Discussion keeps the world turning.
Walk into a growing number of hospital wards in China and you'll find pumps that track their own supplies, beds that monitor movement, and systems that flag risks before anyone else notices.
The technology is impressive, yes, but as smart wards move from pilots to real-world use,
The gap between the hospitals that can afford this and the ones that can't may end up mattering more than the tech itself.
Coming to you live from our studios in Beijing, this is Roundtable.
I'm Steve Hatherly.
Thank you very much for joining us today.
And for the show, I'm with Yushan Mfefe.
First up...
Imagine a hospital ward where an infusion pump alerts nurses when medication is running low, vital signs are automatically recorded, and an AI system that can detect a patient at risk of falling before anyone else notices it's happening.
Smart wards are moving from pilot projects to a growing part of China's health care system.
But as the technology scales, so do the questions, not just about what these systems can do, but about what they actually change on the ground for nurses and doctors and patients and budgets.
For a top-tier hospital in Shanghai, this might not be such a big deal.
It might seem like the next logical step.
But what about a county hospital in the interior part of the country?
That math might look very different.
And that divide may end up shaping the future of Chinese healthcare more.
Than the technology itself.
It is a complicated issue, and we are, of course, going to talk about AI in hospitals and what it means for everyone involved.
Yuxin and Fei-Fei, good afternoon.
But before that...
I want to start here.
Before smart wards became a reality in hospitals, as we do dive right in today, what were some of the bigger challenges facing patients and nurses and doctors in a traditional hospital ward?
This is a question that really worth taking a step back and looking at a much bigger situation because why would we even want to have a smarter wards?
So that's, I think, because the world needs more nurses, but many healthcare systems are already struggling to provide enough nursing care.
And that's actually proved by a lot of stats worldwide.
The stats from the World Health Organization's State of the World's Nursing 2025 report is showing that the global nursing workforce grew from around 28 million in 2018 to 30 million in 2023.
That's, of course, we know, a significant growth.
But the world still had an estimated 5.8 million nurse shortage in 2023.
Growth, but not enough growth.
Yes.
I actually have another set of data coming from actually a study from Peking University.
They published a paper in 2020, and in that paper, they cited a database established by the National Center for Nursing Care Quality Control.
They looked at the staffing of nurses in the Chinese mainland in 2017, and they found, for example, in China,
Large general hospitals in China, a nurse takes care of about eight patients during the day, but 23 patients at night on average.
And it's an average number.
But when it comes to when we are looking at different departments, for example, some departments are handling over 30 patients at night per nurse.
So the point you're making there is that the hospitals are understaffed.
Yeah, especially in nursing.
In terms of nurses, yeah.
So, I mean, what does this look like for nurses?
It could mean when they are understaffed, it could mean that they're making repeated rounds to check patients' health.
Blood pressure and temperature and other vital signs, and then, of course, recording those measurements and updating medical records, all of which would be done, I guess, by hand a long time ago, and then maybe just data entry into a computer not so long ago.
And this would be really time consuming.
Nurses are required to respond to nurse call requests.
I remember staying in the hospital when I was a little kid and having that little button that you hold.
And pressing that button with your thumb and it calls the nurse.
Of course, they have to come and check who needs assistance.
In terms of time, there are some numbers here actually.
Research found that recording a complete set of vital signs,
Took about 215 seconds on paper compared with 150 seconds by an electronic system.
Now you might think, okay, what's the difference?
That's 65 seconds.
That's a minute and five seconds.
Big whoop, right?
Well, multiply that by the number of patients that they have to see, by the number of nurses too.
With the patients they have to see, it's a lot of time and it adds up for sure.
And also for patients, this also could mean waiting for help or for information because they were just, as you said, pushing that button.
And if the nurse was busying on other patients, then it could be a very, very, I think...
Anxious waiting time because that is the moment that you actually need a lot of help.
And if they need assistance, they may have to press that button and wait for a nurse to respond.
And family members might also need to ask staff about the results, medical bills, or treatment arrangement rather than getting that information directly.
So there are just a lot of different types of work that nurses need to do, and still they are in shortage.
Yeah, and there are stats, again, from the WHO.
I don't think we need to go into these, but there are statistics that prove that, that if the nurses can't respond quickly enough, that it's obviously detrimental to the patient's health.
