Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Trekkin]

Automation?  In nursing? (stifles mad laughter)

I'm just gonna quote this for posterity. Everybody thinks their job is too special to automate until suddenly it's not. Nursing in particular is an attractive target because of how standardized it is; it's easier to come up with a single evaluatable set of capabilities to build into a robot RN with calculable value to the industry than it would be to do so for something like a lab tech. That in turn helps get projects funded.

More to the point, though, they don't need to replace 100% of RNs. It's still profitable to replace just the repetitive or tedious or precise parts, firing nurses all the while, until the job of the one RN in the hospital is to sit and be shouted at by angry sick people who are mad at their robots. Then the pool of people who can do it widens and salaries drop.

[wierd]

Quite, but malpractice rates also go up, because the nursing staff lacks the appropriate skillsets to evaluate when the robots are malfunctioning. If you haven't seen what "being insufficiently gentle" when performing cath-care on a man does to a man's penis, You really should.  (and that's just cath-care. There's a whole host of other, "You can seriously fuck up a patient permanently if you do this wrong, so fucking do it right the first time, every time goddamit!" activities that happens in nursing.) The malpractice suits would be outrageous. Possibly enough to completely negate the cost savings from using the robots.

There's also the increased risk factors of sanitization of a robot that has more contaminatable surfaces. While you could theoretically autoclave a robot, that's not always sufficient.


I am not saying that robot nurses are impossible (ever), just that they are currently not realistically feasible in the foreseeable future.
(even then, it is quite likely that humans would strongly prefer to be cared for by an actual human, in much the same way that humans much prefer to be cared for by actual humans in telephone support settings. The difference is that in this new setting, the robot on the line can literally fuck you up.)

[Trekkin]

And I'm saying don't look for robot nurses. Look for increasingly sophisticated labor-saving devices for nurses, because if you can save X% of nurse time you can fire X% of nurses, and repeat that process ad nauseam until the parts of the job that are left don't look anything like the original job.

[wierd]

You see that now with doctors, and things like robot surgeons.

However, robot surgeons are not magic bullets, and there are malpractice claims concerning their use.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5687202/


The number of things a nurse does, (and the frequency!) are in a different league.  Doctors tend to do very sophisticated procedures, requiring very detailed specialist knowledge; Nurses do routine after-care, which is much more frequent. Often many times a day.

Providing tools in the nurse population increases the risks demonstrated in robotic surgery many times over.  There's a reason why doctors want nurses to use a manual stethoscope and sphygmomanometer when taking blood pressures, rather than an automated cuff, for instance.   

[Il Palazzo]

Can you think of any part of your typical workload that could be automated with current or near-current technology, with reasonably little risk? If yes, then what percentage (time-wise) of your workload would such automation take over?

[wierd]

Cleaning rooms and delivering room trays.

Possibly cleaning soiled beds (with patient removed.)


Anything related to actual hands-on care on a patient? No. Not safely.

[Il Palazzo]

So, what percentage of your typical work day do these activities occupy? (because that's approximately the percentage of nursing staff that can be made redundant)

Also, isn't there a job that normally requires more than one nurse, but could be done just as well by at least one fewer, if equipped with a technologically-feasible device?

[wierd]

Depends on the setting;  In long term care, quite a lot. (People are very incontinent all the time, but the human in attendance of the robot still has to clean and care for the patient, while the robot takes care of trash and linens. Most of the time spent (more than 70%, upward of 90 if the patient is uncomfortable and needs reassurance) is with hands-on patient care for those doing grunt work. Patients with dementia may be obstinate about refusing care, but still most definitely require it; the care giver has to convince such patients to submit to receiving the care, and that process can be quite lengthy. This is why you would need a strong general AI straight of of science fiction to replace nursing staff. (Office and desk nurses could probably be completely automated, if not for the requirement of needing quality communication with doctors. See also "Call center robot hell, hospital version" and now people die from it.) 

"Nursing staff" covers CNA work (which mostly focuses on pericare, dealing with belligerent patients, and the gruntwork of getting room trays and delivering specimens. The latter two being ad-hoc errands that are infrequent but annoying, the former being the mainstay of occupation) and RN work (Foley insertions, wound dressing and care, setup and operation of IV units, Skin and wound assessments, detailed medical documentation collection and communication, and management of CNA workers).  Since there is a shortage of CNAs, RNs often do CNA duties too. (It is not uncommon to see an RN cleaning up puke and shit.)

Robotic assistants could cover a portion of CNA work, but very little of RN work.   


For acute trauma care, even less. RNs do much more because there is more that needs their higher credentials.  Working a trauma ward has stricter requirements, and is harder on the mind.

For psychiatric facility care, nurses tend to need CNAs more, and "human interaction" is needed more by the patients. (A robot is likely to be destroyed by the patient in a fit of rage, since the patient is likely to be literally psychotic in this setting.)


For something like "assisted living", where the patient is rarely incontinent, and just needs help remembering to take their pills, put on clean clothes, and eat regularly-- Robots could probably do a whole lot.  Assisted living is a vanishingly small demographic though.  This is the demographic that the Japanese are targeting.

[Il Palazzo]

Well, that does not sound like a fully robot-proof occupation, but still pretty safe, relatively speaking. At least in your lifetime.

In my field it looks more like some 90% jobs going redundant, in a decade or two.

[Trekkin]

Well, that does not sound like a fully robot-proof occupation, but still pretty safe, relatively speaking. At least in your lifetime.

In my field it looks more like some 90% jobs going redundant, in a decade or two.

It sounds to me like it's the sort of occupation where the more fundamental obstacles to automation that aren't good candidates for industrial funding can be discretized into federally fundable projects, and the nursing shortage is a clear motivating factor. I'd be surprised if it had a decade or two.

I'd say my field was resistant to automation, but it's more that we keep building the robots that can take our jobs and then doing more science with them, so I suppose it's more automation-compatible.

[dragdeler]

-

[wierd]

Well, that does not sound like a fully robot-proof occupation, but still pretty safe, relatively speaking. At least in your lifetime.

In my field it looks more like some 90% jobs going redundant, in a decade or two.

It sounds to me like it's the sort of occupation where the more fundamental obstacles to automation that aren't good candidates for industrial funding can be discretized into federally fundable projects, and the nursing shortage is a clear motivating factor. I'd be surprised if it had a decade or two.

I'd say my field was resistant to automation, but it's more that we keep building the robots that can take our jobs and then doing more science with them, so I suppose it's more automation-compatible.

I think you forget just how much the GOP hates funding science, Trekkin.   They often  view AI as "Pie in the sky" science fiction, rather than actually needed research.  They also tend to focus on military applications rather than civilian ones, which they view as the domain of private enterprise. (often religiously so.)  They are much more likely to fund skynet than baymax.

(Unless of course, it's a military combat medic robot, and is exempted from a great deal of medical ethics considerations, because the military is not nearly as culpable for malpractice, and consent for procedures is not nearly as enforced. But then you end up with a scary robot being pushed into civilian hospitals, and then you get all the malpractice claims.)

[dragdeler]

-

[Egan_BW]

We'll make a combat medic robot, then realize that it has a patient mortality rate of 100% and make it a combat robot.

[Trekkin]

I think you forget just how much the GOP hates funding science, Trekkin.   They often  view AI as "Pie in the sky" science fiction, rather than actually needed research.  They also tend to focus on military applications rather than civilian ones, which they view as the domain of private enterprise. (often religiously so.)  They are much more likely to fund skynet than baymax.

Yeah, they hate science so much they gave the NIH alone a 5% budget increase in 2018, the fourth consecutive increase in a row.