We combine cutting edge AI and broader technologies with human expertise to re-engineer a better model of healthcare
👋 Please click ▶️ to expand
🧭 Mission and Product
2500 years ago, in the ancient city of Babylon, folks needing medical advice would gather in the town square and share thoughts on treatments for common illnesses. This is one of the earliest examples of democratizing healthcare - It’s no surprise that Babylonians enjoyed the longest life expectancy across the ancient world
Our mission is to take this simple ideology, apply radically modern technology, and put an accessible and affordable health service in the hands of every person on Earth🌎 🙌.
To make healthcare accessible, we deliver it through the devices people already own.
To make healthcare affordable, we use technology to help ease the burden on our doctors.
By automating their routine tasks, we allow doctors to focus on what they do best: give care to the patients who need it most.
We also use our tech to step in early and help people become more informed about diseases — so they can make informed decisions and stay healthier for longer.
We operate in the healthtech sector and the mission of the company is to provide affordable and accessible healthcare services to everyone globally.
We do this through three core product offerings
Video appointment service
This allows patients to have a video appointment remotely with a doctor, at home, at work, or on the move. I think the benefits of this have been recognized even more during the COVID pandemic.
This allows patients to present their symptoms to our symptom checker; this is essentially an AI chatbot. This chatbot utilizes artificial intelligence to give the patient a probable diagnosis based on the symptoms they enter.
These first two products are focused around sick-care and when a patient is actually ill.
Focuses heavily around predictive and preventative healthcare.
This allows patients to build a personal health profile for themselves. They can answer a series of health related questions, they can also input data about themselves covering things such as diet, lifestyle choices and past family medical history alongside syncing it with devices to track things such as daily steps and heart rate.
All of that data is then used and utilized using AI and machine learning to predict future illnesses or ailments that the patient is potentially more at risk from in the future, and it then works to give the patient a list of preventative measures that they can take, or lifestyle changes that they can make to reduce those risks in the future.
Joseph Enguehard, Dan Busbridge, Adam Bozson, Claire Woodcock, Nils Hammerla
To address their temporal nature, we treat EHRs as samples generated by a Temporal Point Process (TPP), enabling us to model what happened in an event with when it happened in a principled way. Our proposed attention-based Neural TPP performs favourably compared to existing models, and provides insight into how it models the EHR, an important step towards a component of clinical decision support systems.
Published in 2020 | ML for Health Workshop, NeurIPS 2020
Adam Baker, Yura Perov, Katherine Middleton, Janie Baxter, Daniel Mullarkey, Davinder Sangar, Mobasher Butt, Arnold DoRosario
We performed a validation study of the accuracy and safety of the babylon AI triage system and human doctors using a set of identical clinical cases. Overall, we found that the AI system is able to provide patients with triage and diagnostic information with a level of clinical accuracy and safety comparable to that of human doctors.
Published in 2020 | Frontiers in Artificial Intelligence Medicine and Public Health
Yuanzhao Zhang, Robert Walecki, Joanne Winter, Felix Bragman, Sara Lourenco, Chris Hart, Adam Baker, Yura Perov and Saurabh Johri
AI-driven digital health tools often rely on estimates of disease incidence or prevalence, but obtaining these estimates is costly and time-consuming. We demonstrate that context-aware machine learning models can be used for estimating disease incidence. These methods are quicker to implement than traditional epidemiological approaches. We therefore suggest it complements existing modelling efforts, where data is required more rapidly or at larger scale. This may particularly benefit AI-driven digital health products where the data will undergo further processing and a validated approximation of the disease incidence is adequate.
Published in 2020 | Frontiers in Artificial Intelligence Medicine and Public Health
Claudia Schulz, Josh Levy-Kramer, Camille Van Assel, Miklos Kepes and Nils Hammerla
A promising application of AI to healthcare is the retrieval of information from electronic health records (EHRs), e.g. to aid clinicians in finding relevant information for a consultation or to recruit suitable patients for a study. This requires search capabilities for beyond simple string matching, including the retrieval of medical concepts (diagnoses, symptoms, meditations, etc) related to the one in question. We open-source a novel medical concept relatedness benchmark, which is six times larger than existing datasets and consists of concept pairs that co-occurr in EHRs, ensuring their relevance for medical information retrieval from EHRs.
Jonathan G. Richens, Ciarán M. Lee & Saurabh Johri
Machine learning promises to revolutionize clinical decision making and diagnosis. In medical diagnosis a doctor aims to explain a patient’s symptoms by determining the diseases causing them. However, existing machine learning approaches to diagnosis are purely associative, identifying diseases that are strongly correlated with a patients symptoms. In this paper we show that this inability to disentangle correlation from causation can result in sub-optimal or dangerous diagnoses. To overcome this, we reformulate diagnosis as a counterfactual inference task and derive counterfactual diagnostic algorithms. We compare our counterfactual algorithms to the standard associative algorithm and 44 doctors using a test set of clinical vignettes. While the associative algorithm achieves an accuracy placing in the top 48% of doctors in our cohort, our counterfactual algorithm places in the top 25% of doctors, achieving expert clinical accuracy. Our results show that causal reasoning is a vital missing ingredient for applying machine learning to medical diagnosis.
Vitalii Zhelezniak, Aleksandar Savkov, April Shen, Nils Hammerla
Some of the top approaches to semantic textual similarity rely on various correlations between word embeddings, including the famous cosine similarity. We show that mutual information between dense word embeddings, despite being difficult to estimate, is another excellent candidate for semantic similarity and rivals existing state-of-the-art unsupervised methods.
