Future technologies in anesthesiology

[NumTacos]

I was pondering how safe anesthesia is these days, with a couple of technologies we all know that brought us to this point ie. pulse ox, capnography, video laryngoscopy, ultrasound, and even non-invasive cardiac output monitoring. Our drugs are great and reliable. I was trying to think of anything that would be revolutionary in our field.

One thing that obviously comes to mind is the use of AI, which is already being used to assess LV function based on ECG via machine learning. However, it seems like anything that would be a huge advance would likely be decades away. Things I think of would be AI generated decision support and predictive analytics, closed loop delivery systems, robotics, etc. However, it seems like for those things to be affordable and useful for actual clinical care are a pretty long way off. Does anyone have any ideas? Just kind of fun to think about.

---

Members don't see this ad.

 

[dabears505]

I think soon ultrasound machines with software that highlights nerves as you scan and needles that communicate with probe to generate picture even when not perfectly aligned.

The only thing that would stand in the way of a technician doing a pre op block is the issue with liability.

Also could see one day where you supervise 4 computers managing the gas, delivering drugs, etc. One doc does the procedure of incubating, placing lines, etc. Computer manages vent, gas, etc. Iv hooked up to a device which has common drugs and administers. Doc comes in to extubate, manage out of ordinary things.

 

[Crabbygas]

I think soon ultrasound machines with software that highlights nerves as you scan and needles that communicate with probe to generate picture even when not perfectly aligned.

The only thing that would stand in the way of a technician doing a pre op block is the issue with liability.

Also could see one day where you supervise 4 computers managing the gas, delivering drugs, etc. One doc does the procedure of incubating, placing lines, etc. Computer manages vent, gas, etc. Iv hooked up to a device which has common drugs and administers. Doc comes in to extubate, manage out of ordinary things.

But then who is going to make the table go up and down?

 

[NumTacos]

I think soon ultrasound machines with software that highlights nerves as you scan and needles that communicate with probe to generate picture even when not perfectly aligned.

The only thing that would stand in the way of a technician doing a pre op block is the issue with liability.

Also could see one day where you supervise 4 computers managing the gas, delivering drugs, etc. One doc does the procedure of incubating, placing lines, etc. Computer manages vent, gas, etc. Iv hooked up to a device which has common drugs and administers. Doc comes in to extubate, manage out of ordinary things.

Saw that, Admir Hadzic has a video showing how it works. Pretty amazing.

 

[T-burglar]

I have to say this is a timely post because I’m actually planning to go part time before the end of the year while salaries are high and self study software engineering .

Some of these AI applications will never see the light of day in my opinion but if there’s one area where anesthesia stands to grow explosively in innovation , it’s software . I’d bet on that before I’d bet on any new drugs in the pipeline .

 

[NumTacos]

I have to say this is a timely post because I’m actually planning to go part time before the end of the year while salaries are high and self study software engineering .

Some of these AI applications will never see the light of day in my opinion but if there’s one area where anesthesia stands to grow explosively in innovation , it’s software . I’d bet on that before I’d bet on any new drugs in the pipeline .

Curious, what do you plan to do with your studies in software engineering? Are you learning a specific programming language? I've thought about it myself, just for fun though.

 

[PpfSuxTube]

Does anyone have any ideas? Just kind of fun to think about.

Promise you won't run off and try copy the ideas?

 

[PpfSuxTube]

Saw that, Admir Hadzic has a video showing how it works. Pretty amazing.

It really isn't. It's just pattern recognition and has been around for decades in radiology, just got slapped onto an USS device. Except in rads, their x-rays and CTs don't have to deal with oodles and oodles of fat and skin folds that USS has to surmount. That will always be the issue with USS.

It is so far away from being able to have a machine automously perform a nerve block it's not even funny. I was involved in a test development plan for project just like this a few years back. Denied funding due to so many reasons...

 

[T-burglar]

Curious, what do you plan to do with your studies in software engineering? Are you learning a specific programming language? I've thought about it myself, just for fun though.

