Lab Medicine Rounds

In this episode of “Lab Medicine Rounds,” Justin Kreuter, M.D., sits down with Justin Juskewitch, M.D., Ph.D., and Andrew Norgan, M.D., Ph.D., both assistant professors of laboratory medicine and pathology at Mayo Clinic, to talk about informatics for the pathologist.

Show Notes

What is Lab Medicine Rounds?

A Mayo Clinic podcast for laboratory professionals, physicians, and students, hosted by Justin Kreuter, M.D., assistant professor of laboratory medicine and pathology at Mayo Clinic, featuring educational topics and insightful takeaways to apply in your practice.

(electronic music)

- This is Lab Medicine
Rounds, a curated podcast

for physicians, laboratory
professionals, and students.

I'm your host, Justin Kreuter,
the Bow Tie Bandit of Blood,

a transfusion medicine
pathologist at Mayo Clinic.

And today, we're rounding
with Dr. Justin Juskewitch

and Dr. Andrew Norgan,
both assistant professors

of laboratory medicine and
pathology here at Mayo Clinic,

to talk with us about
informatics for the pathologist.

They're both leaders in this field,

and I'm grateful for both of
you guys joining us today.

- [Justin] Good morning.

- [Andrew] Thanks for having us.

- [Host] All right, so
let's kick off with kind of,

you know, I think as
pathologists, maybe we're used to

calling the help desk at our institutions

or sometimes dealing with IT
for support on our projects.

But, from your vantage point,
why should pathologists

care about informatics in 2022?

- I think actually informatics
has been in the core

of pathology and lab medicine
for a very, very long time.

Our field was the first
to spearhead the use

of information systems to collect data

and to help facilitate laboratory testing

and anatomic pathology
consultations for decades now.

And we've been able to use that
technology up to this point

to really help optimize the
care we bring to patients

and the workflows within our lab.

And so, informatics, really,
in some respects in healthcare,

got its start in the medical laboratory

and has been able to empower
us to be able to use that data,

to use that information and
the knowledge we glean from it

to really transform clinical practice.

So, I would actually reframe
the question slightly as to

why should pathologists continue
to care about informatics?

Because we have been, in
large part, thought leaders,

especially during the early
inception of this field.

What do you think, Dr. Norgan?

- [Andrew] Yeah, I agree
completely and I just

extend on that, I think
healthcare organizations

have done a great job up to
this point in sort of recording,

cataloging, warehousing
their information, and maybe

we've been less successful
at taking that information

and creating value for the
patient from that information

and bring that back into the practice.

And if we look at every other
industry, they're probably 10,

maybe 20 years ahead of
us in using information

to do whatever it is they do better.

And so, I think as we look
at informatics going forward

now in time, and especially
with artificial intelligence

and machine learning coming to the fore,

and with cloud technologies
making it easier

to bring information
together, compute on that,

find some value, and again bring that back

into your practice, it's a
great time for pathologists

to be digging in and understanding

well, how can I use this
data and this information

to improve whatever aspect
of the practice I'm doing,

be that anatomic pathology,
clinical pathology,

or something related to either of those.

- [Host] I really appreciate you guys

putting me on the right track.

We're really continuing our
interests in informatics.

But you guys really are,
you know, "informaticians",

if that's the right word, but,

and I love picking your
brain, but, you know,

not all of us can be
information experts like you,

although we certainly are
using it in our practice.

I was wondering if you could kind of share

with our audience, what are
some of the fundamentals

that pathologists in practice,
pathologists in training

should be aware of and
might be thinking about?

- You know, fundamentally,
I think it's critical

for people to understand
the systems that exist

for managing and organizing the data.

And so, in a healthcare institution,

at sort of a first approximation,

that's the laboratory information system

and the electronic health record.

Those are the systems that
are really going to allow

communication of information
to the clinicians.

And so understanding, how does information

flow into these systems,
how does information

flow out of these systems, and again,

at what point are there
opportunities to do value-added work

with this information to improve
whatever it is we're doing

as a pathologist, whether
you're running a laboratory,

or very much, you know, doing case-based

anatomic pathology work,
there are opportunities

to use that data to do that work

better with higher quality
and to bring, again,

new technologies to the
work that we're doing.

