Data Science & Analytics

Episode 22: Ethics and Health Insurance data with Onyinye Daniel from Highmark Health

In this episode Onyinye Daniel, Vice President, Data & Analytics Strategy at Highmark Health, talks about her career in healthcare from the research side, to working at a hospital and now on the payer side in health insurance. Listen to the episode to learn how Onyinye’s background in health policy and bioethics has shaped her career.

 

Transcript

this is buffalo state data talk the

podcast where we introduce you to how

data is used and explore careers that

involve data hello and welcome back to

another episode of buffalo state data

talk i'm your host heather campbell and

thank you for joining us for episode 22.

today we'll be talking to oninie daniel

vice president of data and analytics

strategy at highmark health welcome to

the show

thank you heather very very happy to be

here

so just to start us off could you give

us a general overview of what you do as

a vp at highmark

my responsibilities are pretty broad so

one

i'm responsible for the data and

analytics strategy kind of communicating

that throughout the enterprise

developing it and then seeing to evolve

it and communicate it throughout the

enterprise really trying to articulate

who we are as a data and analytics

organization

both relative to high mark but also

relative to the broader healthcare

industry

and so that's really important to be

able to translate what we do and what

our goals are and how they could

potentially support

enterprise priorities but also how they

really

push highmark

to be a leading organization

um in you know data and analytics and so

so that's a really huge part of my

of my responsibilities i have

accountability for our data and ai

ethics

strategy and operationalization plan

so really making sure that it's not just

like what we can do with data and how we

use data but

should we

be leveraging the data in this way

should we

be leveraging this type of data what are

the ethical implications for the use of

that data for our patients members the

community

part of it is around analytics

governance and making sure that we're

leveraging analytics in a consistent way

part of it is like the algorithms that

we use in our models are explainable

transparent

that we're not introducing new biases

part of it is just like standardization

making sure that we are very intentional

about

communicating why or how a model is

intended to be used right rather than

just kind of a free-for-all so there's

some consistency and standardization

i also have our enterprise data and

analytics research agenda

and so really

um

pulling together like as a data

analytics organization what are the

areas of research that we should be

exploring right like you know some of

the areas around voice analytics a lot

of remote monitoring

things that may not necessarily be

driving business value today but have

the potential to improve health outcomes

down the line or improve efficiencies or

cost savings or

better engagements engagement

that is definitely a lot that you and

your team are working on i'm i'm

impressed how would you say that your

time is split between working

collaboratively with people on your team

and working independently

i do i spend a lot of time in meetings

um

i spend a lot of time in meetings and i

would say

um

yeah probably like 80 20

of my time is in

and i build in

try to build in when i can like 30

minute blocks

like really thinking through

frameworks and processes as somebody who

is a vp and who is more on the

management administrative side of things

can you tell us what kind of data

highmark is analyzing i know that you

probably spend more time on the

management side but what are your data

analysts

analyzing oh there's so much

so we have as a health

organization a blended health

organization because we have we're a

payer but we also have a hospital system

we have lots of data

lots of clinical data

right we have a lot of like diagnoses

data procedures data if you had like a

an arm surgery like we have all of that

information

we have

data around providers like provider data

clinicians

aside from that one area that we've been

really scaling out

since last year has been

really expanding the diversity of our

data sources right so

such as social determinants of health

data so

you'll see that we have expanded our

publicly available data sets research

data

kind of state and federal data set data

sets

environmental data

grocery store data just a myriad of

different sources of data that we

believe

help to craft a more precise story

around

where our patients and members live

how they you know interact with their

environment and some of the things that

can potentially impact

their health outcomes for better or

worse right and so it's really important

for us to be very intentional about

expanding the diversity of that data

because that also feeds into the models

that we create right and

an algorithm our model is really only as

good as the data that

you feed it right so

that is you know really key and so in

order to like really tune

and refine our models

and algorithms you really

should have

better sources and more sources of data

good data

to build them yeah and with all this

patient and provider and you know health

data in general what are you guys doing

with it

yeah i mean i would say that in general

we are so committed to creating

like a remarkable health experience for

our our patients and our members and

everything that we do is geared towards

that so

all of the data that we're using all the

algorithms that we're building all the

