New machine could one day replace anesthesiologists | Hacker News Reader

New machine could one day replace anesthesiologists

washingtonpost.com

33 pointspetethomas24 comments

Non-anesthesiologist ICU physician here so slight bias should be noted.

I'm a huge fan of this. Machine will be much more anal about minor issues that some anesthesiologists / CRNAs may ignore. Additionally, this will greatly bring down costs for healthy routine procedures that need sedation. Machines don't "get bored" and if things fall out of parameters they happily yell to get attention. Properly coded they should also not make medication errors. Additionally, they should be able to record data and find patterns in patients that unexpectedly don't do well.

Michael Jackson would probably still be alive if a machine had been used to help monitor his propofol!

Dialysis, for example, used to be a much more manual process. Now, it is mostly driven by nurses and machines. Heck, people even do self-directed dialysis at home.

This isn't the end of anesthesiology by any means. The more difficult patients such as those who are obese or have a bunch of health problems will still need more aggressive observations. However, I suspect these machines will move into bigger and bigger surgeries to help assist the gas-passers provide safer analgesia and sedation.

Just like robotic assisted surgery became too popular too fast and outcomes suffered, I suspect these type of machines will not be trouble-free. How long before the HN articles about one of these being hacked start appearing? Nevertheless, medical machines like this with tight feedback loops are essential for improving the safety and efficiency in the future practice of medicine.

From the name, I used to think that anesthesiologists just "put people to sleep" or stopped patient's pain.

My view is now different. Their job (for what I'm told) is also to keep people alive and stable. They administer drugs and employ machines and techniques to control body temperature, blood pressure, oxygen levels, heart rate and others. Sometimes it's a simple colonoscopy (like in the article), but other times they are treating people who just had a stroke or a bad accident. They have to react in seconds to the response they SEE on the patient (data coming from machines, but also body movement, change in skin color, eye movement, bleeding, and many subtle changes), connecting knowledge and years of experience. It's also an art to be able to put tubes and needles on all kinds of "strangely" shaped bodies (specially in old people).

A more precise title would be "new machine could one day replace some anesthesiologists". I think it's going to take a while before we can trust machines to keep us alive during emergencies and operations.

My sister is an anesthesiologists. She sometimes anesthetize heart or brain surgery patients without putting them to sleep, because they're too old or unhealthy but need the surgery asap. These people normally cannot come back to life from anesthesia, which is just making someone dead by stopping heartbeat and respiratory functions with anesthetic agents and then animating them with machines. So the creepy state of having your heart worked out while you're awake. I believe this is something that can be fully automated by a machine in the future.

But she also holds the hand of a child to calm her while preparing her to a scary surgery, or talks to relatives waiting outside of intensive care, sometimes tells people that they lost their loved one, or a miracle just happened and they're good.

>These people normally cannot come back to life from anesthesia, which is just making someone dead by stopping heartbeat and respiratory functions with anesthetic agents and then animating them with machines.

Hearts do not stop beating during ordinary general anaesthesia.

You're right. My bad.

My mother was a nurse anesthetist like Bill Clinton's mother, and she had some horror stories that some of her colleagues had experienced, like a child who hid some gum under his tongue who then required some furious suctioning after being put under and it migrating to his airway. Add the providing comfort function that gokhan mentions and that I remember from when my tonsils and adenoids were removed in the late '60s when I was in first grade, or that my brother mentioned from surgery in the last decade or so, and it's clear there will still be a human involved even if a machine takes over some functions.

ADDED: your impression of what they do, their focus on a patients vital signs, easily measurable or subtle, their focus on keeping the patient alive, is correct as I understand it from my mother.

An even more precise one would be "new machine could one day replace some functions anesthesiologists perform".

Like most of the knowledge worker automations, it is not a matter of completely replacing the worker, but simply assisting him. I always roll my eyes at "the robots are comming for your jobs". If you are a cashier, maybe. A manager? Nope.

"But now, after the procedure, she said she saw a potential upside, too: There was no human error, either."

I knew it. Programmers are not human.

I'm saddened by the glorified view on non human systems. When they'll reach a large scale enough, there will be too-complex conditions, confusion and bad (contextually optimal) decisions and ordering, just like what happen to humans.

Code can reach close enough to perfection for errors to pretty much not happen with a process like that NASA coding team

The technology that goes into a self driving car is possibly far more sophisticated than this new machine. Anesthesiologists must have lobbied hard to stop this machine from a more wide spread usage. However the writing on the wall is clear. Machines in many cases are better than Doctors and they will continue to become better more rapidly than say Doctors with their experience. I wish US FDA gives a chance to these innovations and slowly tackle the rising healthcare costs.

Passenger aircraft can take off, fly to their destination, and land all on their own, but it hasn't removed pilots from their jobs. Automation doesn't catch all edge-cases, and in life-or-death situations, there's always going to be human oversight.

Yes, some standard procedures on ASA1 patiens do not require a lot of sophistication. However do not underestimate how hard anesthesiology is, while 90% routine, 10% is incredibly complex stressfull scenarios in trauma/emergency, surgical complications, adverse events. Its much more than just heart rate, RR, oxymetry etc. Its knowledge about medicine interactions, organ systems, physiological systems and the skill to perform under immense pressure (trying to provide a permanent airway through facial trauma or large bore IV access for fluid replacements on a patient in volumetric shock) And while uncommon it can happen during standard procedures in ASA1 patients. I'd rather have the anesthesiologist close please.

As a quick sedation system these sound wonderfull though, for example for resetting joint dislocations in the OR where all emergency staff and equipment is already in place.

ASA1 == American Society of Anesthesiologists physical status classification system "Healthy person", e.g. https://en.wikipedia.org/wiki/ASA_physical_status_classifica...

With an increasingly graying population in the US, rdudekul's hopes ... well, I suppose if you can decrease e.g. the ASA 1 need for anesthesiologists while they're still on tap if things go south, it'll overall help. But it's probably not going to get big, and will create triage like situations where people will needlessly die when there's no anesthesiologist to spare.

As someone who has "been under" a few times, I am torn on this. I like the idea of less errors due to a computer, but there is something comforting about knowing that a human expert is "at the wheel".

I know it is irrational, but I feel similar about airplanes and cars -- much more comfortable when you know there is a human at the helm.

Maybe it is the mortal peril part of the equation.

What is the part of trust vs new ? How many devices were scary and non automated before which are now invisible to our fear ?

The advantage of humans over machines, they have a broader view on things; for now.

I could foresee a setup like airline pilots - the autopilot truly does most of the work, but humans monitor and are available in case of emergencies.

If the machine could detect the awareness well enough then it could reduce the terrifying "accidental awareness" during surgery:

http://www.iflscience.com/health-and-medicine/large-study-ex...

http://www.dailymail.co.uk/news/article-2750201/Three-patien...

so does the machine also open the patients airway and safely insert the breathing tube ?

Perhaps that could be done by one of the many other people there.

i think it's quite a skilled job and one that if done incorrectly can be fatal. this monitoring of the airways and being able to react is one of the reasons why i think the machine will only ever be a complement and not a replacement for an anesthetist.

Or answer codes, sign out medication, or take the patient their consent forms? It's cool, but no replacement for a person.

The article is telling specifically that the machine is already replacing a person in some circumstances.

The kind of circumstances that it can manage are not going to shrink.

well obviously not. Don't let the article's title get you too riled up. This tool won't be replacing anesthesiologists completely. But this will help with this part of the job. Machines are useful tools and should be though of as such. An anesthesiologists shouldn't worry about this machine anymore than an airplane pilot worries about the autopilot.

Reply on news.ycombinator.com