For the medical staff too, workload was a major challenge.
There are studies too that have linked higher nursing workloads and lower staffing levels with negative patient outcomes.
I mentioned falls in the opening.
That's included in some reports.
Medication errors, that's also been reported on.
This is all from PubMed Central, by the way.
And the problem was that doctors and nurses were spending a lot of time on repetitive tasks and had to react to problems as they appeared or after they appeared.
Yeah.
And this is where I think smart when it comes to smart technology being put in place in places like wars can be really, really helpful when it comes to taking over some of the repetitive tasks and also collecting certain datas and also give so that give medical staff more time to focus on the patient themselves.
Well, let's talk about it.
So when a traditional ward turns into a smart ward in a hospital, what does that actually mean?
What does it look like if we compare the two?
A traditional hospital ward is largely reactive.
And nurses make rounds, they check vital signs, respond to call bells, and also record information and deal with problems as they come up.
But a smart ward tries to make that whole process more connected, continuous, and also proactive.
So what would an example of that be?
For example, infusion.
I think we've all been taking some infusion, at least watched people taking infusion before.
Like an IV bottle.
Yes.
And in a traditional setting, the patients or the family members, we need to keep an eye on the IV bottle and then press the nurse call button when the medication is almost run out.
But what happens when you, for example, fell asleep and you lose sort of your track of that time?
I was so scared when I was having that kind of thing on my phone.
Yeah, you need to.
I was scared of, you know, because it is like dripped out of liquid.
Probably it would suck out my blood into the bottle.
Yeah, yeah.
When it's finished and you didn't plug out the needle on time, it will slowly, you can see a strain of blood going back to the tube.
That is very frightening.
Mm-hmm.
But in a smart infusion system, that process can be monitored in real time and the nurse can receive an alert when the attention is needed.
So it's more automatically.
You don't need from the patient's and also the family member's point of view, we don't need to keep an eye on the IV bottle all the time.
Because sometimes that can take more than two hours and the patient is really tired.
They may fall into sleep and don't really have the energy or time to keep an eye on the IV bottle.
But now we have a smart system to help with that.
There's a hospital in Guangzhou in the southern part of the country where the infusion progress for the patients is displayed in real time.
They have a display at the nurse's station and then family members, yeah, like they don't need to call the nurses anymore.
To change an IV bag or the patients don't need to press the button.
The nurses, they see the information in real time right at their nurse's station.
Another example would be vital sign monitoring, maybe even more important than the IV bottle.
Yes, because in traditional wars, the nurses repeatedly will go back to the wars and taking a patient's temperature, blood pressure, and then manually they enter the numbers on a sheet or in their booklets.
But in a smart setting, these connected devices can automatically collect and transmit this data to a sort of hospital information system.
So that means less repetitive paperwork.
And more importantly, doctors and nurses can see changes in a patient's condition more quickly.
So they don't have to work on a chart themselves.
The system is basically doing that for them.
And that, of course, saves a lot of time when you have 100, 200 patients sitting, living in that ward at the same time.
And also to prevent patients from falling.
I think it is also a very important thing for a lot of nurses because I actually experienced, because my mom stayed in hospital for a while before,
And there used to be, there was like almost she hadn't like fallen.
And the nurses was, wow, that actually scared me, Auntie, because if you actually fall, then there's a serious, it is a very serious consequence that probably fall on the whole department.
So I think the hospital, I think all of them, they're trying to prevent these kind of incidents.
And a lot of these smart sensors can monitor patients who may be at risk of falling and automatically send an alert.
How does that work?
How do they get alerted when somebody's about to fall?
Yeah, that was a very interesting question.
I actually checked on that.
I was thinking, how can you predict and prevent a person from falling?
And apparently, they have a system called the millimeter wave radar system that basically monitors
Things like heart rate body position 24 hours a day and also that technology is completely touch free so the patients doesn't need for example a sticky electrode patches or wires connecting to their body but through for example cameras installed in the wars yeah um this radar cameras
They are used in a lot of these, you know, auto driving cars.
Oh, interesting.
Vacuum robots.
They have these sensors to detect the whole environment.
It's the same tech.
Kind of, yes.
Very similar.
So that system can sort of collect and monitor, for example, posture changes of the patients.
