Robot-assisted Surgery: An Example of Growth Ambiguity

As everyone who reads this blog knows, I worry a lot about whether innovation is proceeding as fast as people think it is. That’s especially true in healthcare, where years of apparent advances have not cut death rates for middle-aged Americans.

But some innovations are just tough to assess.  Back in the late 1990s,  a new medical technology was introduced with enormous fanfare: Robot-assisted surgery. The idea was that it was possible to partially automate surgery and drive down healthcare costs, just like manufacturing automation had driven down the cost of producing goods.  In addition, robot-assisted surgery promised to have better outcomes.

The technology took some time to get traction, and then gained ground quickly.  The best way is to track the stock price of the market leader, Intuitive Surgical. The company went public in June 2000. At the end of 2000, the share price was $17. At the end of 2003, the share price was  $16.97.  Then it went shooting up (see stock chart here). Now the latest quote is an astounding $332.

But just this weekend a story  appeared in the NYT about robot-assisted surgery. The story,  entitled “Results Unproven, Robotic Surgery Wins Converts“,  points out that the technology has been very successful commercially.

Last year, 73,000 American men — 86 percent of the 85,000 who had prostate cancer surgery — had robot-assisted operations, according to the robot’s maker, Intuitive Surgical, the only official source of such data. Eight years ago there were fewer than 5,000, Intuitive says.

Here’s the key sentence:

But robot-assisted prostate surgery costs more — about $1,500 to $2,000 more per patient. And it is not clear whether its outcomes are better, worse or the same.

The article then goes on to say that:

Medical researchers say the robot situation is emblematic of a more general issue. New technology has sometimes led to big advances, which can justify extra costs. But often, technology spreads long before investigators know whether it is worthwhile.

As a ‘skeptical technological optimist’, the case of robot-assisted surgery is utterly perplexing for me. It’s exactly the sort of technology that should pay off big, either in cost or in health outcomes or both. But then I would expect the  gains to be big enough to see. Or, is it just too soon?

I think I’ll give the folks at Intuitive a call.


  1. The problem is measuring everything by just one statistic, death rates, rather than other, better quality of life statistics. For example, a lot of older people are getting hip and knee replacements nowadays, which isn’t going to make them live any longer, but will allow them to be much more mobile than they otherwise would have been. However, measuring this would mean coming up with a statistic for mobility, something that is a bit tougher to measure and that would require more data collection. It can be done but people are just too lazy to do it. Another drug mentioned in the article was Viagra, how many older people didn’t have as good a quality of life without it a decade or two ago? As for robot-assisted surgery, I’m skeptical that it does much more than create smaller incisions, with possibly worse accuracy, so the jury’s out on that one. As for medicine increasing quality of life overall, I’m skeptical that that’s happening much either, or at the very least that the giant cost increases are doing much more than improving quality of life by small amounts. However, I don’t see the statistics to back up either that argument or the opposite.

    Mike, you may want to add a wordpress plugin to highlight recent comments in the sidebar, that was actually a useful feature on your old blog.

  2. Innovation doesn’t have to be technology, it might be process, distribution etc. You may enjoy reading the Checklist manifesto which highlights the lifesaving use of checklists in reducing inline infections a dramatic result. I reviewed the book here in the context of hedge funds, but it is an innovation:

    As a former chief analyst for Europe’s equivelant to the MIT, medialab, I know how people can get carried away with technology masked as innovation. Real innovation is outcomes based.

  3. There is always the question of how many of these operations per unit time the robots can complete compared to their human counterparts. I’d imagine that even if the surgery was cheaper done by human if the hospital could make more money from the volume of surgeries done by robots compared to the profit margin on a PhD’s operations, well, who wants low throughput human instead of a high throughput robot?

  4. Mike: As for being perplexed, the first question you have to ask is, why precisely do you expect the robot device to be superior to a competent surgeon, aside from general stereotypes that “computers are always faster and better”?

    And having worked in an automation industry, it is critical to understand the sales pitch and pricing model for any labor saving automation. There are roughly two types of automation equipment/process – those that can do things that are physically infeasible for humans to do, or to do with reasonable speed or yield, and those that replace humans producing the ostensibly same quality/throughput (where the automated activity is entirely feasible for one or more humans).

