Do you feel like the future is coming at us much quicker than our brains can comprehend? If you answered yes, you are not alone. Technology is advancing at a pace faster than we can truly understand.
Just 50 years ago, NASA used multiple computers the size of a car to get a man on the moon. Today, a single iPhone is millions of times more powerful than all of NASA’s combined computing power in 1969.
Today, on the blog, we are joined by Dr. Joel Selanikio who is a physician, futurist, TED Speaker, innovator, emergency responder, and technologist. Dr. Selanikio answers questions about AI and what it means for healthcare.
Our Conversation with Dr. Selanikio
AMP: For many, the idea of AI still seems hard to grasp. With that, can you share an example or two of AI being used in healthcare today?
Dr. Selanikio: AI is barely used in clinical practice today. One of the only examples, which is very new, is the IDx-DR system: an autonomous system that diagnoses diabetic retinopathy from retinal scans, and is the first autonomous AI system approved by the FDA. Other examples include diagnostic systems/symptom checkers like Babylon or Ada, or Zebra Medical’s radiology AI, which is just moving beyond the pilot stage.
Outside of clinical care, there are more examples: ZocDoc, for example, uses AI to check your insurance when scheduling your appointment.
AMP: Do you have a vision of how AI and or Machine Learning can tackle the frustration many providers deal with daily as it relates to EHR systems?
Dr. Selanikio: I don’t think that the problem with EHRs is primarily a technology issue, it’s a business model issue. The problem is that the EHR providers don’t sell, for the most part, to clinicians using their software. The user is not the customer. This misalignment of incentives means that, predictably, EHRs don’t get clinically better as a priority. Getting our healthcare data outside of the EHR, for example, using the Apple Health system, will enable better-incented companies to provide us with better utility for that data — both for patients and clinicians.
AMP: Big Data by itself is useless. For that data to be useful, it requires the data to be analyzed, understood and acted upon. As it relates to healthcare in the United States, where is there data sitting and not being used? What can come from acting on this data?
Up to 73 Percent of Company Data Goes Unused for Analytics.
Forrester Reserach
Dr. Selanikio: Well, in healthcare that would be most clinical data: the process for “doing something with it” is so lengthy and cumbersome that most data in EHRs is never used. As to what could be done with it, the examples are many: how about providing every clinician with their performance figures compared to their peer averages (e.g. for antibiotic prescriptions, tests ordered, etc). That requires no AI, just computing, but it is rarely done. How about providing sensible data visualizations in the EHR and the patient portal that would enable clinicians and patients to better understand labs, etc?
AMP: Right now, a family medicine doctor doesn’t receive new vitals or stats on their patients until they come back into the office. How can AI, wearables, and data be leveraged to catch potential health risks before they arise?
Dr. Selanikio: A good example is the Apple Watch’s ability to detect atrial fibrillation. But this is really just the beginning. Wearables are a good example of the process of consumerization of technology, whereby we have experienced a 50-year shift in who information technology is produced for and sold to. In the 60s and before, the customer was the government, the universities, and large corporations. After the “personal computer” revolution, followed by the introduction of the mobile phone (the real personal computer revolution), it’s plain that the overwhelming majority of “computers” on earth are made for individual consumers.
Why does this matter? Because computers produce data, and the consumerization of tech means that the vast majority of data — and health data specifically — is produced by consumer computers — including, now, wearables. For most people, their phone records more information relevant to their health than anything in your medical record. As an example, my phone knows how much and how quickly I move every day: strongly correlated with my general health. My smart speaker knows what I say, allowing it to discover when I’m depressed or stressed. My wearables can tell my heart rate, and likely soon my blood pressure and blood glucose, enabling them to detect a variety of health conditions much earlier than possible before.
Without the recent flourishing of artificial intelligence, which has come about because of the huge increase of data and the plummeting price of computing power (both of which we owe to consumerization), all that data would still require us to engage with the healthcare system: we would need a doctor’s expertise to make sense of it all. But with AI, we can now envision a near-future in which ever-increasing bits of a doctor’s expertise is cheaply available to us via our phones, or our smart speakers, or our wearables.
AMP: What security issues are at play with AI or does AI actually improve cybersecurity?
Dr. Selanikio: Well, all the same security issues at play with everything else connected to the internet are applicable here. Right now hospital security is, if anything, even worse than the EHRs, so clinical data security needs to be beefed up. But the vast majority of health-related data is already non-clinical, non-hospital data for most people . . . so the question really just relates to how we keep all of our information secure. We’re still in the early days of even deciding what data is ours, who owns different data points, etc.
Dr. Selanikio’s Bio
Dr. Joel Selanikio is a physician, TED speaker, inventor, emergency responder, and consultant working in the fields of technology, healthcare, artificial intelligence, entrepreneurship, social innovation, big data, child health, and disaster response. He is the winner of both the Wall Street Journal Technology Innovation Award for Healthcare and the $100,000 Lemelson-MIT Award for Sustainable Innovation for the development of the Magpi mobile data collection system, the first cloud-based application created for global health and international development.
Dr. Selanikio has consulted and/or spoken at Davos, Foo Camp, WHO, UNICEF, IFRC, Harvard, MIT, Stanford, Google, DARPA, CNN, Fox News, the Clinton Global Initiative, the Royal Society of Medicine, and for the pharmaceutical and healthcare industry – and has been profiled by the Guardian, Wired, Forbes, TED, The Economist, The Wall Street Journal, Fox News, the BBC, NPR, Information Week, and the Washington Post, among many other publications.
An emergency responder and former CDC epidemiologist and outbreak investigator, in December 2014 – January 2015, he was the lead physician at the IMC Ebola Treatment Center at Lunsar, Sierra Leone. As an officer of the Public Health Service, Dr. Selanikio served as Chief of Operations for the HHS Secretary’s Emergency Command Center after the 9/11 attacks. In 2005, he was given the Haverford Award for Humanitarian Service for his work in treating tsunami victims in Aceh.