Very generally - learn to write web based applications that can provide solutions for workflow issues. I’m teaching myself HTML, CSS, JavaScript, PHP and Python to start.

I’m also getting Epic EMR builder training to hopefully be able to interface these two areas. Creating web based applications that interface with the Epic API. Will this ever pan out into anything? Idk maybe not. But I like it and I’m bored of being in the OR every day.

 

[PpfSuxTube]

Also could see one day where you supervise 4 computers managing the gas, delivering drugs, etc. One doc does the procedure of incubating, placing lines, etc. Computer manages vent, gas, etc. Iv hooked up to a device which has common drugs and administers. Doc comes in to extubate, manage out of ordinary ththings.

You know this has been tried and failed massively already. Machine couldn't even manage sedation for the easiest of colonoscopy cases.

 

[NumTacos]

You know this has been tried and failed massively already. Machine couldn't even manage sedation for the easiest of colonoscopy cases.

I agree; we are not even close to realizing some of this stuff. Just because it's been tried doesn't mean it won't be tried again. I'm curious how far off we might be. Looks like decades at least, though.

 

[pgg]

Very generally - learn to write web based applications that can provide solutions for workflow issues. I’m teaching myself HTML, CSS, JavaScript, PHP and Python to start.

I’m also getting Epic EMR builder training to hopefully be able to interface these two areas. Creating web based applications that interface with the Epic API. Will this ever pan out into anything? Idk maybe not. But I like it and I’m bored of being in the OR every day.

I've done some of this. It is interesting and even fun.

There is absolutely demand for physicians who both understand what kind of tools are helpful for daily work flow, and can build those tools. I think the right person could become truly wealthy with the right ideas and execution.

But the software isn't the hard part. The compliance bits are where things get hard. From roughly 2014-2022 I built and maintained a web based anesthesia scheduling app for the Navy (far better, IMO, than any of the commercial **** out there), and easily 90%+ of my time after the first year or so was spent on compliance tasks. And that was for a system that had minimal/trivial connections to other systems that stored PHI/PII.

I found the hurdles to be essentially insurmountable, absent at least a couple of full time people dedicated to nothing but compliance. For a general frame of reference, there are well over 300 NIST security controls for a low confidentiality, low integrity, low availability system. That's without privacy overlays or HIPAA stuff. You start getting into the moderate and high security tiers that come into play with even the tiniest shred of sensitive data, and it's absolutely mindblowing how the requirements start to grow. Penalties for failures to comply are huge.

We ended up outsourcing most of the compliance tasks to cloud.gov, by paying for their platform-as-a-service to run our app and database. This alone increased our hosting costs more than 100-fold. Literally 10s of thousands of dollars for a virtual server and connectivity that any commercial host could provide for $300/year.

I wrote software before going to med school. When I took on that project in 2014 I had thoughts of re-entering that industry as a sideline or alternative to medicine. The coding was cool and fun. Everything else sucked in a way that grated on my soul. I'm just going to be a doctor.

Good luck though. The building and creating is fun and I do miss it.

 

[anbuitachi]

i think it has to have cost benefit. if a system provides bit of better care on top of our already super safe system, it better be worth the cost. i dont see hospitals purchasing AI software for hundreds of thousands etc.

once in a while reps come and show their new toys, seems cool and potentially useful, but usually not at the cost they say

 

[pgg]

End-tidal propofol is a monitor I'd like to see available. I wouldn't call it game changing the way capnography or sugammadex were, but it'd be pretty awesome. The biggest hassle with longer duration TIVAs is the difficulty of timing the wakeup, and the biggest risk of a TIVA is the bad IV or other issue that risks awareness. End-tidal propofol neatly solves both of those issues.

 

[sleepallday]

I think the application of AI will be more useful as tools for diagnostic purposes rather then actually administering or controlling the anesthetic. Edward lifesciences came out with some sort of predictive hypotension software. I'm hoping to get it at my place to play around with it.
Acumen Hypotension Prediction Index software

 

[NumTacos]

I've done some of this. It is interesting and even fun.

