analytics excellence: current and future trends for big data utilization in pharma

17
Analytics Excellence: Current & Future Trends for Big Data Utilization in Pharma Best Practices, LLC Strategic Benchmarking Research Conducted

Upload: best-practices-llc

Post on 16-Jul-2015

232 views

Category:

Health & Medicine


0 download

TRANSCRIPT

Analytics Excellence: Current & Future Trends

for Big Data Utilization in Pharma

Best Practices, LLC Strategic Benchmarking Research Conducted

2

Table of Contents

I. Executive Summary pp. 4-19

Research Overview pp. 4

Universe of Learning pp. 5-6

Big Data Team Overview and Key Study Insights pp. 7-8

Quantitative Key Findings pp. 9-13

Qualitative Key Findings pp. 14-19

II. Defining Big Data pp. 20-26

III. Structure pp. 27-34

IV. Governance and Capabilities pp. 35-43

V. Partnerships pp. 44-50

VI. Budgets and Staff pp. 51-56

VII. Data Types and Sources pp. 57-69

VIII. Applications pp. 70-80

IX. Communicating Results pp. 81-84

X. Performance pp. 85-94

XI. About Best Practices, LLC pp. 95

Best Practices, LLC, conducted a customized study to better understand the growing influence of Big

Data in the biopharmaceutical sector and how it can impact medical, HEOR, and commercial

operations in the U.S.

Best Practices, LLC engaged 22 leaders

from 18 pharmaceutical companies

through a benchmarking survey. There

were multiple responses from 3

companies but they each represent a

different functional area (medical,

commercial, or HEOR).

Research analysts also conducted

seven deep-dive executive interviews

with selected benchmark participants.

Research

Goal

Research

Methodology

Produce reliable industry metrics on

current and future trends for Big Data

utilization across medical, commercial

and HEOR groups.

Topics Covered

Types of Big Data Projects Used to Support

Medical, Commercial and HEOR Decisions

Big Data Capabilities and Governance

Types and Value of Data Used for Big Data

Projects

Big Data Staffing and Budget Levels

Value Rating of Partnerships on Big Data

Projects

Policies and Procedures Governing Big

Data Activities

Investigate data types, data partnerships,

and staffing/budget levels companies

are using as they move to a more

analytically based approach to

commercial, HEOR & medical decisions.

Research

Overview

Research Project Objectives & Methodology

Benchmark Class:

Eighteen Companies Participated in the Benchmark Study

22 analytics, marketing and HEOR leaders from 18 different companies participated in this study.

Participants were recruited because of their presumed investment in Big Data analytics. Three companies

had multiple responses - each representing a different functional area (medical, commercial, or HEOR).

BEST PRACTICES,

®

LLC Copyright © Best Practices, LLC 5

Value of Establishing a Big Data Team or Function

The following are reasons that study participants cited for establishing Big Data capabilities.

5

Copyright © Best Practices, LLC

KEY QUALITATIVE FINDINGS

The biggest benefit is that it finally makes various departments within your company focused

on the questions they're trying to answer with data. Pharma has a lot of trouble trying to

develop questions that can be empirically answered. They don’t do that very well. It turns, in

my opinion, it turns into a lot of just comparison -- because they have a clinical trial

mentality -- and that's not what this is.

“The biggest benefit is that it finally makes various departments within your company focused on the questions they're trying to answer with data. Pharma has a lot of trouble trying to develop questions that can be empirically answered. They don’t do that very well. It turns into a lot of just comparison -- because they have a clinical trial mentality -- and that's not what this is.” – Director, HEOR

“A big advantage is you are positioned, especially in the U.S., to accommodate the changing healthcare landscape. So by using big data you are more likely to be able to address questions that emerging partners have like ACOs, who are going to vertically integrate systems, as well as to position yourself for patient information.” – Senior Director, Managed Care

“So it is going to inform this business - it's going to inform how many, when and what. It’s not going to inform why people make the decisions - which is always going to be in the qualitative market research arena - but it’s certainly going to help you narrow down where you want to focus your qualitative research. As we move forward over the years, I think we’re going to see a lot less binary quantitative research and more use of this type of data to get those types of answers.” - Associate Director, Global Market Research

Help Focus

on Critical

Questions

Attract

Partners

Market

Research

Aid

BEST PRACTICES,

®

LLC Copyright © Best Practices, LLC 6

Most Have Centralized/Dedicated Big Data Team or Function

A majority of study participants said they have a centralized/dedicated team or function to support

Big Data projects.

N=19

Q: Do you have a centralized/ dedicated group of individuals to support Big Data projects (i.e. Big Data team or function)?