And also in the most advanced version of this technology, they also have the ability to make anticipations.
So from monitoring, for example, the steps this patient is taking, the changes of their postures, their heart rate, their blood pressure, they're able to sort of forecast whether this person have the risk of falling if they're moving by themselves.
That's fascinating.
So from everything I've heard so far, the previous system was basically check, record, respond when something happens.
The smart ward improvements are monitor, everything stays connected.
And everyone is alerted either immediately when something happens or even better before something is about to happen.
And if you're wondering, you know, how much of this is happening as we speak, well, IDC forecasts...
That by 2029, 20% of new hospitals or major hospital renovation projects in China will have AI-powered smart wards.
So this is, like I said in the beginning, this is going from the theoretical to the pilots, and now we're seeing it in full practice in real hospitals around the country.
Hmm.
Alright, so now let's talk about how the workload might be reduced.
We talked a little bit about what the smart systems mean.
We gave you a few examples there.
Can it actually reduce the medical staff's workload?
Or I'm wondering if there's a flip side to this coin.
If the system is poorly designed or if the staff is incapable of operating the system, if the tech could actually work against the staff in the sense that if it doesn't fit their clinical routines...
Could it actually create more work for them?
Well, I think we need to answer the first one, actually.
I think we have some reviews of health information technology found that some system, because there are a lot of different systems developed by different tech companies on the market right now,
And some systems reduce the time nurses spend on particular tasks and shift their time toward direct patient care and communication.
For example, barcode medication administration reduced the time spent administering medication by about 33% in the study revealed.
So we are seeing this smart system, smart ward is helping to reduce some workload that mostly the nurses and doctors need to shoulder in the past.
And I think on some level, you know, if you lived in a Chinese outpatient department before, I think you would know that.
A lot of the time the nurses are not happy because they're too busy handling too many patients.
They never stop moving.
Yeah.
So the emotional factor, I think, is also something I want to point out at least is that when the nurses don't need to spend a lot of time on this repetitive work, they can be happier.
So they have more patience.
Communicating with a patient that is also very anxious and happy because of the sickness and pain they're enduring in the world.
And then there's the snowball effect, too, when your nurses are overworked.
I mean, this applies to anybody anywhere, but in this context today, when nurses are overworked,
Um the quality of the work is going to suffer and that's just inevitable there's nothing you can do about that
And also think about the examples that we just listed before um the timers the automated alarms that kind of thing
I don't think these are you know really 20 39 high tech they're just
I think they totally make sense in 2026.
How can we still not have an automated alarm when the IV bottles empty?
Yeah, maybe thinking, why did it take so long to come up with this system anyway?
Yeah, that's what I'm thinking.
So I think a lot of these things are just...
A good upgrade of all of these hospitals and they reduce the workload of all of these repetitive, maybe simple work that nurses need to spend time on.
And then I think that backs to the time add up thing that we discussed before.
Of course, probably for one patient, it's one minute.
But if we add up that for a whole department of patients, for every nurse, then it's probably minutes and hours and days of workload.
There's a hospital in Shenzhen, the Shenzhen Traditional Chinese Medicine Hospital.
It's a top-level hospital in one of the biggest cities in the country.
And it has developed an AI-enabled smart ward that combines the best of everything, really.
5G, IoT, Internet of Things, I mean, big data, AI, everything's AI in there.
And the system does all the things that we talked about.
It collects vital signs and it detects the possible falls and also abnormal physiological readings.
And it supports bedside nursing through mobile devices and connects different hospital systems all together.
And it's made a big difference.
In a pilot, we're talking about time.
The hospital reported that the automated vital sign collection alone saved about 20 to 21 minutes of caregiver time per patient per day.
And automated documentation in a different example saved another 28 to 29 minutes.
Hmm.
That is really significant.
If you can save almost 21 minutes a day per patient only through Vital Sign Collection,
And again, you said it, you shouldn't.
We mentioned it before.
You multiply that, right?
And that time becomes really significant.
Now, what kind of difference would that time actually make to the nursing staff?
Well, it would make a ton of difference.
And one example would
Simply be the workload would feel less overbearing.
I actually did the calculation.
So 20.8 plus 28 plus 0.7, that's 49.5 minutes.
