Weighed in the Balances

Paul Taylor

In the summer of 2018 I had a chance encounter with an acquaintance who knew I worked in health tech. He told me that he knew Ali Parsa, the CEO of Babylon Health, the company behind the app GP at Hand. We talked about Parsa’s extraordinary career and it was clear that my friend was impressed not so much by Parsa’s achievements as by his daring. He recalled Parsa reflecting that while Babylon could become a unicorn – a billion-dollar tech startup – it might also end in complete failure. Perhaps both will be true.

The company was valued at $4.2 billion when it was listed on the New York Stock Exchange in October 2021, but its performance since then has been, in Parsa’s own assessment, an ‘unbelievable, unmitigated disaster’. The shares have lost 99 per cent of their value and the Securities and Exchange Commission has been informed that one of the backers will be exercising a right to take the company back into private ownership. The filing notes that ‘the Take Private Proposal will be implemented through the appointment of administrators by the English courts.’

Earlier in 2018 I’d seen Parsa give a keynote address at a digital health conference. He was an impressive presenter, passionate and articulate, ignoring the lectern and striding across the stage, using his phone to project dazzling graphics onto the screen behind, describing how Babylon’s AI was about to revolutionise healthcare. The core of the technology was a huge database of connected facts representing information about diseases, symptoms, risk factors and their associated probabilities. This was used to drive a chatbot that would elicit a patient’s history and advise them whether they should consult a GP. Part of the app was a video conferencing tool so that any patient who decided to see a GP could get an almost instant consultation.

The glitz quickly came off the presentation. Some of the technology that Parsa appeared to be showcasing that summer – including a machine learning algorithm that could interpret facial expressions – never saw the light of day. A key part of the presentation, a claim to have demonstrated that the app was as accurate in its assessments as a qualified GP, was called into question by independent academics, unconvinced by Babylon’s account of its internal evaluation.

The company was already controversial. It had declined to furnish regulators with evidence showing that its algorithm was safe and accurate, arguing instead that it should not be subject to this level of scrutiny, since it was intended to be used only as a source of information not advice, a piece of legalistic hair-splitting undermined by Babylon’s own marketing material. It seemed clear to many doctors that the system wasn’t safe. Tales circulated on Twitter of ludicrous or dangerous misdiagnoses. For a while none of this seemed to matter to Parsa or his backers. Matt Hancock, the then health secretary, was determined to promote the company, much to the irritation of those working in the NHS and dealing with the disruption it created.

Babylon’s original plan had been to charge subscribers a modest fee, around £5 a month. At some point the business model changed: GP at Hand is now available for free, along with 24/7 access to GPs hired by Babylon. Babylon gets income from this because anyone who wants to use the service has to register as a patient of a GP practice in Fulham, through which Babylon receives the associated capitation fee, the mechanism by which GPs are paid for the patients they look after. This arrangement meant the local health authority suddenly found itself with a £20 million shortfall, because it had to fund care for 100,000 extra patients, most of whom lived elsewhere.

When the app launched, GPs were furious that Babylon, which was targeting young, well-off, tech-savvy people, was planning to take money from the NHS while caring only for the least demanding patients. It turns out that this wasn’t quite the case. GP at Hand users ask for, on average, six consultations a year. The company is currently trying to recruit dozens of GPs, practice nurses and prescribing pharmacists to staff its services. Reporting on the company’s travails last November, the Financial Times quoted Parsa: ‘Our team is killing itself providing [this service to the NHS] . . . because the money is so tight and because the connectivity in the system is so broken.’

We still don’t know whether the app is safe. We do know, however, that it doesn’t work, in the sense that it doesn’t do what it was intended to do: enable Babylon to offer primary care services more cheaply than traditional GPs.


  • 26 May 2023 at 1:30pm
    bentoth says:
    If you were designing a healthcare service from scratch, with no encumbrances, no history, no powerfully vested interests, and with Chat GPT 5 on hand, Babylon would work well enough.