$4B is a Lot for Cherry-Picking and Lemon-Dropping

Babylon Health Makes Public Market Debut Despite Key Unanswered Questions

Yet another SPAC announcement brings us the long-anticipated public market emergence of Babylon Health. With the US market’s seemingly endless appetite for virtual healthcare products, Babylon is likely to make their investors a lot of money, at least at first; their long-term prospects are much less promising. In the UK, the Babylon General Practice at Hand (Babylon GPaH or BGPaH) had an almost ideal patient population, one that prioritized ease and convenience while needing very little in the way of long-term care management. With a population that looks anything like the US median, Babylon is going to need to show a lot of clinical and IT chops, and they’ll need to show them soon.

Key Takeaways

• There have been serious concerns about Babylon’s foundational triage and symptom-checker technology. Even assuming those issues are moot, it is hard to see what they offer beyond standard telehealth and remote encounter technology.
• Chronically ill populations require more monitoring, more engagement, and more interaction than Babylon has ever needed to deliver.
• Babylon aims to be a fully capitated VBC entity partnering with insurers, but it remains to be seen if they can really offer meaningful value.

The Babylon IT Proposition

Babylon markets ‘digital-first end-to-end care,’ with the core value coming from their digital triage and patient navigation engine. Appointments can be delivered via telehealth and messaging, with the goal of reducing in-person appointments as much as possible. But beyond telehealth and their AI, Babylon lacks some of the monitoring and wellness functions offered by leading vendors in their space. Their Babylon Monitor app offers baseline self-reporting and can collect data from other devices, but they have no devices partnerships or offerings themselves. They can distribute education and insights, but don’t offer the kind of direct coaching or regular involvement from a care team that most chronic care management programs rely on.

The largest technological issue at play is one that many AI symptom and triage engines are facing: the need to prove that their algorithms are both accurately assessing conditions and offering the correct recommendation. Some vendors have chosen to shortcut these concerns by including providers within the loop of their algorithm. This improves accuracy and patient confidence, but costs time and raw efficiency. Most, including Babylon, rely on the algorithm for the entire diagnostic workflow, with the goal of preventing any unnecessary provider involvement. As Babylon describes it, the technology “shifts the majority of member interactions to the mobile device,” with provider interaction only when needed.

This places an immense burden on the accuracy and recommendations of their technology, and it may not be up to the task. A 2015 study of AI-driven symptom and triage checkers[1] found that accurate advice was only given in 80% of emergent cases, where algorithms often erred on the side of caution. New research into bias and accuracy in AI/ML algorithms has only increased the scrutiny on these tools, and Babylon itself has come under criticism for these issues. In 2019, Dr. David Watkins offered examples of the Babylon app grossly underdiagnosing female patients. A letter published in The Lancet criticized a number of Babylon’s claims and the analytic methods used to establish their accuracy. For a system intended to both act as the primary healthcare interaction for their patients, and to address a large, diverse population, these issues are seriously concerning.

Cherry-Picking and Lemon-Dropping

The question of their patient demographic is the largest Babylon will need to answer. In the UK, an independent audit of the Babylon GPaH product were vastly younger, more affluent, and healthier  than the averages of either equivalent practices or the nation as a whole. 94% of BGPaH patients were under 45 and two-thirds were described  a “city sophisticates” or “career climbers.” Overall, the audit concluded that “the service is not being used by large numbers of older people, or large numbers of people with more complex health needs.”

Unfortunately for Babylon, those are exactly the patients whose healthcare needs to be successfully managed in their VBC revenue model. Babylon has purchased practices in California, partnered with insurers in Iowa, Nevada, and New York, and in January of 2021 partnered with Home State Health to care for Medicaid and Medicare Advantage patients in Missouri. Babylon certainly isn’t cherry-picking their population anymore. All they need to do now is show they can succeed. For a $4B investment, that’s a lot to prove.

---

[1] Semigran, H.L., Linder, J.A., Gidengil, C. & Mehrotra, A. (2015). Evaluation of Symptom Checkers for Self Diagnosis and Triage: Audit Study. BMJ, 351:h3480 doi:10.1136/bmj.h3480