From disease management to insurance, Google is betting big on using AI to reshape $3 trillion healthcare industry.
Tech giant Alphabet, Google’s parent company, is leaving no stone unturned to shape the future of healthcare industry. A recently released report from research firm CB Insights throws interesting insights into Google’s strategy in leveraging Artificial Intelligence to explore various opportunities in healthcare from disease detection to interoperability to insurance.
The report states that among the five tech giants (Amazon, Facebook, Microsoft, Google and Apple), Google is focusing far more on AI and machine learning as a differentiating factor in ingesting, structuring and processing health data than others. It is investing more on research papers, opening more AI centres around the world and acquiring more AI companies than the rest.
As Google restructured into Alphabet in 2015, it formed three subsidiaries to focus on health AI:
- Verily, which focuses on using data analytics, tools, interventions and research to improve healthcare,
- DeepMind, which is dedicated to finding newer ways to apply AI in healthcare, and
- Calico, which focuses on applying AI on large datasets to learn and combat ageing and age-related diseases.
Google also leverages its other assets such as Google Cloud to support these operations and it further invests heavily in AI-related healthcare start-ups through its venture arm Google Ventures.
The focus of Google’s AI strategy spans the entire healthcare spectrum, including but not limited to –
- Data Generation. This includes acquiring, digitizing and ingesting health data produced by MRIs, wearables and various other imaging devices.
- Disease Detection. This involves designing AI algorithms to identify anomalies in the given datasets that can be used as markers in identifying the presence of disease.
- Disease Management. This involves developing data and research-driven tools to help people diagnosed with or are at a risk of developing a disease to manage their day-to-day life.
The report analysed the publicly available information to identify specific diseases Google is focussing on and these include – eye diseases, diabetes, heart diseases, Parkinson’s and multiple sclerosis.
Eye Diseases. Verily is partnering with Nikon’s subsidiary Optos to detect diabetic retinopathy by analysing retinal images using AI. Google earlier published research proving that its algorithms are as good as ophthalmogists in detecting the condition. Another eye disease where Google is quite active is presbyopia and several Verily patents mention using contact lenses and wearable devices to correct the vision.
Diabetes. Verily partnered with Dexcom to create a small Continuous Glucose Monitor (CGM) that should soon see commercialization and another project with Alcon to monitor glucose levels through tears. The long term objective seems to design a wearable and disposable bandage to detect glucose levels without requiring any finger calibrations.
On the management front, Google is including its Dexcom monitors in Onduo, a virtual diabetes management program it launched a while ago. More interesting ideas Verily is pursuing include a smart syringe that lets diabetics track dosages properly and an automated insulin dispensing system using AI to optimize insulin delivery automatically using monitors and infusion pumps.
Heart Disease. The CB Insights report states that Google is currently working on two ways to detect anomalies in heart conditions. One using both an electrocardiogram and heart rate monitor and the other using optical sensors and machine vision to passively monitor heart by analysing the blood vessels in the eyes.
Parkinson’s Disease. Verily is partnered with Netherland’s Radboud University to identify the underlying causes of Parkinson’s by combining and analysing clinical and patient data including heart functionality, electrodermal activity and inertial movements. By steadying heart rhythm and sleeping patterns, Verily hopes to identify markers that indicate disease onset. Google also acquired Lift Labs, a health AI start-up which created a smart spoon that helps Parkinson’s patients to stabilize their hand movements while eating food.
Multiple Sclerosis. Multiple Sclerosis is a potentially destabilizing disease affecting brain and central nervous system. Its causes are poorly understood and it currently has no cure. Verily is partnering with Biogen and Brigham Hospital for a longitudinal study to understand the onset of the disease. Here too, Verily is developing machine learning algorithms to patient data to possibly detect and understand the underlying causes.
Other disease areas Google is focusing on include Chronic Obstructive Pulmonary Disease (COPD), cancer, behavioral health, aging and life extension.
Google’s second big bet in AI applications in health is developing a new data infrastructure to integrate medical data across various EMRs, apps, medical devices and other health data tracking products. Google acquired Apigee, a health data management company, in 2016 for $625M. It aims to use FHIR technology (Faster Health Interoperability Resources) to create programming interfaces to combine datasets from different systems.
Google’s other subsidiary DeepMind is working on creating AI and analytics applications in healthcare and for that it needs a data infrastructure to combine the siloed medical information from EMRs, devices and doctors.
DeepMind recently released an app to identify acute kidney injuries by pushing patient information and alerts to caregivers to help them identify the severity before it escalates. It also plans to use its data infrastructure to develop more such apps and also allow third party developers to build on its platform.
Apart from accessing data streams of existing health systems, Google is also developing its own data streams that others can use for their research.
Google is developing AI-enabled tools to augment doctors’ expertise. It partnered with Johnson & Johnson to create products involving machine learning, robotic surgery, instrumentation, advanced visualization, and health data analytics.
With a job posting on Verily website for a health plan executive, the CB Insights report suspects that tech giant is also exploring health insurance. The underlying principle seems to be if Google can better detect and manage diseases using various AI and machine learning techniques, then it can also become an insurance company to efficiently manage people’s health risks.
Also noteworthy are the reports that Verily is bidding for Medicaid contracts. And on the other hand, Google Ventures has invested in other health insurance companies such as Oscar, Clover and Collective Health that target individuals, small businesses, Medicaid and self-insured populations.
But Can Google Really Reshape the Healthcare Industry?
The answer is doubtful. Because Google’s past ventures into healthcare have not been successful. Its past projects Google Health and Google Flu Trends failed and were shut down a few years later.
Verily’s ambitious ‘Tricorder’ project to use nanoparticles and magnets to track different proteins and biomarkers in the body in real-time proved to be too ambitious and was shut down for undisclosed reasons.
Also, Google seems to be working in too many directions across too many different arms in a “spray-and-pray” approach instead of concentrating on few key projects. But the report concludes that while this may increase the chances of failure, it also increases the chances of success.