Revenue Anxiety: The Invisible Voice Shaping Clinical Decisions

WEBVTT

00:00:01.066 --> 00:00:01.427
All right, well,

00:00:01.447 --> 00:00:02.447
welcome back to The Operator.

00:00:02.548 --> 00:00:03.289
Last week,

00:00:03.809 --> 00:00:05.931
we talked about AI giving clinicians back

00:00:06.530 --> 00:00:07.251
their presence.

00:00:07.432 --> 00:00:11.294
This week asks an even more uncomfortable

00:00:11.335 --> 00:00:11.814
question.

00:00:12.236 --> 00:00:14.256
Who's actually making the clinical

00:00:14.297 --> 00:00:15.438
decisions in the room?

00:00:15.698 --> 00:00:17.399
Because if you've practiced long enough,

00:00:17.940 --> 00:00:19.460
you know there's another voice in the

00:00:19.501 --> 00:00:20.661
room, not the patients,

00:00:21.321 --> 00:00:22.283
but maybe the payers.

00:00:22.303 --> 00:00:23.184
It's somewhere in there.

00:00:23.204 --> 00:00:23.984
You don't want it to be.

00:00:24.605 --> 00:00:26.687
And if that's influencing what you

00:00:26.746 --> 00:00:28.807
recommend before you've even finished your

00:00:28.827 --> 00:00:29.588
evaluation,

00:00:30.463 --> 00:00:33.026
Larry argues we've got a much bigger

00:00:33.085 --> 00:00:34.968
problem than AI.

00:00:35.008 --> 00:00:35.987
So let's start there.

00:00:36.529 --> 00:00:38.850
You really begin this article in a series

00:00:38.890 --> 00:00:41.773
by saying there's another voice in the

00:00:41.813 --> 00:00:43.173
treatment room, not the patients,

00:00:43.213 --> 00:00:44.755
but the payers.

00:00:44.975 --> 00:00:47.296
When did you first realize that that was

00:00:47.357 --> 00:00:47.658
true?

00:00:48.472 --> 00:00:50.456
Yeah, I would say actually, you know,

00:00:50.555 --> 00:00:53.399
years before I even sort of had words

00:00:53.439 --> 00:00:54.502
or could articulate it.

00:00:54.661 --> 00:00:55.584
You know,

00:00:55.723 --> 00:00:57.566
you feel something oftentimes before you

00:00:57.585 --> 00:00:58.527
can explain it, right?

00:00:58.588 --> 00:00:59.048
Yeah.

00:00:59.088 --> 00:00:59.990
So you're the patient.

00:01:01.356 --> 00:01:03.137
they're telling you where it hurts,

00:01:03.298 --> 00:01:04.858
what they can't do anymore,

00:01:05.338 --> 00:01:06.099
what they're scared of.

00:01:06.140 --> 00:01:07.540
And somewhere in the back of your head,

00:01:07.561 --> 00:01:09.742
there's a second kind of conversation or

00:01:09.802 --> 00:01:10.843
noise going on.

00:01:11.164 --> 00:01:12.025
Is that covered?

00:01:12.965 --> 00:01:14.387
Will I have the time to do that?

00:01:14.427 --> 00:01:15.828
Will that take up too many of my

00:01:15.867 --> 00:01:17.468
minutes and interfere with the next

00:01:17.509 --> 00:01:18.870
patient who's already probably in the

00:01:18.909 --> 00:01:19.391
waiting room?

00:01:19.411 --> 00:01:21.451
Do I need a pre-auth on it?

00:01:21.471 --> 00:01:22.412
Did I document it?

00:01:22.513 --> 00:01:24.034
Do I have functional goals?

00:01:24.073 --> 00:01:25.454
Do I have all these kind of things

00:01:25.474 --> 00:01:26.816
that are going in your head that we've

00:01:26.855 --> 00:01:27.137
kind of

00:01:27.757 --> 00:01:30.581
placed onto medical providers.

00:01:31.941 --> 00:01:33.784
That's the payer uninvited.

00:01:33.944 --> 00:01:36.628
That's the unintended consequences of us

00:01:37.028 --> 00:01:39.290
having payers rule the room.

00:01:39.350 --> 00:01:40.813
They're sitting there with us.

00:01:41.134 --> 00:01:43.417
And here's the part that should bother all

00:01:43.456 --> 00:01:43.716
of us.

00:01:43.816 --> 00:01:45.659
Nobody put that chair in there on purpose.

00:01:46.319 --> 00:01:48.242
you know it got there added one

00:01:48.281 --> 00:01:50.704
requirement at a time just incremental

00:01:50.765 --> 00:01:52.387
over a third year period until it felt

00:01:52.406 --> 00:01:54.929
like furniture in your office right and

00:01:54.948 --> 00:01:56.430
the day that i knew was real was

00:01:56.450 --> 00:01:59.054
the day i watched great clinicians you

00:01:59.073 --> 00:02:01.275
know either shorten a conversation or be

00:02:01.376 --> 00:02:03.177
hurried or do something they wouldn't have

00:02:03.317 --> 00:02:04.299
normally done

00:02:05.099 --> 00:02:06.740
interrupting the patient, for example,

00:02:08.181 --> 00:02:09.901
because they were too busy thinking of

00:02:09.961 --> 00:02:12.081
minutes and coverage and productivity and

00:02:12.282 --> 00:02:13.663
units and all those things.

00:02:13.923 --> 00:02:14.103
You know,

00:02:14.163 --> 00:02:15.543
no one decided to put the insurance

00:02:15.562 --> 00:02:17.683
company in the treatment room, Jimmy.

00:02:17.783 --> 00:02:20.905
It got there the way most things happen,

00:02:21.405 --> 00:02:22.626
one at a time over a long period

00:02:22.645 --> 00:02:24.105
of time until it starts now to look

00:02:24.165 --> 00:02:24.526
normal.

00:02:25.086 --> 00:02:27.306
And those little things contribute

00:02:28.168 --> 00:02:29.207
incrementally to burnout.

00:02:31.296 --> 00:02:32.979
You introduced a phrase that I actually

00:02:33.019 --> 00:02:34.219
hadn't heard before,

00:02:34.259 --> 00:02:35.480
but like you said a second ago,

00:02:35.540 --> 00:02:37.342
sometimes we feel it before we can put

00:02:37.402 --> 00:02:40.425
words to it, which is revenue anxiety.