There is absolutely demand for physicians who both understand what kind of tools are helpful for daily work flow, and can build those tools. I think the right person could become truly wealthy with the right ideas and execution.

But the software isn't the hard part. The compliance bits are where things get hard. From roughly 2014-2022 I built and maintained a web based anesthesia scheduling app for the Navy (far better, IMO, than any of the commercial **** out there), and easily 90%+ of my time after the first year or so was spent on compliance tasks. And that was for a system that had minimal/trivial connections to other systems that stored PHI/PII.

I found the hurdles to be essentially insurmountable, absent at least a couple of full time people dedicated to nothing but compliance. For a general frame of reference, there are well over 300 NIST security controls for a low confidentiality, low integrity, low availability system. That's without privacy overlays or HIPAA stuff. You start getting into the moderate and high security tiers that come into play with even the tiniest shred of sensitive data, and it's absolutely mindblowing how the requirements start to grow. Penalties for failures to comply are huge.

We ended up outsourcing most of the compliance tasks to cloud.gov, by paying for their platform-as-a-service to run our app and database. This alone increased our hosting costs more than 100-fold. Literally 10s of thousands of dollars for a virtual server and connectivity that any commercial host could provide for $300/year.

I wrote software before going to med school. When I took on that project in 2014 I had thoughts of re-entering that industry as a sideline or alternative to medicine. The coding was cool and fun. Everything else sucked in a way that grated on my soul. I'm just going to be a doctor.

Good luck though. The building and creating is fun and I do miss it.

I've done some of this. It is interesting and even fun.

There is absolutely demand for physicians who both understand what kind of tools are helpful for daily work flow, and can build those tools. I think the right person could become truly wealthy with the right ideas and execution.

But the software isn't the hard part. The compliance bits are where things get hard. From roughly 2014-2022 I built and maintained a web based anesthesia scheduling app for the Navy (far better, IMO, than any of the commercial **** out there), and easily 90%+ of my time after the first year or so was spent on compliance tasks. And that was for a system that had minimal/trivial connections to other systems that stored PHI/PII.

I found the hurdles to be essentially insurmountable, absent at least a couple of full time people dedicated to nothing but compliance. For a general frame of reference, there are well over 300 NIST security controls for a low confidentiality, low integrity, low availability system. That's without privacy overlays or HIPAA stuff. You start getting into the moderate and high security tiers that come into play with even the tiniest shred of sensitive data, and it's absolutely mindblowing how the requirements start to grow. Penalties for failures to comply are huge.

We ended up outsourcing most of the compliance tasks to cloud.gov, by paying for their platform-as-a-service to run our app and database. This alone increased our hosting costs more than 100-fold. Literally 10s of thousands of dollars for a virtual server and connectivity that any commercial host could provide for $300/year.

I wrote software before going to med school. When I took on that project in 2014 I had thoughts of re-entering that industry as a sideline or alternative to medicine. The coding was cool and fun. Everything else sucked in a way that grated on my soul. I'm just going to be a doctor.

Good luck though. The building and creating is fun and I do miss it.

Are the compliance issues only because you were working on mil/gov systems?

 

[NumTacos]

I think the application of AI will be more useful as tools for diagnostic purposes rather then actually administering or controlling the anesthetic. Edward lifesciences came out with some sort of predictive hypotension software. I'm hoping to get it at my place to place around with it.
Acumen Hypotension Prediction Index software

That's kind of where I see it going first, predictive decision support, maybe even sophisticated patient specific pharmacokinetic modelling that's good for dosing/timing of drugs.

 

[NumTacos]

End-tidal propofol is a monitor I'd like to see available. I wouldn't call it game changing the way capnography or sugammadex were, but it'd be pretty awesome. The biggest hassle with longer duration TIVAs is the difficulty of timing the wakeup, and the biggest risk of a TIVA is the bad IV or other issue that risks awareness. End-tidal propofol neatly solves both of those issues.