events all analysis capabilities all of

that is geared towards you know how can

we

improve

health outcomes for our patients and our

members how can we improve their

experience right engaging with the

health care system which hasn't always

been the best experience

how can we improve that how can we

create a great health experience but how

can we also create a great experience

for providers like that's a little bit

of a newer

concept let's say like the last five

years or so is you know clinicians

providers really deserve to have a

really good experience and to have a

seamless experience when they're

interacting with patients so they can

focus

on the patients so what types of

capabilities can we create

so that they have that experience and

and they're happy and then we also

leverage that data to see how we can

reduce healthcare costs right so

are there opportunities for us to like

reduce administrative costs associated

with

obtaining care

or you know any myriad of like

interventions like how do we

really kind of

bend the cost curve so that we can at

least slow down the growth of healthcare

costs in the u.s so those are like the

four you can call the quadruple aim

those are like the four

objectives just the bad bone of

everything we do everything we do is

geared towards those

objectives

could you possibly give me like an

example of a success that you had when

you know looking at some of this data

there have been so many successes i

i mean i know we've generated we've

created capabilities around

like around our care coordination and

kind of clinical

operational excellence so being able to

free up

nurses time so that they're not spending

inordinate amounts of time reviewing

patient records

but kind of

introducing some automation into their

workflow so that they're able to spend

their time doing things that are more

productive and that are more knowledge

based rather than kind of roots and one

of the things you are mentioning in your

your job description was some of the

data analytics like software or programs

um i can tell you about the systems that

we're using um

so there's there's a few i mean you can

kind of bucket them and kind of like so

you have some analytics

platforms

you may also have like data specific

platforms and sometimes the two

are used to entertain interchangeably

like sas for example sometimes like

should be used for analytics but

oftentimes gets used for like etl stuff

you know so etl sorry um

so like shaping data like transferring

data loading it so that you can run

analytics on it like that should be done

by a date with a data

platform

um like sql or oracle or you know

something like that not necessarily sas

but that's one example where kind of

because of the pervasiveness success

throughout a lot you know legacy

organizations and all that

people are just comfortable using it for

both

but like some of the analytics tools we

use that you know maybe your audience my

audience might be familiar with you know

r

we're using r we're using um some folks

use python

we've got some folks using alteryx um

still

status obviously

from a visualization perspective

tableau power bi those are the two

main um

uh main

platforms

and then we are kind of expanding so we

have a partnership with google we're

working very closely with them so

they're going to be some tools like

bigquery which i think is more on the

data side it's using some google tools

those are kind of we're transitioning in

to make that more pervasive in the

organization but like those are like the

main

tools on the analytics side and then i

mean

on the data side you know there's denodo

which is kind of like a data

virtualization tool which

it helps you essentially be able to

integrate data without actually having

to go into

each of the disparate data sources but

you can still kind of pull it together

so i hope that's how i was able to

communicate that kind of like a

simplified way

um

and then you have like other tools like

data governance tools so

the organization

you have to

be somewhat mature and well ideally very

mature in how your data is governed

right and

that is part of making sure that we're

doing our due diligence as trusted

stewards of the data

and while a lot of that is heavily

premised on the people component so

making sure that people are

you know aware using the data

appropriately going through the proper

protocols and approvals and all that for

how

to use the data we also have some tools

to support that so i know collibra is

one tool that we're using

um to support it so those are just kind

of a few of the tools that we use for

like analytics data data governance

yeah yeah all the typical data science

tools of course

so could you talk about a challenge that

you have had to overcome in your

position

and how did you solve it i'd say that

one like practical challenge was like

when i came to highmark i had to build

my team you know i really had to create

the vision

um for my team and kind of construct

what what i felt

would help

me you know help us be successful and

deliver on the accountabilities that i

talked about in in the beginning right

and so i had to really think about

the roles

what am i working with like you know i

had a very small team um you know the

assumption of not really getting a

robust budget

so working within like those confines

and still having to construct like what

my team is set to do

the expectations

how we're going to be successful and

really get the ball rolling moving down

that path i think is is a pretty big

challenge and what kinds of um you know

professional development are you finding

