Tuesday, March 31, 2015

I, Dr. Robot

Admit it: when J&J announced that its Ethicon division was teaming up with Google on surgical robotics, didn't you imagine that Google wanted to do for surgery what they are trying to do with (self-driving) cars?

I know I did.

For better or worse -- more on that later -- their ambitions are more constrained, at least for now.  It appears that Google's initial contributions will focus on using its expertise in improving the video feeds surgeon use, in and developing algorithms that add value to the images.  Wired, for example, says Google intends to "do things like highlight blood vessels and display critical information on screen."

They also cite a J&J spokesperson adding that perhaps Google's algorithms might also highlight the best places for incisions, based on the individual patient's medical history: "sort of a Google Maps for surgery," in Wired's words.

The Wall Street Journal similarly says that: "Google reckons it can use its machine-vision and image-analysis software to help surgeons see better as they operate or make it easier for them to get information that’s relevant to the surgery."  Google also told them that, "software could highlight blood vessels, nerves or the edges of tumors that are difficult to see with the naked eye."

Well, that's more like a rear view camera than a self-parking car.  Google could just give its newly patented "smart" contact lens to surgeons if that's all they are bringing to J&J.

Kevin Fogarty of Computerworld, for one, is underwhelmed by the announcement, saying that "Google's contribution will be things it is already good at, not the kind of really new capabilities robot-assisted surgical systems actually need."  As the title of his piece says, they're bringing vision to robots that need help with touch.

Touch is a surprisingly hard problem to solve.  There is work being done on it -- haptic feedback, as it is sometimes referred to -- to create what is sometimes called the "Tactile Internet."  The possibilities such breakthroughs might open up are mind-blowing.

Robotic surgery itself is nothing new.  It has been around about 30 years now, and is most often used for laparoscopic surgeries, especially for things like prostate removals.  The long time leader is Intuitive Surgical, Inc, with their da Vinci surgery product, which can cost $2 million per robot and perhaps $3,500 more per procedure than non-robotic laparoscopic surgeries.

With those kind of high capital and per-procedure costs, one would like to think there would be big improvements in outcomes and/or safety, but that does not seem to be the case.  A 2013 study in JAMA found little added benefit to robotic surgery, despite clearly higher costs.  Moreover, a Kaiser Health News special report that same year raised questions about whether anyone -- patients, providers, or regulators -- are getting the right information.

Some critics worry that the manufacturers spend more time creating consumer demand than training surgeons to use their equipment.  As one ENT told Healthline, "Intuitive are probably the worst at doing this and probably the most responsible. Their direct-to-consumer marketing is just criminal. Their lack of training, in my opinion it borders on criminal."

Yikes!  No wonder that robotic surgery was listed as one of ECRI's 2015 Top 10 Medical Technology Hazards (as it also was in 2014).

Robotic surgery started with the surgeon present, but by 2001 had already moved to telesurgery, where the surgeon was in a different location.  "Operation Lindberg" made international news in 2001 when surgeons in New York successfully removed the gall bladder of a woman in France.  Other such remote surgeries have followed.

The next logical step  would seem to be robots operating independently, starting with routine surgeries but over time graduating to more complicated ones.  Perhaps an autonomous robot would do just as well, or better.  Better vision, ability to make more precise incisions, faster reflexes, more massive and real-time knowledge base, maybe even a wide array of built-in attachments.  Perhaps humans just slow things up.

We shouldn't assume that the purpose of robots in surgery is simply to enhance our capabilities.

If we're going to have robots in health care, though, surgery may not be the area of most need.  After all, there are plenty of surgeons (although not well dispersed geographically) and we probably do too much surgery already.  On the other hand, one health care professional that we do not have enough of is personal care workers, and that shortage is only going to get worse.

Fortunately, robotics companies are already on it.  Toyota, for example, has a whole "family" of partner robots, including ones for health care  Disney even had an animated movie last year -- Big Hero 6 -- that featured such a robot.  That movie was set vaguely in the future, but let's hope the robot of personal care robots is not so far off.

The robots may be ready before we are.  A recent survey in the U.K. by The Institution of Engineering and Technology found that only 33% would use a robot to assist an elderly or disabled relative, while 17% would not even consider it.  Interest was lower when considering assistance for oneself.

As someone who has family members in long term care facilities which always seem to be short-staffed, my only question about using robotic aides to help them would be: where do I sign up?

The question is not really robotic technology will get sophisticated enough to add value in health care -- their role in manufacturing has clearly proven their ability to automate extremely complex processes -- as it is when the supporting software (dare I say "artificial intelligence"?) will allow it to operate with limited or even no human oversight.

Harvard Business Review recently proclaimed that "artificial intelligence is almost ready for business," and specifically cites health care as one of the leading applications, such as IBM's Watson.  Pittsburgh's UPMC, University of Pittsburgh, and Carnegie Mellon just announced their big data collaboration, which UPMC CEO Jeff Romoff says could lead to "doctor-less health care," a quote I bet his PR folks were trying to walk back as soon as the words came out of his mouth.  

Oh, there will still be doctors -- but they won't be our only option, for diagnosis, treatments, even procedures and surgeries.  Not tomorrow, probably not next year, but it will happen.

Of course, the licensing cartel will be a problem.  If organized medicine puts up barriers to doing telemedicine even for physicians who have licenses in other states, I have a pretty good idea how they'll respond to robots getting in on the act.  As was true with telemedicine, I suspect that the technical barriers will be breached well before the cultural and regulatory ones are.

If this kind of future sounds impossible to imagine, keep in mind the words of robotics pioneer Rodney Brooks:
In the future, I'm sure there will be a lot more robots in every aspect of life.  If you told people in 1985 that in 25 years they would have computers in their kitchen, it would have made no sense to them.

No comments:

Post a Comment