Had no idea that was a thing, but it makes sense. I would imagine that perhaps many other drugs could be better titrated using a system like this.

https://onlinelibrary.wiley.com/doi/full/10.1111/aas.14043

 

[NumTacos]

i think it has to have cost benefit. if a system provides bit of better care on top of our already super safe system, it better be worth the cost. i dont see hospitals purchasing AI software for hundreds of thousands etc.

once in a while reps come and show their new toys, seems cool and potentially useful, but usually not at the cost they say

That's going to be how it eventually roles out. A lot of the AI stuff will just be integrated into the newer machines we buy, and most vendors will have something similar. Robotics will obviously take longer. No doubt many of these things will happen, but the time scale to when we will see it is interesting.

 

[Hoya11]

I was pondering how safe anesthesia is these days, with a couple of technologies we all know that brought us to this point ie. pulse ox, capnography, video laryngoscopy, ultrasound, and even non-invasive cardiac output monitoring. Our drugs are great and reliable. I was trying to think of anything that would be revolutionary in our field.

One thing that obviously comes to mind is the use of AI, which is already being used to assess LV function based on ECG via machine learning. However, it seems like anything that would be a huge advance would likely be decades away. Things I think of would be AI generated decision support and predictive analytics, closed loop delivery systems, robotics, etc. However, it seems like for those things to be affordable and useful for actual clinical care are a pretty long way off. Does anyone have any ideas? Just kind of fun to think about.

wireless monitors. i wouldnt have to take them off the patient when i transport, and when i come upon a new monitor in PACU i can link to my patient wirelessly. how far out could it be?

 

[pgg]

Are the compliance issues only because you were working on mil/gov systems?

That's a good question and I don't know exactly. Ultimately it depends on the customer to decide what standards or certifications they demand in the product they're paying for. NIST (National Institute of Standards and Technology) maintains these standards and they are used by civilian industry. The fed gov has its very tedious risk management framework (RMF) process by which systems are evaluated and certified, and if a company wants to market an IT product to any federal agency this long, expensive process is required. I suspect most companies making healthcare products wouldn't want to be locked out of the VA or military organizations.

Much as it pains me to admit it, there are good reasons for most all of the security controls specified by NIST. If you're going to make a serious, non-amateurish go at making an enterprise scale system (vs some one-off custom tool for a small group) then you're going to have to explicitly address all of the management policies, record keeping, log retention, version control, access controls, network security, physical security ... or when the inevitable day comes that a user does something dumb and a data breach occurs you won't have a leg to stand on when the HIPAA police come with their hammers. So whoever the customer is, you've got to be familiar with these standards.

All that said, I'm sure there's plenty of ****ty software out there from fly-by-night companies that doesn't bother with any of that, and behind the polished front it's just a rickety patchwork of hackery. Most every data breach or ransomware attack or other issue that hits the news is because some of these best practices weren't followed.


tldr - whether or not you explicitly document compliance with all of the applicable controls in NIST SP 800-53 Rev. 5 you need to be aware enough of them to understand which really are important and which you might be able to blow off for your app

 

[T-burglar]

Thanks for the insight PGG. Are the regulatory issues less of an issue if you keep the program(s) “in-house” as opposed to trying to commercialize it?

 

[Bertelman]

End-tidal propofol is a monitor I'd like to see available. I wouldn't call it game changing the way capnography or sugammadex were, but it'd be pretty awesome. The biggest hassle with longer duration TIVAs is the difficulty of timing the wakeup, and the biggest risk of a TIVA is the bad IV or other issue that risks awareness. End-tidal propofol neatly solves both of those issues.

Didn't you know? That monitor already exists-

BIS.

🙄

 

[Bertelman]

I wrote software before going to med school. When I took on that project in 2014 I had thoughts of re-entering that industry as a sideline or alternative to medicine. The coding was cool and fun. Everything else sucked in a way that grated on my soul. I'm just going to be a doctor.