like where do you find these things and

where could other data scientists look

for these opportunities

yeah i look everywhere i mean

i google

i look at professional organizations

i'm in a high mark specific

leadership development program right now

i

am

searching for like

areas where i might not be as strong as

i want to be like finance i think that's

an area where i could be a lot stronger

so are there short courses or

certifications that i can take

in that area

i look at different universities mit

um

yale

um

uuic and i also have done

several certifications right

like agile delivery how to deliver large

organizational products really um in an

agile way so safe agile certification

portfolio management these are things

that i feel like high school and college

students can explore and

you would be setting yourself on the

right path because you're understanding

how work

is done and how large products and

projects are delivered in organizations

so you're getting a head start before

you even get there you're like somewhat

familiar and the other thing i do is i

invest in

like i invested in hard skills when i

was still like hands on keyboard

writing code programming and all that i

invested in courses i remember with

tableau was like kind of new

um

now it's not new now but back then it

was like a little bit new so i was like

okay well i'm gonna

invest and take you know these like a

series of two day courses because i was

a sas you know i grew up with stats and

stata for economics and things like that

and i was like well this is like newer

so i could do like r and tableau

and just invest it paid and leveraged

you know if it is tuition reimbursement

i leveraged that um but it was a really

worthy investment especially as i was

building my career and i was still

programming and still like

doing my own analytics that was really

helpful

yeah staying up to date in data science

kid there's always something new to

learn so there's always lots of classes

available to to enhance your skills

absolutely so you mentioned

your background and the education that

you've done

in your background it's actually in

health policy administration so how did

you transition from health policy to you

know data science

yeah so

my uh background is pretty interesting i

feel like it all kind of meshes together

i mean i started off with biology and

chemistry you know back in the day i

thought

i always was interested in health care

period but i used to think that i wanted

to be a doctor

but i would say like in high school um

and after like maybe freshman year of

college i didn't want to be at the

bedside and i also didn't want the

lifestyle like i didn't want to have to

work on the weekend um i love working i

work hard but that's just not what what

i wanted and i didn't want the bedside i

wanted something broader in healthcare

so i thought about like

global health and public health like

those were the things that really

interested me

but then i liked technology i liked data

you know i

it's just something that

had always interested me from college on

but i like the data i like technology

and i like the application of data to

healthcare and i was fortunate i think

to

when i was applying to grad school to

find a program at the university of

illinois chicago that really married

data technology and public health and it

was public health informatics and it was

perfect because it like combined

everything plus

i went into my grad program

um after finishing my master's at

northwestern i went into

my grad program

right as the affordable care act was

like

being you know solidified so 2011. so it

was perfect meaningful use mars

tatum holy cow i mean it was just great

so

clearly very exciting um

and so that program

allowed me to combine all of my passions

um and in my current role i use

everything i read legislation that

relates to data and analytics and health

care and i have to interpret it and be

able to kind of conceptualize

what the technical implications are for

this legislation

on health care systems or on pairs um

i leverage my ethics background data

ethics ai ethics philosophy principles

the health aspect public health we get

to like work on cool things that are

going to improve

you know

patients and members people's health

outcomes

i love

listening to people when they're excited

about their work you could just tell how

much you like this

so you were mentioning a lot about you

know ethics in

data and how it's important for data

management and data science um

but this is an area that people don't

always focus on when people are thinking

about data they're thinking about

analytics or visualization and sometimes

the whole data ethics side gets gets

pushed to the side but it's so important

um so could you tell us you know a

little bit about why data ethics are so

important especially in the field that

you're working in

yeah i mean i think

you know thinking about data ethics it's

almost like you have to think about

just ethics in and of itself and how

it's applicable to any profession

it's really about doing the right thing

and

to me

in the data space and the analytics

space it is

there's always excitement about all

these new and cool things we can do with

data like look at these neural networks

look at these models blah blah blah like

that's all great but at the end of the

day we have to do the right thing with

this data

do the right thing by our

patients and members with this data and

that means we have to ask the questions

before we build the models yeah and we

have to ask the question before we get

to a certain point we really have to do

