This is the third installment of a five-part series titled “Diagnosis Overkill.”
Dr. Jay Parkinson of Sherpaa recently weighed in on the medical chatbot and the slow but steady automation of primary care. He notes the value proposition in deploying bot technology to do things like collect baseline risk assessments, deploy standard patient questions, and so on. This software is a far cry from legacy EMR of yore: A chatbot layer ostensibly offers system-wide, quickly scalable user interfaces (UI) that can talk to clinicians, solicit data from software systems and individuals, and serve as a much more dynamic form of clinical decision support for clinicians.
Today, clinical decision support (CDS) modules serve as a primitive level of artificial intelligence. Companies are already working on souping up CDS with better User Interfaces (UI). This may prove to pave the way for full diagnostic automation tomorrow – though we’re not quite there yet in medical practice today. “I’m a little hesitant to have a bot identify what’s going on and go down a rabbit hole,” notes Dr. Parkinson. “Augmenting the human brain with AI is the way to go – working in partnership.”
As might be expected, chatbot vendors are more bullish on the role that smart UIs can play in disrupting medicine. Ali Parsa, founder of a UK-based chatbots startup called Babylon, has been busy drumming up expectations to pave the way for his company’s success: “An average human doctor does about 7,000 consultations a year. I don’t think [Babylon] is going to be as good as a doctor. I think it is going to be 10 times more precise than a doctor. No human brain is ever going to be capable of doing anything of the sort.”
The clinical market opportunity for a completely closed-loop medical AI will take time, regulation, and piloting to mature – but the consumer apps are already off to the races. For patients, these tools make it easier than ever to self-diagnose, order tests, schedule appointments, and more. Babylon is one of the leaders so far, having signed on clients ranging from large corporates like Cisco to the UK’s National Health Service (NHS). Baidu, HealthTap, WebMD, Buoy, are on a rapidly growing list of vendors are building chatbots to serve as the next incarnation of Dr. Google.
This empowerment represents a huge step forward for the consumer movement in healthcare, which is a good thing for access and patient-centeredness. Yet, it conveniently ignores the central question raise by the self-diagnosis trend: Are we conflating what patients want with what patients need?
When you start feeling unwell, how far do you usually go as a patient before you see a doctor? I…
— Christina Farr (@chrissyfarr) March 23, 2017
Of course, bots aren’t the only instrument in this orchestra. Companies like Lemonaid and Nurx use virtual “doctor review” to approve patient requests for various prescriptions. A recent study suggested that the convenience of telehealth inflates utilization of services. The rise of concierge medicine and direct primary care offers both doctors and patients a hassle-free way to get what they want. CVS’ Minute Clinics are incorporating point-of-care testing to make routine blood draws and some laboratory testing available to consumers under supervision of physician assistants and nurse practitioners. Ample evidence suggests that pharmaceutical companies’ consumer advertising drives up rates of prescriptions, testing, and referrals.
These disruptions all highlight how traditional medicine is falling behind the relentless challenges posed by emerging business and technology trends. We don’t know just yet how consumers will interact with a digital healthcare system that has none of the boundaries (geographical, temporal, administrative), that we’ve grown so used to bumping into.
We do know that healthcare systems will have to adapt significantly to handle a new wave of patients who can self-diagnose on their phone, get tested at home or in a pharmacy, self-refer through their health plan’s provider network, and initiate a consultation about the surgery, medication, or treatment they want. But to paraphrase Bill Gibson, we’re already here, just not all of us. The cultural factors underlying these disruptions – we want what we want, and it’s good business to get it to us when we want it- are part of why we’ve gotten to the point where six million people get hurt due to mistreatment every year.