Will tomorrow’s physicians perhaps use “autodoctors” to handle routine steps of a diagnosis or treatment, while the physician only steps in for critical or unforeseen parts?

A couple of weeks ago now I went to watch the latest Ridley Scott movie – Prometheus. The movie’s main character is a young scientist by the name of Elizabeth Shaw leading a team of explorers to distant parts of the universe on a quest to discover clues to the origins of mankind on Earth.

One might wonder how this can possibly relate to medicine (unless one has taken an unusually strong interest in the study of alien anatomy), but it becomes clear when young Shaw in a particularly gruesome scene of the movie uses a robot to perform an automatic open surgical procedure on herself.

The futuristic Med Pod 720i (a name which indeed sounds like a cheap inkjet printer) is presented as the latest advancement in medical science – capable of performing automatic diagnosis and surgery, and equipped with impressive features such as “Computer Controlled Robotic Surgical Arms, Liquid Spray Anesthetic and Vital Signs Sensors”.

Could this really be a sign of what the future might look like?

From a patient safety perspective, developments in this direction make a lot of sense. In the aviation industry, it is a long known fact that technology which transfers repetitive tasks from humans to machines can vastly improve safety. Together with rigorous training programs, flight simulation and checklists, automated systems such as the airplane autopilot are the pillars of what makes flying today as unbelievably safe as it actually is.

Maybe tomorrow’s physicians will use an analogous “autodoctor” to handle routine steps of a diagnosis or treatment, while the physician only steps in for critical or unforeseen parts?

Vinod Khosla, one of the co-founders of Sun Microsystems, seems to think so. In a recent talk, he argued that Google’s development of a driverless smart car was “two orders of magnitude more complex” than providing the right diagnosis for a patient, and that a good machine learning system would be cheaper, more accurate and objective than a doctor.

Needless to say, Khosla’s comments sparked a great deal of debate, but the age of “big data” is no doubt in the coming and assisted diagnosis systems like the ones Khosla spoke about are already being developed, most famously IBM’s Watson computer – “the world’s best second opinion”.

When it comes to automated treatment delivery, however, technology still has some way to go. For endovascular and minimally invasive procedures, surgery robots do exist but they are so far mostly used for “remote control”. Even if they provide some improved safety in terms of preventing hasty and potentially dangerous movements, a human physician still has to manually carry out every single step of the procedure.

To make the Med Pod 720i become reality, an expert system that can substitute for the physician’s medical knowledge would therefore be needed, to act as a link between medical imaging equipment and surgical robots, capable of telling the robot where to go and what to do based on what it sees.

Curiously enough, at Mentice we use a similar computer algorithm (incidentally nick-named “Dr. Robot”) to automatically navigate catheters and wires and perform basic operations in a simulated endovascular environment. However, it’s only used for the purpose of testing our simulators – to attempt the same thing in real life would obviously be an infinitely more complex task.

Even though we’re still a far cry from the amazing capabilities of the Med Pod 720i, many of the individual components that would be needed to put such an automated surgery system together, at least for minimally invasive procedures, actually already exist. While regulation, safety and ethics concerns and remaining technical shortcomings may hold off this development for another decade or two, don’t be surprised if the next time you visit a hospital, the physician will use his or her “autodoctor” to do the bulk of the work.

Edward Fält