WEBVTT
Kind: captions
Language: en

00:00:14.000 --> 00:00:18.400
AI in healthcare are mostly focusing on imaging,

00:00:18.400 --> 00:00:24.580
because the recent wave of AI development actually

00:00:24.580 --> 00:00:27.390
kind of thanks to a breakthrough

00:00:27.390 --> 00:00:30.800
in the imaging recognition like CNN,

00:00:30.800 --> 00:00:33.460
new algorithms like CNN.

00:00:33.460 --> 00:00:40.640
That's why most people actually focus their AI effort on
medical image processing,

00:00:40.640 --> 00:00:43.880
but again, I would like to emphasize that

00:00:43.880 --> 00:00:49.160
if we use AI in quality and patient safety,

00:00:49.220 --> 00:00:54.780
you know, it might be easier to get
benefit for the patient and faster as well.

00:00:54.780 --> 00:00:59.100
Especially in an environment like
Taiwan, we have a National Health Insurance,

00:00:59.100 --> 00:01:08.440
so every improvement, every tiny little improvement will be amplified by 23 million,

00:01:08.440 --> 00:01:10.500
that's how many patients we have.

00:01:10.500 --> 00:01:12.380
And it's all one database.

00:01:12.380 --> 00:01:19.740
So that's why I think it's about time we have a conference on AI QPS,

00:01:19.900 --> 00:01:22.300
AI in quality and patient safety.

00:01:22.300 --> 00:01:24.520
Okay, before we have our keynote speaker,

00:01:24.520 --> 00:01:28.440
I actually would like to show you a few
slides of one example

00:01:28.440 --> 00:01:33.760
that we dedicate the AI effort into patient safety,

00:01:33.760 --> 00:01:37.620
and this is for improved the medication safety.

00:01:37.620 --> 00:01:41.420
I would like to just spend a few minutes.

00:01:41.420 --> 00:01:45.340
So as you know that medical error 
has been a major problem.

00:01:45.340 --> 00:01:48.920
It was in the year 2000,

00:01:48.920 --> 00:01:56.350
the medical error still ranked number five in the top 5 causes of death

00:01:56.350 --> 00:01:59.680
in the year 2000, every year

00:01:59.680 --> 00:02:04.460
medical error kills about, these are 
deaths, these are not injury,

00:02:04.460 --> 00:02:05.880
those people are dead

00:02:05.880 --> 00:02:12.040
okay, so about 100,000 people dead for
medical error, due to medical error,

00:02:12.040 --> 00:02:20.360
and in 2013, after about 13 years,
this one is published in 2016,

00:02:20.360 --> 00:02:23.140
but the numbers are from 2013.

00:02:23.140 --> 00:02:26.660
So after 13 years of effort, you know, everybody is working on it,

00:02:26.660 --> 00:02:29.480
and feel likes it's a big problem,

00:02:29.480 --> 00:02:33.380
we managed to make medical error,

00:02:33.380 --> 00:02:34.040
you know,

00:02:34.040 --> 00:02:39.240
number three in the top five causes of death.

00:02:39.240 --> 00:02:41.140
If we work even harder,

00:02:41.140 --> 00:02:44.319
it will actually, you know, replace cancer and

00:02:44.319 --> 00:02:46.849
heart disease, and become number one, you
know.

00:02:46.849 --> 00:02:48.670
So that's really a big problem,

00:02:48.670 --> 00:02:53.480
and of course, it's not that we're not
doing enough,

00:02:53.480 --> 00:02:56.440
well actually, we probably are not doing
enough, but uh,

00:02:56.440 --> 00:03:00.040
but again, the Asian population

00:03:00.040 --> 00:03:06.000
make people, you know, uh, it make
the poly-morbidity issue, you know,

00:03:06.000 --> 00:03:08.740
there are more diseases in one patient.

00:03:08.740 --> 00:03:12.160
It's a doctor I've been seeing patient for

00:03:12.160 --> 00:03:17.320
twenty years, and you know, many of my
patients ages with me.

00:03:17.320 --> 00:03:19.000
Okay so when they

00:03:19.000 --> 00:03:21.560
started to see me, they were like sixty in
their 60s,

00:03:21.560 --> 00:03:23.520
now they're in their 80s.

00:03:23.520 --> 00:03:28.560
But they're still kind of relatively healthy,
but again, they have diabetes,

00:03:28.560 --> 00:03:31.470
hypertension, you know, poly-morbidity.

00:03:31.470 --> 00:03:36.629
So me many patients suffer from multiple diseases.

00:03:36.629 --> 00:03:39.040
In poly-pharmacy, because of 
the multiple diseases,

00:03:39.040 --> 00:03:41.180
they need to take a lot of medicine,

00:03:41.180 --> 00:03:43.960
you know, the maximum,

00:03:43.960 --> 00:03:47.400
um, number of drugs that you can 
prescribe to one person.

00:03:47.400 --> 00:03:50.319
Do you know there's a limit for that?

00:03:50.319 --> 00:03:51.400
In our hospital,

00:03:51.400 --> 00:03:55.220
you know,
in many of our hospital, that number was 21,

00:03:55.220 --> 00:03:59.300
okay so in a hospital.

00:03:59.300 --> 00:04:01.200
But there's no limit if you do it,

00:04:01.200 --> 00:04:04.400
if you take your drug from several different
hospitals, right?

00:04:04.400 --> 00:04:07.620
So, um, poly-pharmacy, that's a big problem,

00:04:07.620 --> 00:04:11.620
and more patient visit means
shorter physician time,

00:04:11.620 --> 00:04:16.960
and there are new drugs, new exams, and new procedures.

00:04:16.960 --> 00:04:23.460
Some people estimate that the half-life of medical knowledge is 70 months.

00:04:23.460 --> 00:04:27.140
Okay, if that's the case, that means

00:04:27.140 --> 00:04:32.660
one and a half year after you
graduate from medical school,

00:04:32.660 --> 00:04:36.260
half of your knowledge has already outdated, right?

00:04:36.260 --> 00:04:38.920
So it's really a lot of new things that we
have to learn,

00:04:38.920 --> 00:04:42.000
not to mention genomic factors, environmental factors,

00:04:42.000 --> 00:04:44.040
and other factors are increasing

00:04:44.040 --> 00:04:46.210
the complexity of these things.