Yes 53%

No 47%

Dedicated Big Data Team (Total Benchmark Class)

BEST PRACTICES,

®

LLC Copyright © Best Practices, LLC 7

Benefits of a Centralized Approach to Big Data Team

What do you think are the most significant benefits of a centralized Big Data team approach?

“One is really sharing learnings

amongst everyone. I think

there’s a feeling that with one

team you can be more objective

in reflecting what’s the interest of

the organization versus what’s

the interest of your specific

division. I mean we’re part of

marketing now, for instance. So

that can be a fine line at times

when we say, hey, that program

did not return great ROI.”

--Associate Director

Shared

Goal

One

Information

Source For

Global

“If you're global, it's

probably more

valuable for your core

affiliates because

you're centralized in

supplying information

for them and so they

don’t have to go look

for it on their own.”

--Sr. Director,

Managed Care

Proactive

Strategy

Approach “With a centralized

team we can plan

strategically and

proactively on

product strategy

using real world data,

not just waiting

patiently for the

question to come up

and then start to run

analysis.” – Senior

Scientific Analyst

“I think [it’s important]

from a signaling

perspective for talent

development and

also recruiting.

Because if there is a

centralized data

function, the

company is invested”

--Associate Director

Talent

Development

& Recruitment

BEST PRACTICES,

®

LLC Copyright © Best Practices, LLC 8

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

Rules about what analysis commercial functions can perform using…

Rules about what analysis medical functions can perform with Big Data

Rules on Big Data insights-sharing between commercial and medical

Rules governing disclosure of findings to public

Rules on proactive vs. reactive use of insights from Big Data

Rules on disclosure from regulatory perspective

Rules on publishing

Policies establishing clear ownership for various data types across the …

Policies and procedures for accessing data (e.g., who can see what)

Policies governing protecting identification/ de-identification of patient…

Policies governing clear ownership of IP generated through a…

Policies regarding review/ approval of research protocols

Prevalence of Data Governance Policies (Total Benchmark Class)

N=16

Q: Which of the following policies and procedures are in place at your company to govern Big Data activities?

Most Participants Have Range of Policies Governing Big Data Use

At least half of the study participants use all 12 of the listed policies governing Big Data use. More

than 70 % have policies related to disclosure in place for Big Data activities. These range from

policies protecting the identity of patients to rules around publishing study results.

BEST PRACTICES,

®

LLC Copyright © Best Practices, LLC 9

0.000%

0.020%

0.040%

0.060%

0.080%

0.100%

0.120%

0.140%

Current Future

Non-labor Budget*

Data acquisition Data infrastructure Data analytics Data dissemination Other categories

Data Management is Lion’s Share of Big Data Budget

For participants, data management (data acquisition and infrastructure) represents about 70% of their

Big Data non-labor budget and most of future cost growth.

N=18

Economies of Scale

“It is really cost saving

to the company to have

a centralized team …

we have multiple

products and a lot of

products can share the

same database.”

- Analytics Lead

Data Management

Q: What is the approximate current budget range for each of the following Big Data function spending categories?

*Weighted average based on

midpoints of spending buckets.

Spending set to

increase across

the board

The highest areas for spending

are also the weakest areas for

performance in data management

BEST PRACTICES,

®

LLC Copyright © Best Practices, LLC 10

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Point of Sale (POS) Claims Electronic HealthRecords (EHR) /

Electronic MedicalRecords (EMR) /

HIE (HealthInformationExchanges)

ePrescription /pharmacy fulfillment

Wholesalers /Group Purchasing

Organizations(GPOs)

Government (e.g.,cost data)

Credit card

Impact of Transactional Data Sources (Total Benchmark Class)

Highly impactful Somewhat impactful Not impactful Not used

Transactional Data: Majority Say Claims, EMR Most Valuable

The only types of transactional data that a majority of study participants said were highly impactful

or valuable for Big Data Studies were claims (70%) and Electronic Medical Records (EMR – 53%).

N=19

Q: How impactful (or valuable) has each of the following types of transactional data sources proven to be?

BEST PRACTICES,

®

LLC Copyright © Best Practices, LLC 11

Partnerships Are Potentially Valuable Sources of Data

The pharma industry sees data aggregators, payers, and government agencies as key partners,

but still purchases 40% of its data. While partnerships offer another avenue for obtaining data,

often these arrangements still require payment for the data.