Stoilos, Giorgos and Juric, Damir and Wartak, Szymon and Schulz, Claudia and Khodadadi, Mohammad
The success of logic-based methods for comparing entities heavily depends on the axioms that have been described for them in the Knowledge Base (KB). Due to the incompleteness of even large and well engineered KBs, such methods suffer from low recall when applied in real-world use cases. To address this, we designed a reasoning framework that combines logic-based subsumption with statistical methods for on-the-fly knowledge extraction.
Published in 2020 | European Semantic Web Conference
We create various large-scale datasets for testing whether embeddings correctly encode the similarity between medical terms and test existing state-of-the-art embeddings on these datasets. Our results reveal that existing embeddings cannot adequately represent medical terminology. Our new datasets are thus challenging new benchmarks for testing the adequacy of new medical embeddings in the future.
Knowing that a disease is highly correlated with symptoms, or a drug highly correlated with recovery, is not enough, and basing medical decisions on such information can be dangerous. To truly begin to revolutionise healthcare, AI must learn to distinguish cause and effect. Our work solves this by utilising new physics-inspired ideas about what it means for one variable to cause another, and showing how causal relationships in one dataset limit the possibilities in other overlapping datasets. To illustrate our algorithm, we apply it to breast cancer data, showing how to extract causal relations between two important features despite the fact that they were never measured in the same dataset.
Vitalii Zhelezniak, Aleksandar Savkov, April Shen, Francesco Moramarco, Jack Flann, Nils Y. Hammerla
We push the limits of word embeddings on semantic textual similarity tasks by introducing DynaMax, a novel unsupervised non-parametric similarity measure based on word vectors and fuzzy bag-of-words. This method is efficient and easy to implement, yet outperforms current baselines on STS tasks by a large margin.
Douglas et al. Published in 2017 | NIPS Workshop, NIPS 2017
🗿 Big Rocks
Uniquely positioned to demystify and unify two critical trends in Healthcare
Value Based Care
✅ Movement away from broken fee-for-service model
✅ Aligns system around proactive care
❌ Not scalable traditionally
❌ Challenge addressing many patient types
✅ Leverages technology-driven efficiencies
✅ Bringing care to the patient vs patient to care
❌ Not scaled to address holistic care
❌ Shifts site of care vs addressing overall care and cost
🌎 Scaling Globally
We cover more than 20 million people globally and provide over 5,000 consultations per day.
In 2013 we were founded in the UK where we partner both with private healthcare providers and also our National Health Service. Currently we have around two and a half million registered users in the UK.
Outside of the UK we have also begun expanding in a number of other regions over the last couple of years. This includes South East Asia where we’ve partnered with Prudential who are a large healthcare provider in that region. There we have partnered with them to roll out our Healthcheck product, which is branded as the “Pulse” by Prudential service to 13 countries across South East Asia including Vietnam, Malaysia and Laos. Currently around 12 million Prudential users have access to this service.
We also launched in the US in 2020, where we’re currently operating in four states have around 6 million users currently have access to our service. This is going to be a huge growth area for us in the future and we’re currently scaling our engineering function both in Austin, Texas and the UK.
But some of the most interesting and meaningful work that we’ve done has been in Africa. This is something that many people at the company and particularly the founder is very passionate about, because we really want to be able to offer our services not only to countries that already have a good healthcare infrastructure in place, but also to countries that are perhaps lacking slightly more in that respect.
The work that we’ve done in Africa so far has predominantly been in Rwanda. Here we’ve partnered with the Rwandan government to provide and offer our services to around 2 million people currently.
It’s been particularly beneficial to remote regions of Rwanda that might usually struggle to access healthcare services. We’re able to do this because we tailor our service in Rwanda, so rather than video appointments, we mainly provide text and call based appointments and services. The main reason for this is the lack of infrastructure and that only 15% of the adult population in Rwanda have access to smartphones. Therefore by tailoring our service it allows us to reach and help many more users in that region.
This has been a great blueprint of what we’d like to achieve in other developing nations and is something we’re looking to expand on in the future.
✈ Flight Path ~ IPO
Landing gear RAISED
Babylon has entered into a definitive merger agreement with special purpose acquisition company Alkuri Global. US big data company Palantir has also taken a strategic stake in Babylon.
Following the merger the combined company will operate as Babylon and will operate on Nasdaq under BBLN. This will give us a pro forma equity value of $4.2bn (£2.9bn).
The merger is expected to be completed in the second half of 2021. Several external investors, including Palantir, have helped fund a $230m (£162m) private placement – a sale of stock shares to investors and private institutions. It is worth $10 (£7) a share.
Thoughts from our CEO
“We founded Babylon on a fundamental belief, that it is possible to make quality healthcare accessible and affordable for every person on earth by combining the latest in technology and the best in medical expertise.
“We have achieved one of the highest growth rates every year since our inception, with consistently high clinical outcomes and patient satisfaction. Becoming a public company is just another step in our journey.
“We are at the very beginning of our work to re-imagine our sector, to make it digital-first and prevention-first and shift the focus away from sick care to true health care.”
Series C - Babylon Health
Series B - Babylon Health
Series A - Babylon Health
Seed Round - Babylon Health
There are no rows in this table
“We see this as really the future of what health care can be especially in a health-care market that’s broken and embarrassingly so,” Williams said in an interview. “Our view is that it’s a space that’s ripe for disruption. For too long we’ve tolerated this idea that quality and affordability can’t be balanced in health care. With Babylon we see opportunity to partner with a team that we think can deliver that balance.”
~Groupon’s ex-CEO Rich Williams and former COO Steve Krenzer @ Alkuri Global
Sec Form F-4
Load content from sec.report?
Loading external content may reveal information to 3rd parties. Learn more