00:02:40.825 --> 00:02:43.268
It sounded like you were describing

00:02:43.328 --> 00:02:45.991
something that happens almost beneath the

00:02:46.371 --> 00:02:47.792
clinician's awareness until you can

00:02:47.853 --> 00:02:48.652
actually verbalize it.

00:02:49.354 --> 00:02:50.775
How would you describe what revenue

00:02:50.854 --> 00:02:52.417
anxiety feels like?

00:02:53.103 --> 00:02:54.063
Yeah, it's a hum.

00:02:54.123 --> 00:02:54.724
It's a very,

00:02:54.764 --> 00:02:57.247
very low grade noise or a hum,

00:02:57.429 --> 00:02:59.050
and it never shuts off.

00:02:59.110 --> 00:03:02.474
And it's that underlying issue.

00:03:02.495 --> 00:03:04.176
That's why I call it anxiety.

00:03:05.366 --> 00:03:07.888
that is going on while you're trying to

00:03:07.927 --> 00:03:09.788
be present for a patient right it's not

00:03:09.828 --> 00:03:12.088
the panic of the day it's the background

00:03:12.149 --> 00:03:13.628
dread that you do the right thing

00:03:13.649 --> 00:03:17.509
clinically um but it won't get paid right

00:03:17.530 --> 00:03:18.871
that's that fear if i do it right

00:03:18.911 --> 00:03:20.330
it won't get paid that you'll finish a

00:03:20.390 --> 00:03:22.231
note at seven p.m and it's gonna get

00:03:22.252 --> 00:03:24.913
denied anyhow or that you have to have

00:03:24.932 --> 00:03:27.193
your note reviewed by a peer reviewer or

00:03:27.233 --> 00:03:29.153
that something's gonna come back and as an

00:03:29.334 --> 00:03:29.593
edit

00:03:31.348 --> 00:03:32.829
And the crude part about all of this

00:03:32.889 --> 00:03:34.751
is the therapist can't tell you that it's

00:03:34.790 --> 00:03:35.151
there.

00:03:35.450 --> 00:03:37.272
Ask them and they'll say, I'm fine.

00:03:38.174 --> 00:03:39.474
I love my patients.

00:03:39.574 --> 00:03:40.574
I do patient care.

00:03:40.975 --> 00:03:42.437
But that anxiety is there.

00:03:42.877 --> 00:03:45.079
It's not about the patients and their

00:03:45.118 --> 00:03:45.680
problems.

00:03:46.280 --> 00:03:48.162
It's about the machinery wrapped around

00:03:48.181 --> 00:03:48.822
the patients.

00:03:48.861 --> 00:03:50.502
It's all this,

00:03:51.084 --> 00:03:53.104
the mental tax that goes with

00:03:53.866 --> 00:03:55.707
all that you have about your boss,

00:03:55.768 --> 00:03:56.829
about your company,

00:03:56.908 --> 00:03:58.629
about the insurance company,

00:03:58.949 --> 00:04:00.091
about your peers,

00:04:00.191 --> 00:04:02.112
all those kind of things.

00:04:02.193 --> 00:04:04.414
And the burnout research is interesting

00:04:04.435 --> 00:04:07.396
because it's sort of very clear on this.

00:04:07.456 --> 00:04:09.718
People don't leave the work because it's

00:04:09.778 --> 00:04:11.080
difficult work.

00:04:11.199 --> 00:04:13.521
They go into it for patient care, right?

00:04:13.762 --> 00:04:15.463
They leave because something keeps them

00:04:15.503 --> 00:04:16.584
from doing the work.

00:04:17.485 --> 00:04:20.687
Revenue anxiety is one of those things.

00:04:21.367 --> 00:04:23.848
It doesn't feel like fear.

00:04:23.887 --> 00:04:26.608
It feels like static, noise, hum,

00:04:26.629 --> 00:04:28.749
and you can't do great work over static.

00:04:29.788 --> 00:04:31.310
You can't hear yourself think.

00:04:32.050 --> 00:04:33.709
It's like the good cop,

00:04:33.750 --> 00:04:36.591
bad cop going on in your brain.

00:04:37.531 --> 00:04:40.711
And when you talk to therapists on a

00:04:40.932 --> 00:04:43.033
one-to-one level or medical providers,

00:04:43.213 --> 00:04:44.173
they tell you this, yeah,

00:04:44.192 --> 00:04:45.153
I'm thinking about this,

00:04:45.194 --> 00:04:46.494
but this is going on and

00:04:47.581 --> 00:04:48.382
And, you know,

00:04:48.401 --> 00:04:50.423
we get all this multitasking going on and

00:04:50.442 --> 00:04:52.124
then they got social media in their brain

00:04:52.163 --> 00:04:53.785
and they got text messages and other

00:04:53.845 --> 00:04:55.125
messages coming at them.

00:04:55.826 --> 00:04:57.067
All of that collectively,

00:04:57.127 --> 00:04:58.869
I'm sort of calling revenue anxiety or

00:04:58.908 --> 00:05:00.569
these underlying things that are

00:05:00.610 --> 00:05:03.391
distracting from what really drove you

00:05:03.411 --> 00:05:05.954
into what you love, which is patient care.

00:05:06.014 --> 00:05:06.533
Yeah.

00:05:06.593 --> 00:05:07.915
Well, we would never let a parent.

00:05:08.593 --> 00:05:10.435
into the room and actively say, hey,

00:05:11.076 --> 00:05:12.036
I'd like you to make the clinical

00:05:12.076 --> 00:05:12.497
decision.

00:05:12.896 --> 00:05:14.298
But there was a sentence in your article

00:05:14.338 --> 00:05:16.680
this week where training clinicians to let

00:05:16.721 --> 00:05:19.124
the payer make the clinical decision,

00:05:19.184 --> 00:05:20.345
almost like a Trojan horse.

00:05:20.404 --> 00:05:21.826
They didn't show up and knock on the

00:05:21.846 --> 00:05:22.307
door and say,

00:05:22.326 --> 00:05:23.348
I wanted to make that decision.

00:05:23.968 --> 00:05:25.009
We sort of let them in.

00:05:25.430 --> 00:05:26.870
That's a serious accusation.

00:05:26.891 --> 00:05:29.233
But so how do we end up there?

00:05:30.040 --> 00:05:30.360
Yeah, well,

00:05:30.379 --> 00:05:32.081
I love your analogy of the Trojan horse,

00:05:32.100 --> 00:05:32.581
by the way.

00:05:32.620 --> 00:05:33.781
But we got there slowly,

00:05:33.821 --> 00:05:36.442
which is one of the serious things.

00:05:36.482 --> 00:05:38.083
Why did we ever let this happen, right?

00:05:38.142 --> 00:05:39.783
Nobody ever stood up and announced.

00:05:39.843 --> 00:05:40.423
From now on,

00:05:40.483 --> 00:05:42.463
insurance company decides what the plan of

00:05:42.504 --> 00:05:43.105
care is.

00:05:43.165 --> 00:05:43.524
Instead,

00:05:43.644 --> 00:05:45.706
we built a system where the payer decides

00:05:45.925 --> 00:05:46.985
what gets authorized,

00:05:47.386 --> 00:05:48.425
what gets reimbursed,

00:05:48.485 --> 00:05:50.447
and how many visits are allowed.

00:05:50.487 --> 00:05:52.127
And then we act surprised when the

00:05:52.588 --> 00:05:55.728
clinical plan quietly sort of reshapes

00:05:55.769 --> 00:05:58.009
itself to these three things.

00:05:58.610 --> 00:05:59.370
It's interesting,

00:05:59.810 --> 00:06:02.312
I spoke at a dental conference and there's

00:06:02.333 --> 00:06:04.055
a difference here because in dental,

00:06:04.694 --> 00:06:06.377
a good portion of it is out of

00:06:06.416 --> 00:06:09.360
pocket pay and dental insurance isn't

00:06:09.379 --> 00:06:11.540
really underwriting insurance like it is,

00:06:12.281 --> 00:06:14.564
but it's the exact same thing happened,

00:06:14.624 --> 00:06:17.086
different problems, same identical trap.

00:06:17.125 --> 00:06:19.928
And here's how I put it to them.

00:06:20.028 --> 00:06:20.269
I said,

00:06:20.329 --> 00:06:22.290
here's the piece that almost nobody in the

00:06:22.331 --> 00:06:24.353
building got in their training.

00:06:25.216 --> 00:06:26.495
When we train a dentist,

00:06:26.555 --> 00:06:27.637
and you could say a therapist,

00:06:27.677 --> 00:06:28.697
you could say a doctor,

00:06:28.916 --> 00:06:31.317
to be clinically excellent and then never

00:06:31.398 --> 00:06:33.737
teach them to separate their clinical

00:06:33.798 --> 00:06:35.939
judgment from an insurance company fee

00:06:35.978 --> 00:06:36.499
schedule,

00:06:36.879 --> 00:06:39.779
we're training a clinician to let the

00:06:39.839 --> 00:06:42.120
payer make clinical decisions.

00:06:42.701 --> 00:06:44.180
And think about what that means, Jimmy.

00:06:44.201 --> 00:06:45.862
You've invested in building a highly

00:06:45.942 --> 00:06:47.122
skilled practitioner.

00:06:47.562 --> 00:06:48.822
They sit down with a patient,

00:06:48.841 --> 00:06:51.043
they see clearly what the patient needs,

00:06:52.278 --> 00:06:53.220
And then what happens?

00:06:53.259 --> 00:06:54.901
They open up the benefit portal,

00:06:54.942 --> 00:06:55.901
the insurance portal,

00:06:55.922 --> 00:06:58.264
and suddenly the treatment plan bends.

00:06:58.865 --> 00:07:01.026
That's not clinical autonomy.

00:07:01.047 --> 00:07:03.149
That's sophisticated form of learned

00:07:03.209 --> 00:07:03.949
helplessness.

00:07:04.930 --> 00:07:06.112
It's just dressed up.

00:07:06.331 --> 00:07:07.833
It's just dressed up in a clinical

00:07:07.872 --> 00:07:08.334
language.

00:07:08.353 --> 00:07:10.435
In physical therapy, dentistry,

00:07:10.495 --> 00:07:12.458
pick your white coat, same mechanism.

00:07:13.077 --> 00:07:14.639
You know, clinicians thinking, well,

00:07:14.879 --> 00:07:16.540
I can't do that because it's not approved.

00:07:16.579 --> 00:07:17.560
I like doing this dry kneeling,

00:07:17.579 --> 00:07:18.279
but it's not approved.

00:07:18.300 --> 00:07:19.581
Oh, if I do manual therapy,

00:07:19.620 --> 00:07:22.021
is it too many units now that I'm

00:07:22.062 --> 00:07:22.622
on the patient?

00:07:23.141 --> 00:07:24.802
All of that is going by.

00:07:24.822 --> 00:07:27.184
You don't lose clinical judgment in a

00:07:27.264 --> 00:07:28.125
dramatic moment.

00:07:28.165 --> 00:07:29.504
You lose it the day you think,

00:07:30.125 --> 00:07:31.185
will it get approved?

00:07:31.865 --> 00:07:33.367
Think about that.

00:07:33.406 --> 00:07:34.947
That's not clinical judgment.

00:07:34.968 --> 00:07:36.648
It's the opposite of clinical judgment.

00:07:36.708 --> 00:07:38.389
And if that is clouding your mind and

00:07:38.449 --> 00:07:40.149
we've let it cloud minds,

00:07:40.649 --> 00:07:41.750
that is a problem.

00:07:42.031 --> 00:07:42.230
Yeah.

00:07:42.913 --> 00:07:45.394
I think if we were to do control

00:07:45.714 --> 00:07:47.415
F and see which word would pop up

00:07:47.435 --> 00:07:48.196
the most in this,

00:07:48.216 --> 00:07:50.156
it wouldn't be AI or even insurance.

00:07:50.617 --> 00:07:51.877
I think the word we've said so far

00:07:51.937 --> 00:07:53.038
most is judgment.

00:07:53.678 --> 00:07:54.759
So I'm struck by the fact that this

00:07:54.798 --> 00:07:56.100
article really wasn't about insurance.

00:07:56.139 --> 00:07:57.920
It is about that word judgment.

00:07:58.641 --> 00:08:01.182
Would you say what you're actually – is

00:08:01.202 --> 00:08:02.483
that what you're actually writing about?

00:08:03.144 --> 00:08:03.384
No.

00:08:03.543 --> 00:08:04.865
Yeah, a hundred percent.

00:08:06.115 --> 00:08:08.596
insurance is the villain of the story,

00:08:08.637 --> 00:08:10.658
but judgment is the subject.

00:08:10.677 --> 00:08:11.358
Yeah.

00:08:11.439 --> 00:08:14.060
I don't actually care about billing codes

00:08:14.079 --> 00:08:15.360
for the sake of billing codes.

00:08:15.380 --> 00:08:17.622
What I care about is getting a highly

00:08:17.682 --> 00:08:18.682
trained clinician,

00:08:18.723 --> 00:08:20.685
somebody with a doctorate or a dentist or

00:08:20.704 --> 00:08:21.665
a primary care doctor,

00:08:21.704 --> 00:08:24.086
who's trained for years to be an

00:08:24.127 --> 00:08:26.028
interpreter of the body or interpreter of

00:08:26.048 --> 00:08:26.569
the mouth.

00:08:26.608 --> 00:08:27.689
And they build a plan.

00:08:28.149 --> 00:08:30.071
And then all of a sudden it's being

00:08:30.110 --> 00:08:32.011
reduced to a machine and whether that

00:08:32.052 --> 00:08:33.232
machine says yes or no,

00:08:33.293 --> 00:08:33.994
I can do that.

00:08:34.134 --> 00:08:34.374
Right.

00:08:34.653 --> 00:08:36.395
and so the whole series is one argument

00:08:36.917 --> 00:08:39.679
to protect the clinical judgments you know

00:08:39.740 --> 00:08:42.042
actually sir you know last week's article

00:08:42.062 --> 00:08:45.405
was was really more about being present

00:08:45.446 --> 00:08:48.429
this one is protecting clinical judgment

00:08:48.909 --> 00:08:51.994
you know every one of these ai agents

00:08:52.053 --> 00:08:53.655
and that's where i'm going with this and

00:08:54.495 --> 00:08:56.417
if if ai is to do one thing

00:08:56.456 --> 00:08:59.057
correct is to allow us to do what

00:08:59.116 --> 00:09:01.597
it can't which is be human and so

00:09:01.658 --> 00:09:04.719
week one agent is the scribe week two

00:09:04.918 --> 00:09:08.740
agent is the agent that allows it to

00:09:08.860 --> 00:09:11.400
read all the things around insurance and

00:09:11.500 --> 00:09:13.922
eight minutes and protect it and pre-auth

00:09:13.981 --> 00:09:16.482
and do that in the background so that

00:09:16.523 --> 00:09:18.663
you can focus on doing things clinically

00:09:19.263 --> 00:09:21.083
the scribe allows you to be present

00:09:21.683 --> 00:09:25.505
And AI around insurance needs allows you

00:09:25.546 --> 00:09:28.126
to have clinical judgment without being

00:09:28.267 --> 00:09:30.488
uninhibited or being inhibited,

00:09:30.548 --> 00:09:31.149
I should say,

00:09:31.249 --> 00:09:33.450
from the noise of the machinery around

00:09:33.490 --> 00:09:33.909
insurance.

00:09:34.889 --> 00:09:36.671
Another thing I noticed in this series so

00:09:36.711 --> 00:09:39.572
far is that every AI tool you brought

00:09:39.633 --> 00:09:40.832
up gives something back.

00:09:40.893 --> 00:09:42.514
Last week, you mentioned presence.

00:09:42.553 --> 00:09:43.975
This week, clinical focus.

00:09:45.004 --> 00:09:46.225
We saw a lot of traction, too,

00:09:46.264 --> 00:09:49.265
when I mentioned this on LinkedIn around

00:09:49.306 --> 00:09:50.807
this whole idea of efficiency.

00:09:50.866 --> 00:09:54.328
So why frame AI in terms of restoration

00:09:54.788 --> 00:09:55.809
instead of efficiency,

00:09:55.830 --> 00:09:57.951
which is probably the most used word

00:09:58.010 --> 00:09:59.812
behind AI in that topic?

00:10:00.296 --> 00:10:01.697
Yeah, and we've been warped,

00:10:01.716 --> 00:10:04.559
just like we've been biased by insurance

00:10:04.580 --> 00:10:05.061
companies,

00:10:05.100 --> 00:10:06.942
the vendors are the ones who are selling

00:10:06.982 --> 00:10:07.682
efficiency.

00:10:08.803 --> 00:10:11.226
And that's the smaller part of the story.

00:10:12.047 --> 00:10:14.309
Everybody's out there promising that AI

00:10:14.330 --> 00:10:15.530
will save you time.

00:10:15.890 --> 00:10:18.293
Fine, yeah, I do think that it will,

00:10:18.374 --> 00:10:20.436
but save time is a productivity pitch.

00:10:21.211 --> 00:10:23.754
And this profession has been beaten half

00:10:23.794 --> 00:10:25.735
to death by productivity pitches,

00:10:26.297 --> 00:10:28.078
by unprofessional managers,

00:10:28.619 --> 00:10:30.620
non-PTs in the example of physical

00:10:30.640 --> 00:10:31.001
therapy,

00:10:31.022 --> 00:10:32.942
just pressing down on your schedule

00:10:33.004 --> 00:10:35.466
optimization and open doors and doing this

00:10:35.525 --> 00:10:38.087
and just more throughput, more throughput.

00:10:39.208 --> 00:10:40.009
No, thank you.

00:10:40.330 --> 00:10:42.172
I'm not interested in giving a therapist a

00:10:42.231 --> 00:10:43.312
faster treadmill.

00:10:43.753 --> 00:10:45.114
I'm interested in giving the back the

00:10:45.173 --> 00:10:48.797
things the system took.

00:10:48.836 --> 00:10:49.998
What did it take from us?

00:10:50.457 --> 00:10:51.339
It took presence.

00:10:51.379 --> 00:10:52.320
That was week one,

00:10:52.740 --> 00:10:53.539
all the way in the room.

00:10:53.900 --> 00:10:55.522
Week two, clinical focus.

00:10:56.883 --> 00:10:58.663
I want them to think like a clinician,

00:10:58.724 --> 00:11:00.826
not a coder or a compliance expert.

00:11:00.905 --> 00:11:03.148
Nobody ever became a physical therapist to

00:11:03.168 --> 00:11:04.609
be excellent at compliance.

00:11:05.048 --> 00:11:07.591
Nobody ever came to be the productivity

00:11:07.650 --> 00:11:08.312
king or queen.

00:11:08.792 --> 00:11:10.393
Nobody ever took it to know all the

00:11:10.413 --> 00:11:12.033
insurance rules and regulations.

00:11:12.734 --> 00:11:14.073
Nobody ever went in the profession for

00:11:14.114 --> 00:11:14.293
that.

00:11:14.354 --> 00:11:16.955
We went in there to do patient care

00:11:17.034 --> 00:11:18.596
and to impact people's lives.

00:11:19.135 --> 00:11:21.096
And if AI is to do its job,

00:11:23.177 --> 00:11:25.658
and I think that it can help us,

00:11:25.999 --> 00:11:27.639
those would be what it would be.

00:11:28.220 --> 00:11:32.162
Allow restoration as opposed to a faster

00:11:32.201 --> 00:11:32.542
treadmill.

00:11:33.428 --> 00:11:33.750
All right.

00:11:33.850 --> 00:11:35.931
In your Hemingway moment, I think,

00:11:35.971 --> 00:11:36.631
in this article,

00:11:36.672 --> 00:11:38.613
you wrote this great sentence,

00:11:38.852 --> 00:11:41.134
the machine handles the machinery.

00:11:41.515 --> 00:11:43.817
It sounds like it's just talking in a

00:11:43.876 --> 00:11:44.217
circle.

00:11:44.538 --> 00:11:45.899
I think it's a really beautiful sentence,

00:11:45.918 --> 00:11:46.940
but I want you to help me unpack

00:11:46.980 --> 00:11:47.059
it.

00:11:47.379 --> 00:11:48.921
What does that look like in a clinic

00:11:49.481 --> 00:11:51.722
where it's working the way it should,

00:11:51.763 --> 00:11:54.325
the way you hope the machine handles the

00:11:54.365 --> 00:11:55.125
machinery?

00:11:55.907 --> 00:11:56.246
Yeah.

00:11:56.326 --> 00:11:59.068
So if you remember last week,

00:11:59.089 --> 00:12:02.010
we talked about Sarah and how she's

00:12:02.071 --> 00:12:02.511
working.

00:12:03.677 --> 00:12:05.437
And the reality is she finishes an

00:12:05.498 --> 00:12:07.698
evaluation, she signs the note,

00:12:07.839 --> 00:12:09.599
and that's the end of her involvement with

00:12:09.619 --> 00:12:12.799
the administrative aspect of the patient.

00:12:13.220 --> 00:12:15.801
She doesn't touch code strings if this AI

00:12:15.961 --> 00:12:16.881
agent does its job.

00:12:16.902 --> 00:12:18.601
She doesn't choose a prior app.

00:12:18.621 --> 00:12:20.423
She doesn't get pulled out of a session

00:12:21.062 --> 00:12:23.803
because an authorization disagreed with it

00:12:23.844 --> 00:12:26.684
or an edit bouncing back from something

00:12:26.705 --> 00:12:28.144
you programmed into the EMR.

00:12:28.164 --> 00:12:29.446
I remember when that was really big.

00:12:30.385 --> 00:12:32.447
The system already did it.

00:12:32.587 --> 00:12:34.629
The AI agent already did it with the

00:12:34.669 --> 00:12:36.893
right functional data pulled from her

00:12:36.952 --> 00:12:37.352
nodes.

00:12:37.753 --> 00:12:39.914
The billing happens, the pre-cert happens.

00:12:39.975 --> 00:12:42.236
The payer gets honored.

00:12:42.697 --> 00:12:44.078
The payer gets what they get.

00:12:44.139 --> 00:12:45.681
They just get it from a non-human.

00:12:45.721 --> 00:12:47.261
They get it from an AI agent who

00:12:47.302 --> 00:12:48.823
doesn't typically have a bad day.

00:12:49.884 --> 00:12:52.446
But it happens, I would say,

00:12:53.267 --> 00:12:54.827
the same way that plumbing happens.

00:12:54.868 --> 00:12:57.009
It's in the walls, out of sight.

00:12:57.669 --> 00:12:59.671
She trusts that it's there and that it

00:12:59.711 --> 00:13:00.370
just happens.

00:13:00.390 --> 00:13:02.332
And so she's got presence and now she's

00:13:02.352 --> 00:13:03.974
had clinical judgment all at the same

00:13:04.014 --> 00:13:04.293
time.

00:13:04.774 --> 00:13:05.774
So what the difference is,

00:13:05.835 --> 00:13:08.277
is that Sarah spends her cognitive budget

00:13:09.157 --> 00:13:11.339
The hours that she's meant thinking on

00:13:11.418 --> 00:13:12.539
patients in front of her,

00:13:12.940 --> 00:13:15.282
not on billable units, not on pre-op,

00:13:15.601 --> 00:13:16.643
not on is it covered.

00:13:17.403 --> 00:13:19.284
She spends all of her time doing what

00:13:19.325 --> 00:13:20.405
she's paid to do.

00:13:21.346 --> 00:13:22.847
Not a clinic with fewer patients,

00:13:22.888 --> 00:13:24.629
a clinic where the therapist's brain is

00:13:24.690 --> 00:13:27.052
finally pointed at patients.

00:13:27.909 --> 00:13:28.730
at the right problem.

00:13:28.750 --> 00:13:29.711
And I would say it this way,

00:13:30.110 --> 00:13:32.251
good machinery is like good plumbing,

00:13:32.692 --> 00:13:33.212
right?

00:13:33.253 --> 00:13:35.114
A good agent would be like good plumbing,

00:13:35.134 --> 00:13:36.355
you don't admire it,

00:13:36.434 --> 00:13:38.696
you just you just stop thinking about it.

00:13:39.596 --> 00:13:41.577
And that silences the whole point.

00:13:41.638 --> 00:13:43.558
So when it breaks down, like plumbing,

00:13:43.619 --> 00:13:44.798
you would get concerned about it.

00:13:45.480 --> 00:13:46.379
Right.

00:13:46.399 --> 00:13:48.881
So and hopefully the AI won't break down.

00:13:49.121 --> 00:13:50.702
You only notice it when there's an issue.

00:13:50.763 --> 00:13:52.644
And every other time you just it's it's

00:13:52.683 --> 00:13:53.224
in the walls.

