Can AI solve what’s ailing healthcare?

I ask academics and investors driving the healthtech space

Thrive invites voices from many spheres to share their perspectives on our Community platform. Community stories are not commissioned by our editorial team, and opinions expressed by Community contributors do not reflect the opinions of Thrive or its employees. More information on our Community guidelines is available here.
On stage at Healthtech 2019
On stage at Healthtech 2019

Healthcare is a personal experience that’s intrinsically part of our shared human experience. Even healthy people at some point see a doctor of some sort, get vaccinated for school or travel, be prescribed medicines, undergo tests, maybe even get stitches or surgery. Small wonder, healthcare is the fastest growing sector in our economy. Access to care, and the quest for preventative care and to cure the currently uncurable conditions that affect 40% of Americans are among the biggest problems facing medicine and the society it serves. 

Technology in healthcare as in other user sectors has always been a R&D enabler helping researchers to work faster and collaborate better. Powered by the scale of cloud and built by brilliant human minds to mine and learn from massive data sets, AI applied to diagnostics and therapeutics let alone administrative functions can handle repetitive tasks, patient log transcriptions to advance healthcare innovation. That in mind, I was eager to hear healthtech academics and investors share their outlook on the collision of new technologies and healthcare at Constellar Ventures’ annual conference, where panels were equal parts focused on bringing to light the science and the technologies, particularly AI, shaping the space. 

The current cost model creates cascading problems that are ripe for disruption. As highlighted by Dr. James Wall, Stanford pediatric surgeon and director of the university’s BioDesign Fellowship, “Healthcare is expensive.” US healthcare system has the distinction of spending twice as much per patient as other developed nations with no better outcomes to show. Individually, “practitioners have a very limited scope of experience” compared to “the exponential growth in medical knowledge,” and at the end of the day, the fee for service system doesn’t incentivize delivering best outcomes. Patients pay a copay and then extra for lab work, a prescription, or a procedure regardless of the treatment’s effectiveness. “The convergence of data, connectivity, computing power [can] provide more efficient personalized healthcare. But it has to be coupled with value, to deliver better outcomes, at less cost.” 

Researcher and practitioner entrepreneurs are disrupting every piece of the healthcare system with a sense of urgency. “There’s a ton of pain points in healthcare,” said Jeff Herbst, VP of Business Development, Nvidia and Head of GPU Ventures. “If we don’t do something, the healthcare system is probably going to collapse within my lifetime.” Another systematic problem that healthtech is poised to correct is the reliance on treating the patient after the illness takes over. Fred Toney, Managing Founder at Launchpad Digital Health pointed out, “It’s a sick care system. It’s not a health care system.” As Frank Caufield, Managing Director at Darwin Ventures, elucidated in a prior session, healthtech solutions already underway range from “the mundane” like delivering pharmaceuticals to “the sublime” like using CRISPR to edit DNA that would answer the call for preventative care by rooting out diseases and disorders before they can afflict someone’s quality of life. 

But, innovation comes incrementally. Caufield reflected on his decades of indirect investments across technology and healthcare. “We’re always on the cusp of solving for cancer, for cystic fibrosis, for diabetes, for any number of health ailments. While the trend line is going up, right, it’s not a really steep trend line. I’ve seen so many attempts at therapies not quite pan out but make incremental progress.” While technology has not yet supercharged science to eradicating many serious conditions, survival rates for once deadly diseases and caring for chronic conditions are better. For instance, for diabetes, the devices for testing blood sugar, glucose levels, and insulin delivery have improved the experience for managing the condition but not eradicated the condition itself. And as Herbst noted on a different panel, not all forms of assistance have to be moonshots. “Natural language processing” to accelerate speech recognition “alone is going to be a huge deal” by taking the burden of documentation off doctors’ plates.

Accelerating the cure. Whether solving a specific condition or a systemic feature, there are no walled gardens. A drug breakthrough has to go through FDA approvals and then plugs into insurance and healthcare networks. Flipping the sick care model into preventative care requires medical intervention, personal accountability, and potentially subsidies from employers, insurance, and or the government to sustain healthy lifestyles for preventing avoidable conditions. “The problem is identifying the people that need it the most for intervention. Then, it’s really difficult to get people to change habits, but economic incentives help and those are starting to kick in,” said Toney. 

The need for interdisciplinary thinking extends even as we zoom into specific pieces. For instance, in the realm of hardware, teams need to account for the technical considerations as well as integrating the technologies they are building into the healthcare system. “When you have an approach to human prosthetics, it’s not just mechanical. It incorporates the complexity of the software, as well as the neurology and the interaction with the interface,” explained Calvin Chin, Managing Director of E14 Fund. Citing augmented and virtual reality and wearables for healthcare, he said, “New forms of human computer interface will require lots of invention.” 

Machines are augmenting, not replacing human doctors. Deep learning machines can do the data scouring and cross-referencing but they will need human guidance. Alireza Masrour, General Partner at Plug & Play Ventures, provided some grounding context. “What is happening right now is AI and other technology tools are getting closer and closer and the world is getting smaller and smaller.” Humans can use AI to augment our work in drug discovery and diagnostics, but this is amplification of processes rather than replacement of trained medical professionals. If given a choice between a human practitioner and a robot doctor with no human oversight, how many patients would choose the latter? And even if autonomous vehicles and internet-of-things in daily doses gradually sway the public to trust machines to run our healthcare experience, would insurance pay for treatments prescribed and surgeries performed by unsupervised robot doctor? 

At the moment, the answer at the social and policy levels is no, and even the technologists advancing innovations are focused on refining AI to be good assistants rather than replacement options. “AI in the future is going to help the doctors be much faster and be a bridge for them to communicate [best practices] with each other,” said Masrour, who sees AI supporting “a huge portion of diagnostics” enabling human doctors to work on treatment. This resonated with Chin. “Everything from implantable to prosthetics [creates] novel approaches to human augmentation.”  

So, can AI, the shiniest tool in the healthtech tool belt, fix the systemic and medical problems in our healthcare system? Time will tell. So far, consistent funding backs up the measured optimism that healthcare specialists, academics, and investors have in healthtech’s ability to fix what’s broken in the care system. As new technologies like AI mature, they enable more innovative applications. But AI reaching its full healthcare potential will also depend on outside variables. A system that can crunch large data sets will sit idle if it doesn’t receive enough data and will return results of limited use if samples are too homogeneous in gender or ethnicity, as has been the case with cardiovascular research

What’s certain is the immense need to make processes more efficient, to open up access lanes to care for the unserved and underserved, and to fix what ails us, whether through preventative diagnosis or post-detection therapeutics. Technology has already offered solutions for these areas in the forms of simplified insurance, AI diagnostics apps and CRISPR. To build on that work and turn potential into reality, we will need as much collaboration as we can get amongst the people who know the science, technology, and funding to bring new solutions to market. Events like Constellar’s help foster crucial cross-disciplinary dialogue.

You might also like...


The Need For Physician Leaders For A Better Healthcare System

by Dr. Tomi Mitchell
First Responders First//

Working Together to Transform Healthcare Workplaces

by Lynda Benton

The Future of Healthcare: “We must move away from today’s volume-based model to one that better recognizes and accounts for value” with Patty Fritz, Vice President of U.S. Corporate Affairs for UCB, Inc

by Christina D. Warner, MBA
We use cookies on our site to give you the best experience possible. By continuing to browse the site, you agree to this use. For more information on how we use cookies, see our Privacy Policy.