Intubating robots are coming for your job (someday)
Undergraduate biomedical engineering students have developed the first prototype of a laryngoscope blade whose future iterations could allow novices or robots to intubate patients as well as expert humans.
The new laryngoscope blade has embedded force sensors that calculate its location in the airway. The sensors were trained using machine learning, with tactile patterns generated during intubations of manikins by expert anesthesiologists. LED lights provide instant feedback to the operator to direct subtle adjustments of the blade during insertion into the airway. It acts much like an expert saying “move your wrist back a little,” but more helpfully.
Called the “InTouch” laryngoscope, the gadget won first prize and $20,000 in a competition sponsored by the biomedical engineering division of the N.I.H. and the nonprofit VentureWell.
The device has so far only been demonstrated effective on airway manikins. The design team’s next task is to train the device using data obtained from human patients.
Intubating Robots: Coming Soon?
Although designed for use by human beings, similar technology could be used to facilitate intubation by autonomous machines, without human involvement.
Tracheal intubation is a procedure well-suited to robots. A human remote-controlled intubation system has already been successfully used at McGill University in Montreal. It’s a short step technologically to replacing the human with sensors to map the airway and guide the intubation stylet into the trachea. Engineers have built such intubating robots that use electromagnetic guidance (Ohio State) and infrared light (Hebrew University).
The eventual availability of such an intubating device seems inevitable; how quickly it arrives depends mostly on perceptions in the venture capital world of the size of the potential market.
Expect initial deployment by first-responder services, where a high failure rate of intubation seems to result in excess mortality, or in risk-averse or understaffed hospitals as a backup system for failed airways. Robots could do all their intubations "awake" -- eliminating the risks from the temporary coma and paralysis induced during rapid sequence intubation -- requiring instead only some fentanyl I.V. followed by atomized lidocaine to the airway.
Speed and breadth of the technology’s adoption in health care would depend on its cost, ease of use and efficacy. With time -- let’s guess 20 years conservatively -- all those metrics will likely favor intubation by the machine rather than most humans. Robot intubations will become common, at first in underserved areas, then more widely.
A vicious cycle will emerge (virtuous, to the robot): fewer intubations by humans lead to generally lower skills, while robot intubations improve, both in a real sense (through AI machine learning using data inputs from all robot intubations) and relatively (as average human skills decline). Airway experts will continue to equal or beat the robot, but average operators will not, especially on difficult airways.
To maintain skills, trainees and attendings will be allowed to occasionally intubate. Where physicians in training once competed only with each other for intubation experience, in the future everyone might have to compete with the robot.