00:13:54.100 --> 00:13:54.860
Another line I liked,

00:13:54.980 --> 00:13:56.923
verification is the product.

00:13:56.982 --> 00:13:59.005
This one sort of stings a little bit

00:13:59.046 --> 00:14:00.788
because it feels a little true.

00:14:01.207 --> 00:14:05.633
That sounds more like more than an AI

00:14:05.692 --> 00:14:06.293
principle, though.

00:14:06.333 --> 00:14:08.395
That sounds more like a systemic issue.

00:14:08.416 --> 00:14:10.378
It almost sounds philosophical if you want

00:14:10.398 --> 00:14:10.778
to get it to.

00:14:11.119 --> 00:14:14.322
So what does verification is the product

00:14:14.422 --> 00:14:14.602
mean?

00:14:15.393 --> 00:14:17.092
Yeah, it's a philosophy.

00:14:17.133 --> 00:14:20.614
It's the one that keeps this from really

00:14:20.673 --> 00:14:21.673
going wrong here.

00:14:21.734 --> 00:14:23.455
So here's the trap with AI in healthcare.

00:14:23.495 --> 00:14:25.355
People think that the output is the

00:14:25.394 --> 00:14:27.075
product, the generated note,

00:14:27.475 --> 00:14:30.956
the auto-selected code, the prior auth,

00:14:31.316 --> 00:14:32.777
the compliance check.

00:14:32.937 --> 00:14:35.116
Any tool could generate an output.

00:14:35.157 --> 00:14:37.658
The thing you're actually buying,

00:14:37.778 --> 00:14:39.278
the one thing that makes safe

00:14:40.171 --> 00:14:41.351
is that it is checkable,

00:14:41.732 --> 00:14:44.115
that a human being can look at every

00:14:44.235 --> 00:14:45.956
one of those decisions, re-understand it,

00:14:46.096 --> 00:14:47.958
and in theory, override it, right?

00:14:48.619 --> 00:14:50.461
The agent doesn't get to be a black

00:14:50.500 --> 00:14:52.682
box that quietly decides how to get paid.

00:14:52.863 --> 00:14:54.004
It does the legwork,

00:14:54.544 --> 00:14:55.745
and it shows the work.

00:14:55.826 --> 00:14:57.388
The clinician stays accountable.

00:14:58.028 --> 00:14:59.590
The audit trail is the product.

00:14:59.610 --> 00:15:01.652
The verifiability is the product.

00:15:01.672 --> 00:15:04.094
So ultimately, the clinician always has

00:15:05.014 --> 00:15:07.035
you know the benevolent dictator right

00:15:07.196 --> 00:15:09.456
over a verification check of an ai it

00:15:09.475 --> 00:15:11.797
just doesn't again the output isn't the

00:15:11.836 --> 00:15:13.496
product the fact that a human can check

00:15:13.537 --> 00:15:17.038
it understand it and overrule it that's

00:15:17.077 --> 00:15:18.898
the problem it's the human that can

00:15:18.918 --> 00:15:22.078
overrule it verification you know is is is

00:15:22.119 --> 00:15:24.139
the whole ball game yeah well you

00:15:24.159 --> 00:15:26.320
described sort of shangri-la right which

00:15:26.379 --> 00:15:28.760
is a world where the clinician gets to

00:15:28.941 --> 00:15:29.240
asks

00:15:30.048 --> 00:15:32.490
what does this patient need instead of

00:15:32.931 --> 00:15:34.471
this little voice, this hum,

00:15:35.392 --> 00:15:37.452
will insurance approve this thing that I'm

00:15:37.493 --> 00:15:38.333
already thinking of?

00:15:38.953 --> 00:15:40.894
Do you think we've forgotten that those

00:15:40.934 --> 00:15:43.336
are supposed to happen in that order?

00:15:43.397 --> 00:15:45.018
Because I feel like it does happen.

00:15:45.038 --> 00:15:46.879
It just happens in the other order.

00:15:47.616 --> 00:15:47.756
Yeah,

00:15:47.777 --> 00:15:49.318
I would say that we haven't necessarily

00:15:49.359 --> 00:15:50.460
forgotten about it.

00:15:50.720 --> 00:15:51.780
We've been trained.

00:15:52.201 --> 00:15:54.844
We've been trained to learn helplessness.

00:15:54.884 --> 00:15:55.644
We were trained out of it.

00:15:55.664 --> 00:15:56.466
There's a difference.

00:15:56.485 --> 00:15:57.927
And the difference really does matter.

00:15:58.648 --> 00:16:00.350
You know, forgetting would be innocent.

00:16:00.529 --> 00:16:01.831
You know, this is an innocent.

00:16:01.870 --> 00:16:04.474
It's it's conditional behavior that we've

00:16:04.494 --> 00:16:06.556
been trained in to deny a clinician enough

00:16:06.615 --> 00:16:06.975
times.

00:16:07.636 --> 00:16:09.418
make them redo enough notes,

00:16:10.197 --> 00:16:11.697
claw back enough visits,

00:16:11.738 --> 00:16:13.379
and they'll start pre-checking their own

00:16:13.438 --> 00:16:15.158
judgment against the payer's rules.

00:16:15.318 --> 00:16:16.980
We've just been normalized into the

00:16:17.019 --> 00:16:17.720
payer's rules.

00:16:17.740 --> 00:16:19.080
And before they can say it out loud,

00:16:19.120 --> 00:16:21.961
it's a survival reflex, if you will.

00:16:22.782 --> 00:16:24.662
Those two questions are supposed to happen

00:16:24.701 --> 00:16:25.381
in that order.

00:16:25.822 --> 00:16:27.783
Need first, coverage second.

00:16:27.903 --> 00:16:29.043
What does the patient need?

00:16:29.062 --> 00:16:30.703
What does my clinical judgment tell me?

00:16:31.283 --> 00:16:32.644
You figure out what the human being

00:16:32.684 --> 00:16:33.285
requires,

00:16:33.865 --> 00:16:35.746
and then you go fight the administrative

00:16:35.767 --> 00:16:36.707
fight to deliver it.

00:16:36.748 --> 00:16:38.629
And it's okay to have adult conversations

00:16:38.649 --> 00:16:40.811
with patients about coverage issues and

00:16:40.831 --> 00:16:42.032
that that's not your bag,

00:16:42.072 --> 00:16:43.212
but this is what you need.