    The pricing model for the latter is strictly based on the “saved” human employment cost (or claims to the effect how large the savings are), with a discount large enough to make the sale palatable. There are no cost savings to be passed on to the consumer, at least until the technology becomes commodity and competition drives down prices (if ever). Meaningful competition can usually only emerge when the market for the product is large enough to support enough players.

    Another consideration esp. in the medical field is liability, and the enormous development costs to engineer around it.

  5. I would guess the big impact of robotic surgery is repeatability. I suspect that, much like drag racers have moved towards automatic shifters, surgery will move towards robotic surgeons to reduce the rate of malpractice suits based on error in surgery. I’d be curious how robotic surgeons affect malpractice rates: that would be a good leading indicator.

    • I don’t claim to know anything about surgery. There are probably some types of surgery that are more routine than others, and lend themselves better to automation or mechanization. Maybe prostate surgery is one of those.

      In automation of cognitive tasks (feature detection, classification, interpretation) there is a concept of “reject” which roughly means “can’t do”, and the fundamental quality tradeoff is between error and reject. Typically to bring the error rate down a bit you have to reject good cases at a higher rate. When the reject to meet an error limit gets so high that the required recognition performance is not met this means the task cannot be automated (using the algorithms/technology under consideration). With robotic (assisted) surgery there must surely be a similar phenomenon.

      • I’m sure you’re right about that, but I would also expect there to be a sizeable amount of surgery amenable to robotic assistance. It would be interesting to see how much is not, though.

  6. Expanding on others’ comments, it seems likely that the answer might lie hidden in some of your past articles, Mr. Mandel.

    You demonstrated that the health industry is the exception to a steady decline in return on assets, even in the recent decade of generally low capital costs, suggesting that surgery times and/or hospitalization times were shortened in a way that enabled redirecting resources toward higher profits.

    Before you spotted a recent disruption, you showed us that real wages for highly educated high earners were holding or increasing, i.e. surgeons, contrary to the pattern for the vast majority of earners. If I recall, a reader took exception by suggesting that benefits were increasing, which may well be the case, that benefit being the dramatically inflating cost of health insurance, since it seems that the health care industry has a captive audience that enables it to pass along cost increases.

    If you can expand this into a paper that points the way to getting us out of the health care cost trap we are in, that would be good, until someone accuses you of trying to take away their high tech surgery.

  7. Dean Jackson says:

    I have a sister who is an OB GYN and uses the Da Vinci. She says it is very superior to conventional surgery. Patients spend less time in the hospital and recover so quickly compared to conventional that it is amazing. Less invasive equals a less societal cost.

  8. The Fifth Horseman says:

    The savings is not in surgery costs. It is in recovery costs, which are slashed by two-thirds. This is because the incision is so much smaller (just an inch, rather than an incision that has to be large enough for the surgeon’s hands).

    So it is not surgeons getting their jobs automated away, but nurses.

  9. Mike Mandel says:

    >So it is not surgeons getting their jobs automated away, but nurses.

    That’s a very interesting point. Thanks.

  10. The Fifth Horseman says:

    Let me also add that major heart surgeries done via machines from Intuitive now can go between ribs, rather than the traditional practice of sawing the breastplate in half and opening it.

    Again, the surgeons hands are no longer needed, so things can be done without sawing any bones at all.

    As a result, hospital stays of 15 days are down to just 5 days. At-home recovery times of 9 weeks are now just 3 weeks.

    The capacity of nurses and hospital bed-days consumed is slashed. Productivity lost while the patient recovers is slashed. The unappealing prospect of sawing bones to get to the heart is eliminated.

    And lastly, of course, the Da Vinci will drop in price, just like anything heavy in electronics quickly does.

    If your look at revenue and earnings growth rates, Intuitive is one the of best performing companies of the last 3 years. In the same rarefied tier as Apple and FirstSolar.

  11. Technology innovation could lead to so many health benefits such as the development of useful equipment for treatments. Although, I’m aware that robotics is programmed to help, we can’t predict when malfunction will occur. So, it’s better to conduct several trials first to avoid incidents that can worsen the patient’s condition.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s


%d bloggers like this: