Analytics Staffing Models of Health Systems That Compete Well Using Data

41
we create thinking data® Analytics Staffing Models Health Systems That Compete Well Using Data Greg Nelson, CPHIMS, MMCi Monica Horvath, PhD

Transcript of Analytics Staffing Models of Health Systems That Compete Well Using Data

Page 1: Analytics Staffing Models of Health Systems That Compete Well Using Data

we create thinking data®

Analytics Staffing ModelsHealth Systems That Compete Well Using Data

Greg Nelson, CPHIMS, MMCi Monica Horvath, PhD

Page 2: Analytics Staffing Models of Health Systems That Compete Well Using Data

As a result of this “perfect storm”, healthcare organizations that continually improve how they manage data, develop insights and operationalize analytics are best poised to succeed in the

new healthcare economy.

§ Transition from pay for service to pay for performance

§ Increased competition (including impact of ratings, social media, healthcare marketing, pricing, transparency)

§ Expectations of integrative and accountable care§ Employer based strategies for improving health

§ Technology modernization in healthcare (EHR, Imaging, Cloud, Mobile, Security)

§ Data challenges for healthcare (volume, variety, velocity, veracity) and transparency

§ Ubiquitous access to information (patients, physicians) and the rise of the “data scientist”

§ Growth in patient involvement§ Increase in payer negotiating power§ Increased focus on product economic value

propositions (e.g., formulary, bundled payments, employer contracts)

Maturing & Evolving Markets, More Competition

Increasing Consumerism, Payer Consolidation, Changing Economic

Value

Tension between “Service” and “Information” economy

Need for integrating analytics competency into

the value chain

Analytics is a Key Competency Needed to Survive Future Challenges

Page 3: Analytics Staffing Models of Health Systems That Compete Well Using Data

Inefficient,inconsistentversionsofthetruth

Foundationofdataandtechnology

Relatingandorganizingthecoredata

Efficient,consistentproduction

Efficient,consistentproductionandagility

Measuring&managingevidence-basedcare

Takingfinancialrisk

Takingmorefinancialrisk&managingit

Contractingfor&managinghealth

Source: Adapted from Health Catalyst, 2014

Analytics Maturity

Page 4: Analytics Staffing Models of Health Systems That Compete Well Using Data

Our Perspective around Analytics Maturity

Analytic Organizational Models

Discussion and Implications

1

2

3

Page 5: Analytics Staffing Models of Health Systems That Compete Well Using Data

Our Perspective1

Page 6: Analytics Staffing Models of Health Systems That Compete Well Using Data

Understanding Health Analytics Excellence

• 5 core competency areas• Focus on knowledge, skills,

abilities, and behaviors to create a learning health organization

• No ladder or implied hierarchy of core competencies

• Helps us conceptualize what it means to be ‘the best’

Page 7: Analytics Staffing Models of Health Systems That Compete Well Using Data

Information Management &

Reporting

Predictive & Prescriptive

Analytics

Information Security &

Data Privacy

Engagement with

Organizational Strategy

Data Management

& Warehousing

• Data Warehousing• Enterprise Data Management• Data Quality• Data Model• Data Sources• Data Currency• Data Capture• Data Integration

• Metadata• Information Governance• Information Development• Enterprise Content Management• Enterprise Search• Data trustworthiness• Master Data Management

• Analytics• User Profile (data democracy)• Adoption Profile (operationalization

of data and metrics)• Business Intelligence• Reporting• Analytical tools• Aggregation & Measurement• Performance & Improvement

• Data Privacy• Data Security• Anonymization• De-identification• Data Transparency

• Information Strategy• Information Value• Metrics• Initiative mapping• Culture• User adoption• Change management• Internal consulting• Collaboration• Analytic services

Perspectives of Analytic Maturity

Page 8: Analytics Staffing Models of Health Systems That Compete Well Using Data
Page 9: Analytics Staffing Models of Health Systems That Compete Well Using Data

Source: Accenture, Counting on Analytical Talent, 2010

Analytics Project

Analytics GroupSolid line indicates a direct line of authorityDotted line indicates a a partial lien of authority or funding

Corporate

Centralized Decentralized

Function

Corporate Corporate Corporate Corporate

Center of excellence Consulting Functional

Business Unit

Function Business Unit

Function Business Unit

Function Business Unit

Function Business Unit

COECOE

Analytic Staffing Models

Page 10: Analytics Staffing Models of Health Systems That Compete Well Using Data

Analytics Organizations2

Page 11: Analytics Staffing Models of Health Systems That Compete Well Using Data

Centralized

Carolinas HealthcareUnity PointHenry FordAllinaHealth Alliance HospitalHCAMemorial Hermann