I love that you managed to find an industry more soul-sucking than modern medicine. Then recommitted to medicine because it was actually better.

Though to be honest, I don't find Anesthesiology to be nearly as toxic as some other specialties have become. But I'm currently on the other side of a truly ****ty 7 years in AMC hell.

 

[Simba1711]

I would love wireless monitors like someone else said. They’ve had wireless infusion pumps that automatically transfer data to the emr in japan when I rotated there about 10 years ago. Not sure why we don’t have it over here.

 

[NumTacos]

wireless monitors. i wouldnt have to take them off the patient when i transport, and when i come upon a new monitor in PACU i can link to my patient wirelessly. how far out could it be?

Great thought, it would be so practical, and not just in the OR like you said. Now that I think of it, don't they do that already on the wards for remote telemetry? I think the problem is that the equipment is bulky.

What I’d like to see is some serious tricorder stuff, or at least miniaturized sensors that can be placed anywhere on the body and give you everything, wireless of course. Some day maybe!

 

[pgg]

Thanks for the insight PGG. Are the regulatory issues less of an issue if you keep the program(s) “in-house” as opposed to trying to commercialize it?

If you can get the locals who run your place to OK it and sign off on the risk, you can do whatever they let you do. I think that's totally sane and reasonable if what you're doing is low risk. Our app was a scheduling system that doesn't contain any patient data, so the consequences of any kind of unauthorized access, data corruption, data loss, or downtime wouldn't be a legal risk. (We do import PHI-stripped surgical schedule data because that's key to the module that assigns people to ORs each day. That process has a few layers of safety, and was scrutinized very carefully. The main thing that kept me awake at night was the possibility of a dumb user manually entering prohibited data into our system, despite the warnings and safeguards. There's just no sense of humor or oopsie-daisy-oh-well in that area. )

It may be simpler if you run it on their intranet (where their IT people are doing all the network and physical security) vs asking to put "their" data on a commercial server someplace. They'll probably think any data of any kind that's even tangentially related to hospital operations is "their" data. But if you put it on an Amazon server because you want people to access your web app from outside the hospital, without a bunch of VPN hoops to jump through, they may be less willing.

Of course the people that enforce HIPAA aren't going to cut you any slack if you screw up.

Brief problems with reliability, downtime, data loss might be tolerable if it's you and your own group, less so if it's a commercial product. But even friends and colleagues can be a little demanding if the thing you built, that they've come to rely on, suddenly has an issue, even if it's not your fault.

 

[dumbmedstudent]

I heard electronic charting will be available soon at my hospital

 

[PpfSuxTube]

I was a computer engineer before medicine, worked for about 6 years in a hardware design company that did automotive but also Healthcare applications and home entertainment. Massive multinational. I was smart enough but orders of magnitude behind the top guys that were brain storming new ideas, and there were a lot of them.

Anything that we can think of doing has probably already been done or started. There are very few low hanging fruit left.

 

[NumTacos]

I was a computer engineer before medicine, worked for about 6 years in a hardware design company that did automotive but also Healthcare applications and home entertainment. Massive multinational. I was smart enough but orders of magnitude behind the top guys that were brain storming new ideas, and there were a lot of them.

Anything that we can think of doing has probably already been done or started. There are very few low hanging fruit left.

That's the thing, it will only be adopted if it's affordable and does it actually change patient outcomes or make care more efficient. Lots of stuff I’ve seen is gimmicky and/or just too expensive. I occasionally do those marketing interviews, and in one interview they asked my thoughts about a BP cuff that can give a cardiac output every 5 minutes. It's like, why?

 

[2Fast2Des]

I would love wireless monitors like someone else said. They’ve had wireless infusion pumps that automatically transfer data to the emr in japan when I rotated there about 10 years ago. Not sure why we don’t have it over here.

Because we can't have nice things in this country...

 

[IMGASMD]

Because we can't have nice things in this country...

We are also more litigious than other places.