So, fundamentally understanding,
what are those data types,

how do those systems talk to one another,

how do those systems
allow third party systems

to interface with them and add value,

be that again, middleware,
or other kinds of things

that we see deployed in our systems,

I think that's a fundamental piece

that people need to understand.

I think the other piece, and
I'll let maybe Dr. Juskewitch

talk about this a little
bit more is understanding

how to use that data to
develop verifiable insights

so, you know, when we
look at publishing now,

pretty much every paper has a
statistical component to it.

Every machine learning paper certainly has

statistical aspects to it.

And as we're looking to
bring those technologies

into our practice and they become part of

what it means to, quote
unquote, "do medicine"

or "do pathology", I think
it's critical that we have

enough understanding
of those technologies'

systems approaches, and
fields that we're able to

adequately judge, you
know, what is good work,

what is maybe not so good work,

where are the flaws in the
methodology of this study,

and so on and so forth.

- [Justin] Yeah, I completely agree.

So, what Dr. Norgan is talking about

has been kind of the involvement
of evidence-based medicine,

I would say across most
fields of medicine,

if not all over, like,
the past decade or so.

The next level up is
not just the application

of those, you know, those skills

to now a new area of research,
evaluating machine learning

and artificial intelligence and stuff.

But I think the next level up
that will be for pathologists

and particularly pathology
trainees coming in

is going to be data literacy.

So, it's knowing how
your data is captured,

in what form it is captured, and then

based off of those forms,
how do you do, you know,

as you're taking data out of your LIS

or hopefully in some sort of
replicated source from that,

how do you evaluate the data critically

that you're getting out, find any errors

or missing data elements,
use the appropriate technique

to fill those in and then be able to

take that data to start
generating insights.

And it may be the most
simple statistical comparison

to begin with, and
evolving that way upward

towards machine learning.

But to be perfectly honest,
for many of us operationally

in the lab or in the
anatomic pathology space,

being able-just to be
able to access our data,

manipulate it, and present
it in ways that can inform

even operational things
in real time in our labs,

have been, and continue
to be huge value-add gains

for the laboratory and our patients.

- [Host] That's interesting.

As I hear you guys explaining
these fundamentals,

in my head, I'm really almost
constructing a map of, like,

you know, the system, you
know, and the flows, you know,

there's rivers or something
flowing through there.

But that's really an
interesting way to really also

highlight, you know,
understanding the system

that we work in and how
it can affect change.

Is this kind of the same
strategy you guys take when,

I know you guys are teaching

our Inner Pathology Residency
program, teaching informatics,

is this kind of the same
approach you take there

or is there a different approach

and why might you go that way?

- [Andrew] So, we sort
of have two approaches

that we've gone down.

One is a didactic approach where we cover

kind of a core curriculum
that's been developed by CAP

and by the Association
for Pathology Informatics.

It's known as the pathology
informatics for residents,

or pathology informatics
essentials for residents.

And that curriculum
lays down really, again,

kind of getting to your earlier question,

the baseline level of knowledge
that would be expected

of a pathology resident in sort
of the realm of informatics.

So everything that Justin
and I have been talking about

in terms of data literacy systems,

how systems talk to one another,

the standards that are
used in the industry

to facilitate that communication,
et cetera, and so on.

So we have a didactic
curriculum built around that.

And then we have a sort of
application-based curriculum

where we try to get
people to actually dig in

and work with the data and
perform some of these functions.

Again, not to become
computer programmers, per se,

or anything else, but to have sort of that

experiential learning of
encountering data-based problems

and then strategies for
solving those problems

in order to accomplish the kinds of tasks

Justin was talking about
earlier, be it an analytic task,

or putting a system into place
that's going to take data in

and then put out some sort of altered data

to facilitate a laboratory
function, and so on.

We expect, you know, a
pathologist or laboratorian,

you'll be sort of in charge
of making these things happen

but you probably won't be
doing all of them yourself.

And so the facility we're
hoping people walk away with

is to know what's possible.