Physician NetworkVirginia Commonwealth

Univ. Health

Decentralized

Partner’s Health

Kaiser Permanente

Center of Excellence

AtriusChildren's

Hospital Wisconsin

Center of Excellence Heavy (functional)

UPMCIntermountain HealthUniv. of MichiganMayo ClinicGeisingerYale New HavenMount SinaiPenn Medicine

Center of Excellence Light (consulting)

Cleveland ClinicSeattle Children’sDuke Medicine

• Deep history of IT innovation / much 'best of breed' effort

• Active mergers and acquisition history• Large expectation of business unit autonomy• Large political stake in research• Many established yet siloed analysis teams

• Very strong executive leadership demanding high strategic alignment

• Analytics may be relatively new• Planning on a high degree of self service

Analytic Staffing Models

Page 12: Analytics Staffing Models of Health Systems That Compete Well Using Data

Top-down approach to managing organizational analytics

• Can match skills to requests• Enterprise view into analytics

activities• Can measure ROI

4

Centralized Staffing Models

Page 13: Analytics Staffing Models of Health Systems That Compete Well Using Data

Analystcommunicationincreases

Strategicalignmentisensured

Duplicateworkavoided

Datamanagementtimesarereduced

“Ithinkthebiggestlessonlearnedisthatpriortocreatingthisnewinfrastructure,workwasbeingdoneinaverysiloedfashionacrosstheorganization,butnowthatwehavecommittedtoactuallyconsolidatinganalyticsinonearea,enterprise-wide,thingshaveworkedmuchmoreefficiently.Focusingonaneffortasbigasthisfromacentralizedperspectiveisreallykey.”

VicePresidentforAdvancedAnalyticsCarolinasHealthCareSystem

Consistenttools,training,process

Shared Capabilities of Centralization

Page 14: Analytics Staffing Models of Health Systems That Compete Well Using Data

Case: University of Utah Health Sciences

Strong commitment to sharing4 teaching hospitals

10 community clinics$1.04B operating revenue (2013)

Page 15: Analytics Staffing Models of Health Systems That Compete Well Using Data

Value-Driven Outcomes System

Multidisciplinary team • Quality• Biomedical informatics• IT

• Finance

Costing datamart• Covers all aspects of clinic activity• Granular to 1 minute time intervals• Gold standard = general ledger (total

direct costs must be within 20%)

“We determined who we needed to help us design the system, and sequestered them. We took some our

brightest minds and found them an open office space, gave them a three-month task.. We did not hire any

additional people.” CMQO Robert Pendleton, MD

University of Utah Health Sciences

• Execution– Pilot 6 months: 8-16 people, 60-100% effort– Subsequent work: 20 people, 20-60% effort

• Results– Physicians leading improvement initiatives

based on data– Direct costs per medical condition reduced 20-

30%– Inpatient lab utilization reduced 20%

Page 16: Analytics Staffing Models of Health Systems That Compete Well Using Data

Centralization Challenges

Maintain alignment to departmental needs

Executive ‘just-do-it’ Analytics

Extra focus needed on communication

Culture of independence for departments, medical groups

Large researcher presence

May stifle innovation

Page 17: Analytics Staffing Models of Health Systems That Compete Well Using Data

Decentralized Staffing ModelFor organizations that have a deep history of data analysis expertise

Page 18: Analytics Staffing Models of Health Systems That Compete Well Using Data

Shared Capabilities of Decentralization

Suits complex organizationalstructures

Fosters deep knowledge of the business

Tool agnosticPermits business arms to define their analytic approaches

“We have an environmental landscape that is challenging….We have absolutely endorsed a decentralized analytics model…. My customers are the analytics teams across the health system. Partners is organized as a set of individually operating hospitals and physician groups…. To be able to consolidate analytics would be an impossible, impossible task.” - Associate Director Enterprise Data Warehousing, Partners Healthcare

Page 19: Analytics Staffing Models of Health Systems That Compete Well Using Data

Innovation Thrives Among Decentralized Models

• Sells the ‘Geisinger’ experience• Consulting, analytics, data

management• Analytics for the volume to value

transition• Care transformation suggestions• 23 provider customers and

counting

• Text mining extracts intel from EHR

• Adds risk scores on top of EHR records

• Venture-backed launch in 2013

• Quick to provide Ebola app for users in Nov 2014

Page 20: Analytics Staffing Models of Health Systems That Compete Well Using Data

Case: Innovative Decentralization Enables Agility

• Partners Healthcare is a large organization in the Boston area with a wide variety of clinical informatics applications

• Text mining extracts intelligence from EHR-sourced data lakes• Adds risk scores on top of EHR records• Venture-backed launch in 2013• Quick to provide Ebola app for users in Nov 2014

Page 21: Analytics Staffing Models of Health Systems That Compete Well Using Data

ReportContent

Consistency

Training & process

consistency

Analyst Communication

Resource Hoarding

Tribal Knowledge

Change Management

Strategic Alignment

Tool Bloat

Decentralization Challenges

.. but these are manageable if recognized and addressed head-on!

Page 22: Analytics Staffing Models of Health Systems That Compete Well Using Data

Center of Excellence Staffing ModelBalancecentralizationofserviceswithdepartmentalautonomy

Ifwedon’tmanage it,allwewilldoisacceleratetheability tomakebaddecisions-- Director,ClevelandClinic

Page 23: Analytics Staffing Models of Health Systems That Compete Well Using Data

“Our medical groups have their own analysts who have their own fiefdoms, but we're all hitting on the same shared database” -- Joe Kimura, CMO, Atrius Health

Shared Capabilities of a Center of Excellence

Maintains strategic alignment, even for

large orgs

Blend resources, not budgets

An analytics community is created

Consistent training, tools, methods can be evangelized

Best practices can be socialized to the enterprise

Page 24: Analytics Staffing Models of Health Systems That Compete Well Using Data

• 11 hospitals + sites in FL, Las Vegas, Dubai• 5.1M visits / 157K admissions annually (2013)

• HIMSS Stage 7 Ambulatory Award (Dec 2014) • 6.5B operating revenue (2013)

Case:

Volume to Value

• Bundled payment arrangements negotiated

• Lowes, Boeing have contracts for cardiovascular procedures

Patient Engagement

• Expanded patient surveys

• Analytics uncovered unexpected satisfaction drivers

• Changes made, patients happier– 16% boost in high ratings

Utilization Reduction

• Analytics to pinpoint overused lab tests

• Developed alerting to block orders

• IP: $244K cost avoidance

• OP: $1.72M revenue avoidance

Analytics to transform the business

“One must take risk and experience failure to foster innovation; without a baseline failure rate, innovation isn’t happening” – Toby Cosgrove, MD, CEO

Page 25: Analytics Staffing Models of Health Systems That Compete Well Using Data

Role confusionEmbedded vs enterprise

analysts

Prioritization confusionEnterprise or

departments?

Vendor bloatTool usage hard to

enforce

CultureTakes much planning and strong leadership

SLAs essentialDefine what the enterprise team delivers versus departmental teams

Analytics stylesVended vs internal development? Modeling strategy?

Challenges with a Center of Excellence

Analytics

COE

Dep

artm

ents

Page 26: Analytics Staffing Models of Health Systems That Compete Well Using Data

Discussion and Implications3

Page 27: Analytics Staffing Models of Health Systems That Compete Well Using Data

All or Nothing?

Page 28: Analytics Staffing Models of Health Systems That Compete Well Using Data

Commonly Centralized Data Domains

Traits•Enterprise-facing activities•Analytic platforms•Expensive data•Controversial data

Examples“Core clinicals”Patient satisfactionData brokers / consumer spendingGeospatial data packsPopulation health platforms / HIE dataPatient-reported outcomesActivity-based costingBiobank metadata

Analystcommunicationincreases

Strategicalignmentisensured

Duplicateworkavoided

Datamanagementtimesarereduced

Consistenttools,training,process

Why?

Page 29: Analytics Staffing Models of Health Systems That Compete Well Using Data

Creative Operationalization of Predictive Models Takes Advantage of Centrally Managed Resources

Readmission Risk Score• Every IP gets a risk score that is update

daily• 30 highly predictive variables• Risk scores are calculated using

warehoused data and uploaded to EHR and patient census dashboard

• Transition coaches and transition ‘conferences’ performed for most difficult patients

• At any given time, 10+ interventions are being tested

Readmission Risk Score• Daily reports divide patients into high

(~35%), medium (~16%), and low (~12%) risk groups

• Pushed into IP records as well to primary care MDs after discharge

MRSA risk model• $500K saved annually by only testing

half of newly admitted patients for MRSA

Undiagnosed hypertension model• 50% predictive value; finds 50 new

hypertensive patients per month

• Alerts pop up in OP EHR

Page 30: Analytics Staffing Models of Health Systems That Compete Well Using Data

Case: Use of Data Brokers

UPMC Health PlanCarolinas Health System

“..people with no children in the home who make less than $50,000 a year are more likely to use the ER rather than a private doctor” – Pam Peele,

Chief Analytics Officer

“What we are looking to find are people before they end up in

trouble,” -- Michael Dulin, Chief Clinical Officer for Analytics and

Outcomes Research

Editorial: the problem with these data is the sensitivity– is your doctor spying on you? Responding to this should be part of the data strategy.

