In the original Star Trek series of the late 1960s, Dr. Leonard “Bones” McCoy relied on technology to diagnose and treat every health condition the crew encountered. But the tricorder didn’t replace the doctor. It enhanced his abilities.
At this week’s Digital Medtech Conference in San Francisco, I will take part in a panel discussion on the benefits and risks of AI and technology in health care. We will discuss how medical technology companies could help physicians improve operational efficiency and the customer experience (and boost their productivity) by streamlining processes and administrative tasks. Health care has become more complex as medicine and science have advanced. But technology, particularly artificial intelligence (AI) and cognitive analytics, could help to simplify some areas by bridging specialty silos.
During this session, we will also consider some of the possible risks of technology, such as the potential to displace physicians and other caregivers who have been trained in certain medical specialties. We are already beginning to see the impact that technology can have on jobs in other industries. Car-sharing services, for example, are experimenting with autonomous cars that could replace drivers.
Medical knowledge has historically existed in the hands of the few. But AI and other technologies could make that knowledge accessible to many. In health care, we might reach a point where a tech-savvy general practitioner can complete tasks that can now only be performed by specialists.
AI could help determine treatment options
We recently looked into the role AI and cognitive analytics could play across five areas in health care—wellness, diagnoses, treatment decisions, the treatment itself, and post-acute and chronic care. Of these five areas, it seems that advanced technology has the most potential in helping to make diagnoses and treatment decisions. AI can quickly sift through and interpret mountains of data, something that would be impossible to a doctor, or even a team of doctors, to do. Consider this: In 2015, the National Library of Medicine indexed 806,000 articles.1 A human brain could not absorb even a fraction of that information. But a computer can quickly digest it, categorize it, and determine which data are relevant or useful.
Unlike computers, physicians are able to look at, touch, and talk to a patient, which helps them diagnose a patient’s condition. But technology is getting better at detecting some of the same nuances, and could eventually surpass what we can do as humans. AI, for example, can tap into massive databases to identify slight correlations between different biomarkers that a physician might not notice. Recent technologies have been developed to identify diseases such as cardiovascular disease and dementia by detecting changes in a patient’s voice tone/pitch/pattern. Piecing together multiple bits of data together to develop a composite picture of the patient is an area where technology is excelling.
Moreover, images produced by many smartphone cameras have enough resolution that pictures can be used by dermatologists to accurately assess a condition. Even before contacting a doctor, a consumer can scan a mole with a smartphone. Using an app, that image can be compared against millions of others to determine if a dermatologist should be contacted. While such technology wouldn’t replace the dermatologist, it could replace some unnecessary visits, and allow people to screen themselves if they don’t want to visit a doctor.
Are we entering the decade of medtech?
In recent years, we have seen many biopharma companies develop and launch therapies that cure disease rather than just treat symptoms. If we are now in the decade of biopharma, I think we will soon enter the decade of medtech. AI-enabled medtech could fundamentally change the way health care is delivered. However, payment mechanisms aren’t yet set up to reimburse companies that develop advanced technologies. Some physicians, and other clinicians, might feel threatened and slow down the adoption of some innovations.
What about that tricorder?
We do seem to be approaching the world that Gene Roddenberry envisioned when he created Star Trek. In 2012, XPRIZE and the Qualcomm Foundation launched a global competition that challenged entrepreneurs to develop a working diagnostic device similar to the tricorder Dr. McCoy used aboard his starship.2 The mobile-integrated, non-invasive device would need to diagnose and interpret a defined set of 13 health conditions, while continuously monitoring five vital health metrics. It also had to weigh five pounds or less.
Last April, a Pennsylvania-based start-up won the $2.5 million top prize from. The aptly named Final Frontier Medical Devices—led by an emergency room physician and his brother (a network engineer)—developed DxtER. The AI-enabled prototype met nearly all of the diagnostic benchmarks outlined by the contest, according to an April 13, 2017 press release.
Cognitive technologies have been evolving over decades, and we have reached a point where it is possible to automate tasks that require human intelligence. While much of this technology is already being used in other industries, many medtech companies are just beginning to take a look.
We might still be a few years from AI playing a significant of a role in medicine, and decades away from technology taking the place of physicians. Health care is a difficult environment for technology to thrive—it is a vastly specialized, fragmented, and heavily regulated sector. Maybe the biggest impediment is a reluctance among some physicians to rely on technology, such as AI, rather than their own judgement.
I think it is unlikely that AI will ever fully replace a human physician, but I think we could see a future where AI becomes part of the stethoscope of the 21st Century.