00:16:43.413 --> 00:16:44.393
Let them decide.

00:16:44.953 --> 00:16:46.835
And what happens is most patients will

00:16:47.355 --> 00:16:48.937
typically decide to do what's right,

00:16:49.096 --> 00:16:50.378
not what the insurance covers,

00:16:50.457 --> 00:16:51.618
at least in my experience.

00:16:51.638 --> 00:16:52.859
It's just that we've been conditioned

00:16:52.918 --> 00:16:53.299
elsewhere.

00:16:53.980 --> 00:16:55.441
And the whole point of the agent is

00:16:55.480 --> 00:16:57.261
to take the second question off the

00:16:57.302 --> 00:16:58.283
therapist's plate.

00:16:58.562 --> 00:17:00.583
So the first question gets the clinicians

00:17:00.644 --> 00:17:02.164
full and undivided attention,

00:17:02.225 --> 00:17:03.166
clinical judgment.

00:17:03.586 --> 00:17:05.207
Get the order right and everything sort of

00:17:05.267 --> 00:17:07.449
downstream.

00:17:07.949 --> 00:17:09.490
Everything downstream gets healthier.

00:17:09.529 --> 00:17:11.250
It's not that we forgot, you know,

00:17:11.290 --> 00:17:12.172
the right order.

00:17:12.491 --> 00:17:14.232
We got trained out of it one denial

00:17:14.292 --> 00:17:14.853
at a time.

00:17:14.953 --> 00:17:17.434
Need first, coverage, treatment second,

00:17:17.654 --> 00:17:18.076
or, you know,

00:17:18.155 --> 00:17:19.316
treatment coverage from the insurance

00:17:19.355 --> 00:17:19.957
company second.

00:17:19.977 --> 00:17:22.157
Flip those two and what you've done is

00:17:22.178 --> 00:17:23.358
you've given your judgment away.

00:17:24.817 --> 00:17:27.057
So you brought Sarah back from from week

00:17:27.178 --> 00:17:29.259
one from article one in this series,

00:17:29.798 --> 00:17:31.140
you bring her back at the end with

00:17:31.180 --> 00:17:33.701
something you call this clinical focus.

00:17:34.621 --> 00:17:36.642
How would she know she'd gotten it back?

00:17:36.682 --> 00:17:38.583
What would feel different about her day?

00:17:38.603 --> 00:17:39.903
Would it be the thought process,

00:17:39.923 --> 00:17:41.364
the order in which or what would it

00:17:41.544 --> 00:17:42.585
look and feel like?

00:17:43.334 --> 00:17:43.614
Yeah.

00:17:43.634 --> 00:17:44.835
And so you're a clinician, Jimmy.

00:17:44.855 --> 00:17:47.476
So you have to go back and reimagine

00:17:47.496 --> 00:17:47.996
this a bit.

00:17:48.036 --> 00:17:49.817
But she'd feel it in her body at

00:17:49.856 --> 00:17:50.537
the end of the day.

00:17:50.576 --> 00:17:51.396
That's specific.

00:17:51.436 --> 00:17:52.698
This, you know, hollow.

00:17:53.417 --> 00:17:55.118
We get the sort of hollow tiredness,

00:17:55.159 --> 00:17:56.519
the one that comes from fighting the

00:17:56.558 --> 00:17:58.980
system, this fighting and fighting.

00:17:59.019 --> 00:18:00.441
It's very, very stressful.

00:18:00.461 --> 00:18:01.641
And it doesn't come from treating

00:18:01.701 --> 00:18:02.080
patients.

00:18:02.141 --> 00:18:02.981
It comes from fighting.

00:18:03.041 --> 00:18:04.221
And you're tired.

00:18:04.261 --> 00:18:05.403
In this case, she's tired.

00:18:06.403 --> 00:18:09.505
Now she's tired from clinically treating a

00:18:09.545 --> 00:18:09.964
patient.

00:18:10.025 --> 00:18:11.566
It's a different type of field.

00:18:11.586 --> 00:18:14.406
The kind you earn from actual work.

00:18:14.426 --> 00:18:15.527
They say, like,

00:18:15.887 --> 00:18:17.587
honest sweat doesn't smell.

00:18:17.929 --> 00:18:19.750
Honest tired feels good.

00:18:21.049 --> 00:18:23.211
I work today helping patients get better,

00:18:23.290 --> 00:18:24.491
not fighting the system.

00:18:25.051 --> 00:18:26.834
when sixty seventy percent of your time is

00:18:26.854 --> 00:18:28.715
fighting fighting system that's that's

00:18:29.036 --> 00:18:31.397
that's you know that's a problem she

00:18:31.458 --> 00:18:33.440
noticed she stopped rehearsing denials in

00:18:33.480 --> 00:18:35.000
her head you know that this noise and

00:18:35.040 --> 00:18:36.722
this hum you know started to go away

00:18:36.762 --> 00:18:37.844
without these little internal

00:18:37.864 --> 00:18:39.365
investigations going on

00:18:41.467 --> 00:18:44.328
And here's the quiet one, you know,

00:18:44.348 --> 00:18:46.510
she went home and she didn't need to

00:18:46.550 --> 00:18:48.031
bring her clinic home with her.

00:18:48.271 --> 00:18:49.653
It's not the second job.

00:18:49.972 --> 00:18:51.734
I'm not talking about notes that you have

00:18:51.755 --> 00:18:54.296
to do, but, oh man, units per visit,

00:18:54.336 --> 00:18:55.636
what was my report card telling,

00:18:55.717 --> 00:18:56.938
all these kind of things, you know.

00:18:57.898 --> 00:18:59.799
So you know it not from a dashboard,

00:18:59.839 --> 00:19:00.980
but you know it from the feeling that

00:19:01.020 --> 00:19:01.481
you had.

00:19:02.357 --> 00:19:04.357
And you know you've got your focus back

00:19:04.498 --> 00:19:06.878
the day you're tired from the work,

00:19:07.519 --> 00:19:09.440
not from fighting the machinery around the

00:19:09.460 --> 00:19:09.700
work.

00:19:09.740 --> 00:19:11.680
And every healthcare provider knows the

00:19:11.720 --> 00:19:12.079
difference.

00:19:12.099 --> 00:19:13.480
They know the feeling of the difference.