it as a baseline first step to say have

we thought through the ethical

implications for the use of this data

standalone but also for how it's going

to be used in the models and for the

potential outcomes that could be

generated and the unintended

consequences from the use of this data

so it just makes more sense to ask the

critical questions

and take a position beforehand

um then to try to like do it after you

messed up

and now you're on the front page

right so yeah that makes a lot of sense

yeah so it's just about it's about doing

number of different areas within

healthcare um including you know on the

research side at the hospital and now on

the payer side in health insurance

so can you tell me a little bit about

how the responsibilities are different

in these different sectors and what made

you decide to focus on the health

insurance payer side yeah

um yeah i started in clinical research

data management academic research and it

was nice um it wasn't very challenging

to me

well i would say it was obviously

challenging the beginning but then it

got really like

i got really good at it um

so i was comfortable and i don't like

that's not my preferred state of being

so i left the clinical research and went

to the hospital side so i was working um

doing clinical analytics for a large

catholic health system at the time and

since like merged and been acquired

multiple times since i was there so i

don't even know what it's called now

i would say that on the hospital side it

was a little bit slower a little bit

less technologically savvy

um a little bit more challenging a lot

of bureaucracy

but lots of opportunity for improvement

and for growth and for kind of

modernizing how

things are done on the hospital side and

that's one of the benefits i think on

that side of the house

or the industry was just

that there was so much opportunity for

growth you can really see the horizon of

opportunity in terms of improving how we

collect data how we share data

how we leverage emrs how we improve page

provider workflows and things like that

emr's electronic medical records i left

the hospital system and went over to the

vendor side which is where i think

i built more of my technical chops i

feel like that was

really valuable because it really

brought me out of just like the

analytics programming capabilities and

really i was able to work more on the

data side and you know with oracle and

redshift and all that it just expanded

my technical knowledge which i think

helps

um in terms of inputs to like

when i'm managing a team i don't

i'm not operating off of theory right i

have the practical hands-on

um

experience to at least be able to speak

to like what's being developed

um

when i left there i came to the payer

side and to be honest i think the payer

side was the most fun like it just

i don't know i just really liked it um

i was managing uh analytics consulting

and so

getting more into like

health outcomes for provider groups and

you know analytics around that and how

can we generate

insights that are going to be actionable

that these provider groups can leverage

to

better target interventions for certain

conditions

you know closed care gaps and things

like that that was really exciting to me

which is why i really liked the payer

space so this is my favorite

space really um

just because of like the connection to

both the payer and the provider and the

patient

and members and just the wealth of data

we're able to use so if one of our

listeners is interested in becoming a

data scientist um specifically you know

in the area that you work in what

resources would you suggest that they

check out looking at like user community

groups for the you know maybe the

applications that you prefer like if

you're an r enthusiast like joining

those community groups or python or

joining those groups and then looking to

see if there are conferences or

opportunities to you know network with

or work with others who have that

general shared interest i think will

help you refine

your you know your goals and like what

you truly want to do

as somebody who is a manager and who has

you know people have worked under her

and has potentially mentored

new or up and coming data scientists

what advice would you give to someone

who's interested in working as a data

scientist

um i would say you know i think it ties

in with

kind of how do you refine

your goals in terms of what you want to

do

for your career

and for that it's you know it could come

through

the different conferences right and user

groups and

you know if you're able to connect with

someone that's in the field or through

an internship or mentorship program

you know types of bridge programs

pipeline programs um some organizations

have like pipeline programs that go all

the way through like to high school

level where you could shadow someone or

intern

for a summer you know short stints

that'll help you get exposure

i think that's probably the most

important thing because

you're always going to be learning the

technology is always going to be

changing and so you want to make sure

you're getting the people aspect

you know you're getting exposure within

the field

oh nine thank you so much for joining us

today

thank you this has been fun i really

appreciate it

and to all of our listeners if you

haven't already check out our previous

podcasts they're available wherever you

listen to podcasts for more information

about starting your career as a data

scientist go to

dataanalytics.buffalostate.edu

and don't forget to subscribe so that

you get notified each time we release a

new episode of buffalo state data talk

[Music]