Almost an hour per patient a day.
And the pilot ward had 26 beds.
So across the whole ward, that adds up about 21 hours of nursing time saved every single day.
Wow.
And that's roughly the equivalent of 2.6 full time nurses working time.
And that's important because nurses are not looking after just one patient.
And China's National Health Commission says that in ordinary hospital wards, each nurse should be responsible for no more than eight patients on average.
So the point isn't that a smart system gives one nurse an extra 49 minutes.
It's when you automate small repetitive tasks across dozens of patients, those minutes,
Start to add up to hours of nursing capacity every day, and that time can potentially be redirected toward things technology cannot easily replace, like talk to patients, observe changes in their condition, providing hands-on care and making clinical judgments.
So these are, I think, more, maybe relatively more important stuff that nurses can actually spend time on.
I want to push back just a little bit.
I'm not saying this isn't a wonderful idea.
I'm not saying that the patients can't benefit or the nurses can't benefit.
You mentioned, what was the number?
Almost an hour per patient per day.
We would have to assume that all of the staff knows how to use these AI systems perfectly.
Mm-hmm.
And that can't, at least immediately, that can't be the case.
Because if you're implementing the best tech available to a staff that never used the tech before, there's got to be at least a little bit of a difficulty in the crossover period.
Yeah.
Yeah, especially in this stage, I think we are only starting to deploy such technology really on a large scale throughout the country.
Facing different hospitals.
Some nurses and doctors may have better technical know-how compared to especially, for example, a 40, 50-year-old nurse.
It may be more challenging for them to navigate an AI system.
And also when it comes to, as I mentioned before, the current system are developed by different manufacturers, different companies.
They also have different standards, different communication protocols.
And so they may create, for example, something called data islands, meaning that this set of data is not transmittable to another hospital or wards.
That can also be a problem if, for example, a patient is transmitted through different wards.
So I think it's very natural in the initial stage for smart wars.
I'm reminded of a drama that I saw recently.
I love this show.
It's called The
Pit and are you thinking what i'm thinking probably because it's an what it is is it's an american drama about a hospital and it's about the emergency ward in that hospital
And the show has gotten a lot of praise from its fans because of how realistic, apparently, the show is.
And there was one particular episode, Yushan, maybe this is what you're thinking of.
The electricity episode.
No, I'm not thinking of that episode.
I'm thinking of the one where the new head doctor, the new boss, implemented an AI system that all of the staff needed to follow.
And in the episode, the staff was saying, oh, what is this?
I don't know how to do this.
What's going on?
Now, in the meantime, the emergency room doesn't stop moving.
Right.
There's patients constantly coming and going and requiring doctors and nurses care.
And then so when they do finally have time to spend with the AI system, they're trying to use like the voice recognition thing for medical chart input.
But the system isn't picking up their voice properly.
Yeah.
And it's making mistakes constantly.
So then the staff was complaining and this is my point.
The staff in the episode was complaining, we have to double check this AI's work.
And the boss said, yeah, you have to double check it because AI makes mistakes.
And the staff was like,
Well, then aren't we doing double the work if we have to go back and check it anyway?
So that's kind of the point I'm getting here, too, is that we assume that the staff is going to love this.
But there's going to be a little time needed for the staff to learn how to use it perfectly and to suit it to their needs.
Yeah, actually, I have a set of data.
Among 20 studies conducted across eight countries on AI-enabled workflows in nursing,
Workload reductions in 10 studies, they have like mixed effects in five of the studies.
No clear differences in two studies and increases in two studies and no reported workload outcome in one study.
So nurses are not 100% happy about having this kind of system in their workplace.
And I guess that's one of the reasons for, for example, for one, the hospitals need to work very closely with the tech companies.
So they need to, for example, first figure out what is the actual nursing work routine in this particular hospital.
But then another question pop up in my mind is about what if, for example, for a smaller hospital that don't have that resources?
Well, that's another part of this puzzle, isn't it?
It's another piece of this puzzle.
So let's talk about that now.
You have the Shenzhen top tier hospital that's fully integrated AI this, AI that.
Wonderful.
I don't have the numbers in front of me, but I can only imagine that that...
Hospital in Shenzhen has a significant budget available to them to be able to implement such a system.
