Here’s an astonishing assertion: “We are convinced the machine can do better than human anesthesiologists.” This statement was made by a doctor. Not only a doctor but an anesthesiologist. Not just an anesthesiologist but a pediatric anesthesiologist.
Not just any old pediatric anesthesiologist but one in charge of pediatric anesthesia research at the University of British Columbia medical school in Vancouver. One can only assume that this guy has a pretty low estimation of what his colleagues can do. Must make for great break room conversation.
The doctor making this statement, one JM Ansermino, is co-creator of a new automated anesthesia system called, cutely, iControl-RP. This machine uses vital sign readings along with oxygen saturation levels and EEG monitoring to give anesthesia to patients undergoing surgery, without any human input. Thanks to my fellow blogger Karen Siebert for pointing this one out. http://in-training.org/why-not-not-go-anesthesia-karen-sibert-md-9276 and http://www.washingtonpost.com/blogs/the-switch/wp/2015/05/15/one-anesthesiology-robot-dips-its-toes-into-whats-possible-this-one-jumps-all-in/ iControl-RP is what the creators call “A scheme-based closed-loop anesthesia system”. Here’s what their paper says about it: (http://www.researchgate.net/publication/264860147_A_Scheme-Based_Closed-Loop_Anesthesia_System)
The system software in its entirety consists of approximately 22K lines of Scheme code and features a client-server implementation interfacing medical devices with portable graphical user interface. The strengths of the Scheme functional language have been leveraged to build a robust maintainable modular system with extensive testing facilities to mitigate the inherent safety hazards associated with the application. Blah, blah, blah… wait. “Inherent safety hazards?” You think? Look. Here is the “problem” that the folks in Vancouver are hoping to “solve”. Intravenous anesthetics like Propofol have dosage and administration guidelines that are, in practice, mere suggestions. People have a very wide range of responses to sedatives and the effects can not always be predicted. Which is why you have a human, an anesthesiologist or CRNA, up at the head of the bed controlling the dosing. This type of sedation, using intravenous agents only, is not usually called “General Anesthesia (GA)”. You can do a GA with just IV drugs but is not very common. Usually Propofol infusions are used in what we call a MAC or “Monitored Anesthetic Care” which is a really stupid name that basically implies that someone is giving sedation and monitoring how it works, but in reality means that the patient is sedated but retains the ability to maintain independent respiration, and someone is adjusting things so that both sedation and spontaneous breathing occur. In other words, you’re in la la land but don’t need to be put on a ventilator. We use it for procedures in which local anesthetic effectively numbs the surgical area so that all that is needed is to make the patient comfortable. There are a number of cases in which doing a procedure under MAC becomes difficult to impossible. For example, if the local anesthetic is not adequate to block the pain at the surgical site MAC is tricky because the level of sedation needed to eliminate the experience of pain usually also causes people to stop breathing. Obese patients are at risk of airway obstruction at even fairly low levels of sedation. A young teenager might still be reactive at very high levels of sedation. People with involuntary movements, tics, restless legs, etc. often do not lose these movements when sedated. Some people get more agitated the less aware they are of their surroundings. People who take drugs or chronic pain medications respond very differently and may prove impossible to sedate adequately under MAC conditions. In some cases the anesthesiologist might decide “switch from a MAC to a GA”, at which time the anesthesiologist takes over breathing for the patient and thus is able to provide an higher level of sedation. This decision is not made easily or lightly. It requires considerations about patient size and state of health, how effective the surgeon is at numbing the surgical area, how fast the surgeon is, what the anesthesiologists threshold is for switching, what operation is being performed, what the duration of action of the local anesthetic is, unexpected surgical findings or complications, and ultimately, the safety of the patient. If the decision is made an actual person, i.e, the anesthesiologist, has to secure the airway and then set the ventilator and gas parameters appropriately. Consider this from the iControl-RP paper:A closed-loop automated drug delivery system that administers drug based on feedback from measurement of a physiological end point can potentially improve the outcomes for patients by addressing the inter-patient variability, providing greater hemodynamic and respiratory stability, more stable depth of anesthesia, and reducing the total dose of drug delivered. Where in that sentence is there mention of any of the variables I just listed? Is there an algorithm for deciding a different approach is needed? Are there different settings for different surgeons? Do you dial in a different number for the super-anxious patient or the one with the Methadone habit? Inter-patient variability is what makes anesthesia part art, part science. Inter-patient variability is the very reason we need actual humans at the head of the bed. Patients are not Scheme code. Which is why… The “operating room scenario” depicted in figure 1 of the creators’ Researchgate article features an attending anesthesiologist. To monitor the iControl.