00:19:14.340 --> 00:19:14.421
Well,

00:19:14.441 --> 00:19:17.020
let's give a little forward tease to next

00:19:17.060 --> 00:19:18.382
week and the next part of this series.

00:19:18.422 --> 00:19:19.741
And when people finish this series,

00:19:19.761 --> 00:19:21.642
what do you hope they'll never again

00:19:21.662 --> 00:19:24.202
accept as just part of healthcare or that

00:19:24.282 --> 00:19:24.583
hum?

00:19:25.338 --> 00:19:25.659
Yeah,

00:19:25.679 --> 00:19:27.480
so what you're going to see is continued

00:19:27.599 --> 00:19:30.421
emphasis on what I believe is the best

00:19:30.560 --> 00:19:32.461
use case for AI in health care.

00:19:32.840 --> 00:19:34.701
Week one is about presence.

00:19:35.122 --> 00:19:37.403
Week two is about clinical judgment.

00:19:37.962 --> 00:19:39.663
And week three will just take this to

00:19:39.703 --> 00:19:40.844
different levels, again,

00:19:40.903 --> 00:19:42.704
more agentic view of where we're going

00:19:42.724 --> 00:19:43.444
with AI.

00:19:43.825 --> 00:19:45.365
And hopefully, by the end,

00:19:45.405 --> 00:19:46.546
they'll get rid of that hum,

00:19:46.625 --> 00:19:48.727
the idea that the background dread,

00:19:49.626 --> 00:19:50.787
the cost of the job.

00:19:50.846 --> 00:19:52.228
I want a clinician to finish.

00:19:53.259 --> 00:19:54.680
reading this or looking at this and

00:19:54.700 --> 00:19:55.941
actually get a little bit angry,

00:19:55.980 --> 00:19:56.780
but in a good way,

00:19:56.820 --> 00:19:58.840
to stop treating the second voice in the

00:19:58.901 --> 00:20:01.321
room, the payer, the seven PM note,

00:20:01.362 --> 00:20:02.582
the revenue anxiety,

00:20:02.602 --> 00:20:04.122
the erosion of your own judgment,

00:20:04.701 --> 00:20:07.542
to stop treating any of that as weather.

00:20:08.762 --> 00:20:10.943
And it's just how it is,

00:20:11.003 --> 00:20:12.723
is the price of admission, right?

00:20:12.763 --> 00:20:13.943
I wanted a little anger there.

00:20:14.443 --> 00:20:15.684
None of this is natural,

00:20:15.825 --> 00:20:16.704
all of it was built.

00:20:17.065 --> 00:20:17.664
And by the way,

00:20:17.684 --> 00:20:19.565
I want the EMR companies to read this,

00:20:19.605 --> 00:20:20.726
because I'm going to be submitting,

00:20:20.806 --> 00:20:21.306
secondly,

00:20:21.346 --> 00:20:22.846
Tim Flynn and I tried to submit an

00:20:22.905 --> 00:20:24.685
article for a professional publication,

00:20:24.726 --> 00:20:25.727
peer-reviewed publication,

00:20:25.787 --> 00:20:27.086
on the use of AI agents.

00:20:27.666 --> 00:20:29.547
And it was along these themes.

00:20:29.567 --> 00:20:31.147
A little harder to explain in the academic

00:20:31.188 --> 00:20:31.508
sense.

00:20:32.167 --> 00:20:32.907
But really,

00:20:32.948 --> 00:20:35.067
the ultimate goal will be another position

00:20:35.107 --> 00:20:37.308
paper that EMR companies look at and say,

00:20:37.828 --> 00:20:39.429
this is really what I need to do

00:20:39.449 --> 00:20:40.249
for a clinician.

00:20:40.769 --> 00:20:43.569
So that's what the hope is here.

00:20:45.589 --> 00:20:46.490
If you think about it, Jimmy,

00:20:46.530 --> 00:20:48.933
nothing about our system is natural.

00:20:49.153 --> 00:20:50.015
It was all built.

00:20:50.234 --> 00:20:51.855
The forms, the denials,

00:20:51.895 --> 00:20:53.097
the second voice in the room,

00:20:53.157 --> 00:20:54.959
and anything built by people can be

00:20:55.079 --> 00:20:56.580
rebuilt by people.

00:20:57.182 --> 00:20:58.303
Stop calling it weather.

00:20:59.063 --> 00:21:02.006
Start calling it a choice and then say,

00:21:02.105 --> 00:21:05.690
can we really use something like AI to

00:21:05.730 --> 00:21:07.171
help us get there?

00:21:07.191 --> 00:21:07.751
That's the goal.

00:21:08.700 --> 00:21:09.440
All right.

00:21:09.460 --> 00:21:10.701
Look forward to this next week.

00:21:10.721 --> 00:21:13.041
Make sure to subscribe on Substack.

00:21:13.061 --> 00:21:14.663
This conversation around AI usually

00:21:14.702 --> 00:21:17.924
focuses on what technology can do.

00:21:17.964 --> 00:21:18.845
You've seen the ads.

00:21:19.204 --> 00:21:20.165
If you've somehow missed them,

00:21:20.405 --> 00:21:21.145
teach me how to do that.

00:21:21.826 --> 00:21:23.287
But we want to keep asking the different

00:21:23.326 --> 00:21:25.667
question, what can technology give back?

00:21:25.788 --> 00:21:27.229
Last week, it was presence.

00:21:27.288 --> 00:21:28.710
This week, it was clinical focus.

00:21:29.269 --> 00:21:30.691
What is the focus of next week?

00:21:31.413 --> 00:21:32.173
Oh, you're going to see.

00:21:32.233 --> 00:21:33.474
It's going to be a surprise to you.

00:21:33.494 --> 00:21:34.194
Spoiler alert.

00:21:35.155 --> 00:21:37.616
If you're enjoying this series called The

00:21:37.656 --> 00:21:38.656
Care of the Age of AI,

00:21:38.696 --> 00:21:40.337
subscribe to The Operator on Substack.

00:21:40.357 --> 00:21:41.557
We'll put that link in the show notes

00:21:41.577 --> 00:21:44.179
so you don't miss the next installment.

00:21:44.400 --> 00:21:46.440
Until next time, keep operating.

00:21:49.342 --> 00:21:50.021
And I'll send you...