, it helped to create a precarious gig work model for clinicians performing GP appointments on demand. This has had a knock-on effect for patients of undermining continuity of care.’’

When it comes to understanding the potential mismatch between start-up culture and public service institutions, we need to remember that ‘‘moving fast and breaking things might seem to justify regulatory evasion for the benefit of gaining investment and growing a business, but not when lives are on the line.’’

This LSE Research Showcase talk was written up by Rosemary Deller, Research Engagement Manager.

LSE Research Showcase is a series of 20-minute talks from LSE researchers to enjoy on your coffee break. Catch up on YouTube.

Subscribe to LSE Research for the World

Interested in more research like this? Sign up to receive our newsletter: a bi-monthly digest of the latest social science research articles, podcasts and videos from LSE.

Subscribe

The London School of Economics and Political Science (LSE) is a world-leading university, specialising in social sciences and ranked top in the UK by The Times and Sunday Times Good University Guide 2026. Based in the heart of London, we are a global community of people and ideas that transform the world.