Someone has to be responsible for any outcomes we don’t like, or more so recently anything we don’t agree with.

 

[nimbus]

They’ve had wireless infusion pumps that automatically transfer data to the emr in japan when I rotated there about 10 years ago. Not sure why we don’t have it over here.

We transitioned from Alaris to Plum 360 pumps in 2019 because the Plum pumps were supposed to have that capability. Epic interoperability was supposed to go live in 2020 but it hasn’t happened yet.


Edit: Apparently both Plum and Alaris have this feature now. Hospital just needs to pony up for it.

 

[Bertelman]

We transitioned from Alaris to Plum 360 pumps in 2019 because the Plum pumps were supposed to have that capability. Epic interoperability was supposed to go live in 2020 but it hasn’t happened yet.


Edit: Apparently both Plum and Alaris have this feature now. Hospital just needs to pony up for it.

That doesn't make me feel good. Just switched to Plum this week, was told "within a year" the programmed drip rates will communicate with EMR

 

[nimbus]

That doesn't make me feel good. Just switched to Plum this week, was told "within a year" the programmed drip rates will communicate with EMR

The plum pumps are not bad. Just kinda bulky and capable of only 1 channel/pump for vasoactive drips.

 

[pgg]

The plum pumps are not bad. Just kinda bulky and capable of only 1 channel/pump for vasoactive drips.

I have a lot fewer air-in-line and other bull**** errors with the plum pumps compared to the Alaris. They're not perfect but they are pretty good.

 

[Bertelman]

Yeah, our last year with Alaris has been pretty painful. But seeing those 8 Plum pumps on a double-pole, with 8 cords, etc. is just silly.

Like is this really the best we can do?

 

[GravelRider]

Yeah, our last year with Alaris has been pretty painful. But seeing those 8 Plum pumps on a double-pole, with 8 cords, etc. is just silly.

Like is this really the best we can do?

I used the plum pumps in residency. I think they make a triple pump that only has one cord. They are definitely bulky, but way less prone to error messages and annoying beeping as compared to Alaris.

 

[Bertelman]

Triples are getting phased out. Plum360 baby, because everything that says 360 is better.

I think my eyes rolled 360 deg when I found out we were getting them.

 

[GravelRider]

My favorite pumps will forever be those syringe pumps that you slapped a magnetic faceplate on of whatever drug you were using, you clicked a few dials, and you were up and running without a beep or a hitch. Sometimes new technology is unnecessarily complex.

 

[pgg]

My favorite pumps will forever be those syringe pumps that you slapped a magnetic faceplate on of whatever drug you were using, you clicked a few dials, and you were up and running without a beep or a hitch. Sometimes new technology is unnecessarily complex.

Last time I saw one of those was in my Navy days doing inventory of a shipping container for a field hospital - perfect, simple, bombproof. Was happy to see them since at the regular CONUS hospitals they'd been rounded up and disposed of as "unsafe" or some such bull****. I guess because they didn't have guardrails and confirm buttons or whatever the new idiotproofing hotness is.

 

[abolt18]

Had Alaris pumps in training. Last couple years they were unbearable with their "channel error", "air in line", and channels that just wouldn't even turn on. New job has Plum pumps. Definitely appreciating their simplicity, even if they are huge if you've got several pumps running (haven't run into that yet).

 

[nimbus]

Oxygen delivery using engineered microparticles

The clinical treatment of profound oxygen deprivation (hypoxemia) is time sensitive and requires skill and specialized equipment. When the airways or lungs become incapacitated (e.g., due to airway obstruction or lung injury), resuscitation is ineffective ...

www.ncbi.nlm.nih.gov

https://www.pnas.org/doi/10.1073/pnas.2115276119

 

[Arch Guillotti]

My favorite pumps will forever be those syringe pumps that you slapped a magnetic faceplate on of whatever drug you were using, you clicked a few dials, and you were up and running without a beep or a hitch. Sometimes new technology is unnecessarily complex.

Baxter pump.