Page 31: Analytics Staffing Models of Health Systems That Compete Well Using Data

Case: EDW Geospatial Infrastructure Development at Duke Medicine

Copyright 2016 ThotWave

Technologies, LLC.

32

Page 32: Analytics Staffing Models of Health Systems That Compete Well Using Data

• Sentiment analysis of customer experiences

• Health system brand management

• Mayo Clinic offers a social media residency

• Social media identities for patients

• Facebook app to remind transplant patients to take meds (University Iowa)

• Understand linguistic phrasings for conditions (UniversityIndiana School of Nursing)

• Emory University: 20TB of vitals data over 3 years; Hadoop technology to analyze 100,000 real time data points a second

• Predicts risk of stroke, heart attack, or other serious conditions at the bedside

Case: Use of Real-time Data to Add Context to Healthcare

Page 33: Analytics Staffing Models of Health Systems That Compete Well Using Data

Omics datais

managedbythe

DataAnalytics

Center

Omics dataisnot

integratedwiththe

PennDataStore

thatserves

enterprise health

systemneeds

Case: PennOmics is the Translational Data Warehouse

Copyright 2016 ThotWaveTechnologies,

LLC.

34

Page 34: Analytics Staffing Models of Health Systems That Compete Well Using Data

Shared Purpose

2

Build Optimize1Explore 4 Measure Impact

Prioritize

5

Demonstrate

3

Create Knowledge

6

Analytics is a Journey Not a Destination

Page 35: Analytics Staffing Models of Health Systems That Compete Well Using Data

Process automated

Process improved

Self-service platform release

Plan to drive usage

Deep mgmt. structure

Fewer individual contributors

Pilot project shows value

Enterprise rollout

LEAN analysis

Long term plan for continuing to measure

Uses existing tool licenses

Real or self-imposed vendor lock-in

Self-service data access permitted

Data literacy

Data governance

Priority of core master data

Great predictive capacity

Impractical plan to act on results

Traps That Can Create Misalignment

Page 36: Analytics Staffing Models of Health Systems That Compete Well Using Data

Rethink Analytics Talent

Lorem ipsum dolor sit amet, consectetur adipiscing elit.

Integer mollis vehicula ligula.

WorldwideDevelop analytics concierge services

New Role: Analytics Strategist

MindsetEmpower analytics ambassadors

SkillsetNon-traditional candidates can transform

ToolsetSupport modern, collaborative, interoperable tools

Staffing practices of the past will be ineffective in the future

Unicorns don’t exist but horses with party hats

do….

Educate Human Resources

Entice them with social impact, challenge them

with the puzzle

Ensure Meaningful Work

“There are not too many people like me in health care… But there are plenty of people who have the necessary statistical knowledge and background. One of the things that is most useful is having the experience of working with a lot of data.” -- Allina’s Sr. Statistician, Jason Haupt, PhD, a particle physicist

Page 37: Analytics Staffing Models of Health Systems That Compete Well Using Data

Shared savings, bundled payment, ACOs, PCMH, population health management

At Risk Contracting

Integrative medicine, service line design, care redesign, patient engagement and commitment

Care Transformation

Design and execution of experiments, innovation, labs that help customers explore

Performance Improvement

Who is most affected?

Page 38: Analytics Staffing Models of Health Systems That Compete Well Using Data

Vision/ AspirationAnalytics Brand IdentityStrategic Goals and Milestones

Strategic PlanningAnalytics Lifecycle“Work-source” effectivenessAdvisory ServicesMentoring

ExecutionTransformational Change ManagementData and Analytics Literacy

Learning Health Organization Roadmap

Change Management

ThotWave: What we do

Page 39: Analytics Staffing Models of Health Systems That Compete Well Using Data

ChangeManagementSupport

30DayPrescription

ResearchSummaries

CuratedContent

eLearning HandsonWorkshops

Classroom

Competency-Based

Literacy

StrategicPlanningforAnalytics

Our Philosophy: The ”T” in Transformation is NOT Technology

Page 40: Analytics Staffing Models of Health Systems That Compete Well Using Data

@healthcare_bi

linkedin.com/in/thotwave

[email protected]

919.931.4736

Contact

www.thotwave.com

Gregory S. Nelson, MMCi, CPHIMSThotWave Technologies, LLC.

Copyright 2016 ThotWaveTechnologies,

LLC.

41

Page 41: Analytics Staffing Models of Health Systems That Compete Well Using Data

we are

we create thinking data®