Automating Medicine? We’re behind the curve if we are still discussing whether Vinod Khosla is right or wrong.
20 percent doctor included - Discussing if Mr. Khosla is Right or Wrong is the wrong discussion to have - An outsider’s perspective
This week I had the pleasure of attending an evening Q&A session at the Health 2.0 Conference. Interestingly titled - Cowboys vs Aliens? Doctors vs Algorithms? The what and the when in the future of data, doctors and patient care - the hour long session was my first encounter, up close and personal, listening to a dialogue on this particular issue.
I could be wrong, but I sensed disagreement in the air. There was an awkward gotcha moment, which a gentlemen behind me found particularly amusing yet the laughter was a bit derisive to my ears. But that was but a brief moment.
For my bit, after the session, I went outside, introduced myself and told Mr. Khosla (yeah I’m old school like that), that in essence, he was right and that I agreed with him. As did the recovering medico I was with. The original post can be found here.
Discussing if Mr. Khosla is right or wrong is actually the wrong discussion to have. The discussion we should be having is the following.
Given that Mr. Khosla (and others) are right about the future of digitized and automated medicine, what institutions, processes and educational practices should we create now so that we, as practitioners, patients, and caregivers can usher in the era of automated and digitized medicine.
Why Mr. Khosla is Right - Understanding the proper context
Its very simple. There are many of us who are aiming to do precisely this. We recognize that medicine is messy and humans, prone to error, deal with messy information to make the best decisions they can under the circumstances. We want physicians to make the best decisions they can and the best decisions are data enabled decisions. They always have been. Data is information and increasingly it is being digitized.
My personal approach is to consider the following:
Healthcare is really divided into Medical Skills and Surgical Skills or any combination of the two, along with the Data and R&D done to introduce new processes and procedures. One has to understand a system to be able to solve a problem within the system and humans have five senses plus a mental apparatus that reacts to those senses.
We touch, taste, see, hear and smell; This is how we interact with the world and this is how health is also delivered, within a system that we continue to learn about (biological), and within a system that we have created, to deliver it through the best means possible (finance, administration and regulation).
For me, our ability to automate medicine revolves around our ability to digitize the senses with which we make clinical decisions and with which we practice our craft.
The Question of Experience
Gladwell in Blink eloquently describes intuition. The more experience one has the better one’s intuitive decision making. There however is a twist to this story; something that I don’t think has played out as yet. Information around us is changing so fast, that professionals I believe, won’t have enough time (without digitized information) to develop the deep intuition that physicians before us developed.
And as our reliance on digital information increases, our experience without digital information will decrease. In the digital age, our ability to make better decisions without the aid of digital information will result in cognitive overload unless we simplify, automate and create decision support systems. This is simply a fact of the way our brains are wired. We grew up in the Serengeti but now live on planet Digita. And we aren’t going back either.
This is playing itself out in medicine already. If you didn’t grow up playing video games (2D hand eye coordination) minimally invasive surgery will be tougher to master. Indian physicians are slowly losing their clinical acumen as they rely more on tests (that have become cheaper and hence more ubiquitous) joining their American brethren, generally recognized as having poorer clinical skills. Correlation of high to low fidelity radiological information is already guiding clinical decision making in low resource settings. NASA is already trying to understand how to perform remote robotic surgeries; how does one overcome lag - pesky physics!
The future of automated medicine is already being created; the issue is how we choose to institutionalize its practice.
That is the discussion we should be having. Where, and who should be included in the creation of the Institute of Automated Medicine? Given the fidelity of information being generated and our switch to a preference for Bayesian statistics, what, if any, priori information can be used for establishing regulatory claims? And given the convergence of business models around preventative health, as our bucket of preventable diseases expands due to our discovery of novel determinants, who is a patient and what is the scope of activities that define health care?
There is no doubt in my mind that medical skills can be largely automated. The question of empathy so often raised in this conference is a psychological question and we already rely on inanimate objects to bolster our psychological requirements. Can machines empathize with humans ? I think very soon machines will do a better job at recognizing both non-verbal and verbal forms of communications and Doctors already seem to be terrible at empathy (I disagree, but I keep hearing this over and over again over the last 15 years) so the question around automation is not if a machine could empathize, because it clearly could.
Its more around, allowing ourselves to become a society that has forced physicians into abandoning the basic sense that led them into being physicians or give them the best tools with which they can go back to being the healers they always wanted to be.
Image Attribution “Rust Bot 1” @plognark - Deviant Art
Bias
I am not a practitioner. I am what I call a reluctant physician. Reluctant because the system of medicine as practiced was something that I simply couldn’t internalize as a career because of its inefficiencies. I very soon decided to get into the management of healthcare, thinking I would have a better opportunity to change the system from the outside than on the inside. Everything about Medicine seemed alien to me. The way it was taught, the way it encouraged unnecessary competition. The way patients were herded around. There seemed to be something wrong and while my colleagues recognized the same issues I recognized, they were too busy practicing medicine.
I also have not practiced clinical medicine in the US. I know how it works. I however have been a patient in the US. I find the US healthcare system particularly onerous from a patient perspective. I find shopping for health care insurance annoying and complicated. And yes while I have established a PCP, someone, once tried to take my pulse by palpating my ulnar artery and when I pointed it out to her (I couldn’t tell who it was… everyone wears pajamas) I was informed that one could palpate on either side of the wrist. No you can’t :) My resting pulse rate is 64 not 88 thank you very much.
I come from a family of physicians. My sister, brother in law, father, uncle, and aunts are all physicians. My great grandfather was a surgeon in the Royal Indian Army. Most of my very good friends and colleagues are also in the medical field either as Practitioners or Administrators or Innovators. Some of them are in the US, some in India, and some in Australia. My perspective to healthcare in general and to healthcare in the US, in particular, is as an outsider looking in. I’m dangerously armed with a very good understanding of medicine as it relates to basic diagnosis, treatment and population health. I understand technology. I have several invention disclosures, most of which are related to automation and digitization of certain aspects of the practice of medicine. Finally, I am also the Founder of a seed stage company looking to automate nutrition management for both personal and professional use.
That is a lot of bias and you should understand where I’m coming from.