So what can be done, what
are the likely issues

that are going to be
encountered in pursuing that,

what are possible
solutions to those issues

so that they can guide
their team in sort of

figuring out the problem,
solving the problem,

and finally implementing that solution.

- [Justin] And circling back
around, I think the sweet part

or the sweet spot for pathologists is that

we, as part of our training,
and then part of our

laboratory directorship roles,

we have a whole lot of
experience in the validation

and evaluation of new laboratory
tests or new offerings.

And these informatics tools,

whether they're software applications

or just simple data visualizations
that help these workflows

can be the same methodologies
and the same way of talking

and evaluating them really do run in,

are pretty much the same.

And so, as part of our
training and our role already,

we already have a leg up on
being able to evaluate these

and assess their impact,
good or bad, on our workflows

and our patient care.

So, pathologists already have
the mindset and the training

to be able to begin to
take these offerings

and do the critical evaluation
and being the leaders

of those groups within their labs

that Dr. Norgan talked about.

- And how do you guys
recommend that pathologists

reach out to informatics
experts for collaboration?

Like you said, you know,
having a general understanding

and what's possible is healthy
and important, but, you know,

how do we interact with
the informatics group?

- Like you would any
other subspecialty expert

you've got in your large laboratory.

I'm particularly partial to
tea, so, a free tea for me

will get you about 30
minutes of good conversation.

Andy's laughing because
for him, it's coffee.

But to be honest, just like,

I don't think there is any magic formula.

It's a phone call, it's
running into a colleague

in the hallway, it's an email.

But coming to that conversation
with a crystallized idea

of what your need or your
desired, you know, product

or functionality is, and some
idea where that data resides

is actually a really good
starting place to start

those conversations because
then it makes it easier

for individuals with an
informatics background to start

scoping in their head exactly
how big of an ask it is

and what sort of data
infrastructure is already in place

on which that offering may be built.

But, to be honest, for most of the time

when I need some help
from the informatics crew

here at our own institution,
it starts with a page,

it starts with a phone call,
it starts with an email,

and usually an offer to buy them coffee.

Right, Dr. Norgan?

- Yeah, I completely agree.

It's just like engaging any
other expert that you would

to help solve a clinical
or procedural problem

in your laboratory or again
anatomic pathology area.

I think, you know, breaking
it out more generally,

it's going to be very
different depending on the kind

of institution you're at,
the size of that institution,

and how developed that
institution's current approach is

towards information
management and informatics.

But every institution's
going to have a CMIO,

many institutions will
have an associate CMIO

for laboratory medicine
and pathology or someone,

you know, over the lab's
areas, and that's an individual

you should be able to engage
with and start to talk about

the issues that you're
facing in your laboratory,

or in your practice, and how the systems

that your institution might
be able to address those.

Most departments at divisions
will have that person

or that group of people who are sort of

the informatics folks.

In DMLP, here, we have a
vice chair of informatics

who has kind of a scope over
the LIS and the interchange

between the LIS and the
EHR who'd be, again,

a wonderful person to address

with those kinds of issues and questions.

And then, new in our department,
we actually have a division

of computational pathology and AI.

So, a whole group of
individuals now who are really

dedicated to that and
who would be resources

at our institution that
you could reach out to.

And I think other institutions
are doing similar things

so, be it someone in
your division department,

or at the associate or actual CMIO level,

there's someone at your
institution who cares about data

and using that data to
solve clinical problems

and you should be able
to reach out to them.

- That's really helpful.

I appreciate both your advices
on this about, you know,

having an idea of what your
issue is and what you want,

where some data is, and also
how you might reach out to,

and the acronym "CMIO",

is that Chief Medical Informatics Officer?

- Chief Medical Information
Officer I think is usually,

and it's, you know, they
sit next to that CIO

who's more of a technical
person or an IT person

and often bring in that sort
of medical aspect to them.

Frequently, the CMIO is going to be

an informatics strained physician.

- I see. So, I'm on team coffee,

that's who I'm buying
coffee for, it sounds like.

(Host and Andrew laugh)

So, we've been talking about
a couple of different things.

I was curious for those of us,
like myself, who, you know,

I remember doing my informatics rotation

in my residency program,
I think it was maybe

14 or 15 years ago.