What happens when we're talking about grassroots hospitals?
What if we're talking about hospitals in rural areas, rural communities?
They, again, I don't have numbers, but I would have to assume, would not be dealing with the same type of budget.
How are they supposed to implement a system like this?
I think the whole situation right now, it's not only happening in China.
It is also still a global issue.
Yeah, an issue everywhere.
Yeah, sure.
I saw also the WHO stats.
Around 78% of the world's nurses are concentrated in countries that account for only 49% of the global population living there.
Many lower income countries with much fewer nurses relative to their needs.
So it's not only about whether we do we have enough nurses, it's also about where they are.
So, it is an issue that's been spotted by a lot of hospitals here in China, and we're seeing a broader expansion.
First of all, new and renovated wards in top-tier hospitals are increasingly adopting smart technologies as standard.
That's one thing.
And then secondary hospitals in China are also following suit on a larger scale and grassroots hospitals are beginning to test the waters to trying to introduce and include these kind of technologies.
Yes.
For example, in county-level hospitals and township health centers, the rate remains below 12%, with many still limited to basic digital infrastructure because, for example, with budget constraints.
So that would come into, I think, for smaller hospitals, they need to be
In a sense, smarter when it comes to what kind of technology they can use.
For example, they can have a more targeted approach rather than building a very large entire AI-powered hospital like what a top-tier hospital in Shenzhen is doing.
For example, there is a 2026 report described a secondary county hospital in Guangxi.
They upgraded all these walls with a smart nurse call and communication system linked with this hospital information system covering more than 500 beds.
So they're only sort of upgrading one part of their nursing working routine, but not entirely.
I mean, that can work between the constraints coming from the budgets, but also coming from they also want to increase their efficiency as well.
But I think it would be fair to say that the county level hospitals wouldn't require the same budget that the Shenzhen hospital might require, because I'm sure it's probably a smaller facility.
Yes.
Right.
So you wouldn't necessarily need all the bells and whistles, so to speak, that a bigger hospital might need.
Okay, so we've taken a look at some of the obvious advantages, perhaps some of the difficulties in implementing a system like this.
So where are we going next?
Or what would it take for smart wards to become fully implemented everywhere and really as useful as they're meant to be?
Yeah.
First of all, the standard is quite important because- Industry standard.
Yeah.
Right now, different devices and software systems may not communicate with each other as we mentioned before smoothly.
If hospitals have to replace an entire system
Whenever they add a new device or smart wards will be difficult to scale.
And then smart solutions need to become more affordable and flexible to make it to be common in different levels of these hospitals, no matter is provincial or municipal or county level hospitals.
Especially for these smaller hospitals, I think they even needed more.
And a county hospital may be more practical to start with technologies that address its biggest needs, such as automatic vital sign monitoring and smart nurse call systems or fall detection, and then expand gradually.
And also, I think the whole thing about this technology, it's about helping nurses to reduce their workload, but it's not like get nurses less trained.
That show that you were mentioning, Steve, that also reminded me of one incident that they had a whole blackout of their technology, technological system.
Yeah, all the computer systems crashed.
Yes.
And then they had to rely on writing on the boards about all the information of the patients.
And that really requires the techniques, the skills of every single staff in that hospital.
So I think the whole thing, it's about upgrading the experience of the patients and the nurses and doctors.
I think that's a great point, actually.
Yeah, it's not a zero sum game.
They still have all the skills that they had before.
It's just they're learning something new.
And yes, my criticism before was not really a criticism because anytime a new system is implemented, it's going to take some time for people to study it and learn it and know how to use it perfectly.
That's a natural thing.
But eventually, they'll know both, and they can make the best use of both worlds.
And they'll figure that out, right?
Where do we use AI to its best?
Where do we use the old system to its best?
And they'll figure out how to do it.
Yeah, I think people, no matter as being the nurses or the doctors or the patients or the family members of that patients will always be central to any of this technological transformation taking place.
I think, as Steve said, it's not a zero-sum game.
It's not saying we have a vital sign monitoring system that means the nurses no longer need to learn how to collect blood pressure, the significance of body temperatures on a daily basis.
It still needs something.
They still need training behind that.
And I think they still need to learn how to collect these vital figures on themselves.
It's just making things easy, making them happier is the basic goals.
