Google’s Search for Direction in Healthcare

Fake news, begone: Reports of Alphabet’s healthcare struggles have not been greatly exaggerated.

That bears repeating: Not been greatly exaggerated.

Indeed, Google’s parent company is foundering a bit in its attempt to claim a significant stake in the industry, but its failures should not be mistaken for waning influence or interest in medicine. Truth be told, Alphabet’s healthcare roots are more extensive and entrenched than most people realize.

“Google has always seen itself as more than a search and advertising company,” concludes a 47-page report from New York, N.Y.-based research firm CB Insights. “It’s…betting that its AI prowess can create a powerful new paradigm for the detection, diagnosis, and treatment of disease.”

That bet is already beginning to pay off. For every stumble like Google Glass, Google Health, Google Helpouts, and the glucose-sensing Smart Lens (shelved last fall over scientific and biological challenges), there are triumphs such as Google Fit, Onduo, G Suite, and Verb Surgical.

Scads more potential coups lurk in Alphabet’s pipeline, thanks to its bevy of healthcare-related acquisitions, venture capital investments, partnerships/collaborations, and patent filings. The company’s venture arm, GV (the former Google Ventures), has backed nearly 60 life sciences organizations over the last decade, including consumer-oriented brands like 23andMe, as well as data management entities (DNAnexus), and disease screening specialists (Denali Therapeutics, Freenome). The latter company likely won Alphabet funding by appealing to the larger firm’s commitment to improve early disease detection (particularly cancer). Freenome is developing technology to detect cancer in its early stages by pinpointing trace amounts of DNA that tumors shed into the bloodstream (circulating tumor DNA).

Similarly, the 2017 purchase of Seattle startup Senosis Health could help Alphabet improve anemia and chronic obstructive pulmonary disease detection. Apps developed by Senosis turned smartphones into simple monitoring devices that could collect health metrics to diagnose pulmonary function, hemoglobin levels, and other critical health information.

One of Alphabet’s most important healthcare-related acquisitions, however, arguably was its 2014 purchase of British artificial intelligence firm DeepMind, a startup whose AlphaGo algorithm beat the world champion of the abstract strategy board game Go in 2016. Last fall, Alphabet absorbed DeepMind’s health division, which is developing tools to scrub EHR data for better, faster health services. DeepMind’s AI prowess is likely to play a key role in helping Alphabet employ artificial intelligence techniques to solve fundamental problems in science and human biology.

“Over the coming years, we expect AI to help scientists make transformative advances on problems ranging from protein folding to image analysis, potentially improving medical diagnosis, drug discovery, and much more,” DeepMind’s founders wrote in a November 2018 blog post. “We’re excited to play our part in that journey both at DeepMind and at Google, in the service of patients and clinicians around the world.”

Although the world of machine learning and AI is still in its infancy, Alphabet is already leveraging the technology to help pathologists spot cancer cells in tumor biopsies, and interpret retinal scans to faster diagnose diabetic retinopathy. Moreover, DeepMind has created an algorithm that can predict the complex, three-dimensional shapes into which proteins can be folded—a key step in understanding biological processes and finding the molecules necessary for developing new medicines.

DeepMind also is working with Moorfields Eye Hospital in London to improve eye care, developing an AI system that can match leading experts in diagnosing various optical conditions, and recommend the correct referral decision for more than 50 eye diseases with 94 percent accuracy. In addition, DeepMind has partnered with both the National Health Service and Cancer Research UK Centre at Imperial College (London) to improve upon tumor identification research. Alphabet’s previous research in this area yielded algorithms that could detect metastasized breast cancer tumors with a 92 percent accuracy rate.

To coordinate its plethora of digital health initiatives, Google last fall established a new healthcare unit that includes a team that previously operated under DeepMind. The company tapped healthcare veteran David Feinberg, CEO of Geisinger Health System, to head the new unit.

Google Health will primarily be focused on creating digital products and services that can be sold to doctors and hospitals—an approach that could help differentiate the search giant from other medtech industry outsiders looking to move into the space (Amazon, Apple). And while the diversity of healthcare initiatives in Google’s crosshairs are promising for the future of medicine, the disparity could ultimately undermine the company’s ability to deliver on many of its innovations, CB Insights warns.

“Google is working on so many initiatives focused on so many different facets of healthcare across so many areas of the company that the chances of failure are high,” the CB Insights report states. “But so is the potential for success. Ultimately, if Google can find effective solutions for any one of the many issues it’s tackling, there’s a potential to apply lessons and successful approaches elsewhere, and create a new data- and AI-driven healthcare paradigm.”