0% 20% 40% 60% 80% 100%

Health Information Exchanges(HIEs)

Consulting companies/ Analyticsoutsourcing companies (e.g.,…

Academic institutes

Contract Research Organizations(CROs)

Academic medical centers

Accountable Care Organizations(ACOs)

Medical Groups/ Health CareProviders (HCPs), Group…

Government groups (e.g., CMS,HHS, NIH)

Health Systems/ Hospitals/Integrated Health Networks…

Data aggregators

Health plans/payers

Highly impactful Somewhat impactful Not impactful Not used

N=18

Q: What percentage of your data comes from each of the following sources? Q: Which of following partners are most valuable for Big Data projects?

Purchased 44%

Partnerships 15%

Interally Generated

37%

Other 4%

Sources of Data

BEST PRACTICES,

®

LLC Copyright © Best Practices, LLC 12

Predictive Analytics an Untapped Game Changer

Despite relative confidence in analytics capabilities, a majority of study types are retrospective

-- based on econometric models that rely only on observations from the same period as the

dependent variable. Predictive models could give pharma a more accurate forward view.

“I work a lot on products, and we were

able to find an opportunity for one product

that didn’t even exist.

“We looked at different sets of patients

and we modeled algorithmically the

patients and put them in different settings,

mathematically. We were able to determine

that there was opportunity for a product

that didn't exist.

“I had the very good fortune to work with

the head of scientific affairs who actually

got what we were doing, and they

generated a product which turned out to

be a big seller for them. And it was unique

because nobody had ever used the data

to tell them where the opportunity was.”

— Director, HEOR

Predictive, 26%

Retrospective, 74%

Q: Please estimate the percentage of your Big Data projects and studies that fall into

each of the following two categories:

N=21

BEST PRACTICES,

®

LLC Copyright © Best Practices, LLC 13

Drug Development Decisions Being Influenced By Analytics

The percentage of techniques covered in the survey employed across a variety of questions is charted

below, showing the relative sophistication of the benchmark class.

Q: Percentage of Big Data projects currently used to support these medical decisions:

“There are a multitude of benefits. One is

you can get information earlier to make

decisions about what you're investing in.

So you're basically going to use these to

figure out what's the burden of information

you need to have to bring a product to

market. You're going to start pushing back

that cycle of waiting till Phase 3 for things

to fail and you're going to earlier on figure

it out.”

– Senior Director, Managed Care

N=21

Mo

re S

op

his

ticate

d

Le

ss S

op

his

ticate

d

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Targets for Drug DevelopmentDecisions

Other Drug Development/Submission Decisions

BEST PRACTICES,

®

LLC Copyright © Best Practices, LLC 14

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Impact on Data Dissemination Targets (Total Benchmark Class)

Highly impactful Somewhat impactful Not impactful Not used

Data Targets: Internal Functions are impacted most by Data Analysis

A majority of study participants felt that internal company functions – specifically commercial,

medical and development – were the target audiences impacted most by data analysis from internal

analytics groups.

N=18

Q: Who are the most impactful target audiences for your data analysis?

BEST PRACTICES,

®

LLC Copyright © Best Practices, LLC 15

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Governance(strategy &

policy)

System design:infrastructure,architecture &

analytics

Dataconnectivity

Dataidentification &

acquisition

Dataorganization &

storage

Data cleansing/validation

Datadissemination

Data analysis& visualization

Datainterpretation/

insights

Mature Somewhat Mature New Capability None/Don't Know

Pharma’s Self-Assessment of Big Data Capabilities is a

Combination of Highs and Lows Most partners say they can visualize, interpret, and disseminate data well, but lack a strategic

vision and the technical horsepower to grow.

Q: Please indicate the level of maturity within your organization for each of the following capability types:

N=17 Low = Strategy + Tech HP High=Analysis + Distribution

BEST PRACTICES,

®

LLC Copyright © Best Practices, LLC 16

Universal Accessibility and Unification of Data Is Beneficial

“I think all the data should be in one

place. So anyone who is interested

knows where to go to find out what’s

there. That's kind of the basic building

block.

Then the next level I think is where it’s

linkable, we should link it. So if you

know what you have, then you go,

‘Wait a minute, these two can link.’

If they’re in different places or if you

don’t know, if you can’t see them all,

then you don’t know which ones kind

of link. And part of the value of Big

Data is when you start linking data

sources.”

- Director, Customer Data

BEST PRACTICES,

®

LLC Copyright © Best Practices, LLC 17

Best Practices, LLC is a research and consulting firm that conducts work

based on the simple yet profound principle that organizations can chart a

course to superior economic performance by studying the best business

practices, operating tactics, and winning strategies of world-class companies.

Best Practices, LLC 6350 Quadrangle Drive, Suite 200

Chapel Hill, NC 27517

www.best-in-class.com

About Best Practices, LLC