I'm sure I'm quite outdated
at this point in time.

I was curious, are there
any informatics resources

that you guys recommend that
those of us could kind of

brush up on our, you know, we've
been talking about systems,

you were using the acronym LIS

for laboratory informatics systems,

EHR for electronic health
record, but then also,

we've been talking about data literacy.

Are there any, you know,
resources you guys recommend?

- So if you had asked us,
what, 10 to 15 years ago,

the list would've been
really, really short

because informatics as a field has been,

just like all of these new
divisions are spinning out

and now we have fellowships,

and now this is a medical subspecialty.

Those resources have been
expanding more and more.

So, on the pathology side,

the American Society
for Clinical Pathology

has some online offerings and
they also have some textbooks

that they now offer for sale,
either electronic or print.

So there are some resources there.

The American Medical Informatics
Association Association,

AMIA also has been really busy spinning up

a whole lot of educational offerings

beyond just the laboratory,

but in the broader
medical practice as well.

And then as Dr. Norgan mentioned,

the Association for Pathology Informatics

now has an annual conference,

and part of that conference
is a lot of workshops.

And then depending on
how deep you want to go

into the weeds as far as data science

and how much do you want
to actually start learning

how to manipulate data on your own,

a lot of the different
programming languages and stuff

now have conferences and
workshops that they offer

and the online resources just
for reading to be able to

start learning how to do
some of that data literacy

and manipulating data and
stuff and making visualizations

those offerings available
online have exploded

over the last decade or so, I would say.

Dr. Norgan?

- Yeah, agree completely

with all of the suggestions.

I think if you want a sort of healthcare

and pathology specific
look at informatics,

the University of Pathology
Informatics by ASCP

is a great resource
and it offers, I think,

over 20 different lectures or sort of

blocks of curriculum that cover, really,

a large spectrum of informatics.

So you can pick and choose.

I think there is some
cost associated with that,

or at least it might be governed by

your institution's
subscription to ASCP resources.

But there are certainly pieces in there

that I think are valuable
and would round out

someone's informatics education.

The online resources have exploded

in the last probably 5 years, especially.

They're not always specific to pathology,

but I think that's okay.

A lot of the problems are
really generalizable problems.

And so, for free, you can get
an education in data science,

data literacy in one of the popular

open source programming
languages Python or R.

And you can do that all
at your own pace online

with no cost at all on any computer.

So I would say it's incredibly accessible

if you want to learn about
some of these things.

If you want to go further
and actually develop

real expertise in one of these
areas, maybe an area related,

again, to your work, say
you're in molecular pathology

and you want to learn
bioinformatics more deeply

or your statistical background
isn't what you wish it was

and you want to beef up on that

or you want to learn about
AI and machine learning.

There are online sort of
learning platforms that now have

what I would argue, I mean the curriculum

is coming from MIT and
Harvard and Stanford,

it's really, you're getting
I think a top tier education

and those would be platforms
like edX or Coursera.

There's often a free option associated

if you don't need credit
and don't need to sort of

be able to take that and use
it for something downstream.

You can subscribe to these
courses and go through them

at your own pace without cost,

or I think there's a
minimal cost associated

with some of them.

So, really great opportunities to go out

and get an education in
an area of informatics

that interests you and that is relevant

to the kind of work that you do.

- [Host] That's phenomenal.
So we're going to,

for our listeners, we're
going to put a number

of those recommendations
in the show notes.

And we've been routing with Dr. Juskewitch

and Dr. Norgan. Thank you
both for taking the time

to talk with our listeners today

and help us to continue our
interests in informatics.

- Pleasure.

- [Andrew] Thanks very much.

- [Host] So to all of our listeners,

thank you for joining us today.

We invite you to share your thoughts

and suggestions by email.

Please direct any suggestions to

mcleducation@mayo.edu and
reference this podcast.

If you've enjoyed Lab
Medicine Rounds podcast,

please subscribe until
our next rounds together.

We encourage you to continue
to connect lab medicine

and the clinical practice
through insightful conversations.

More episodes

Chapters

Show Notes

What is Lab Medicine Rounds?