Big Data and Data Analytics: An Entrepreneur Perspective Detlev H. (Herb) Smaltz, PhD, FACHE, FHIMSS 17 February 2015 AUPHA Leaders Conference
Overview Background The Vendor Landscape Critical Success Factors (CSFs) for Enterprise Analytics Recommendations for Health Services Administration analytics course/program content development Q/A 2
Business Press Coverage of the Topic Organizations must excel at analytics & knowledge management in order to compete in the 21 st century 3
Analytics at Most Hospitals & Health Systems Transactional Systems Independent Data Marts Spreadsheets & DataBases End User Reporting Finance abc12349x Quality abc12349x abc12349x Mgt Eng. Marketing Supply Chain Strategy W External Systems 4 4
Background: Analytics at Most Hospitals UHC Nursing Benchmark Teletracking Census ED DM Ultipro Census Visits Workforce Management Vacancy Jane is responsible for creating the Nursing scorecard as well as providing numerous reports for Magnet certification for the HS Kronos For this Jane gathers data from many systems including Kronos and Ultipro Productive and unproductive work is defined differently in the systems so she has to massage the data to ensure that workforce management and vacancy reports are correct She accesses DM for many reports but does not trust the ED and census data She checks data from different source systems to manually scrub and reconcile the data She also downloads data from UHC for generating reports Her team spends numerous hours reconciling the anomalies in the benchmarking data
Background: Analytics at Most Hospitals UHC Nursing Benchmark X Midas Vendor Teletracking ED Censu RN satisfaction DM Census Visits Nursing Scorecard Ulti Kronos Workforce Management Pressure Ulcers Manual data logs Manual Charts Reports She also needs pressure ulcer data which she collects manually She is also responsible for providing pressure ulcer and other data to Midas She needs data from Midas for her reporting but has no access to it She gets hard copy reports from Midas which she manually integrates with her data She downloads RN satisfaction data from the vendor site for the reports and scorecards Her staff has to do lot of manual chart abstraction for value based purchasing reports She manipulates this all manually in spreadsheets to generate the reports/scorecard
SWOT Analysis Hospital & Health System Analytics Readiness Strengths Almost all hospitals/health systems have experience in basic reporting/analytics at the local/departmental level Increasing awareness of senior management that analytics is a needed new core competency Opportunities Increasing amounts of data in digital form ACO initiatives can be an excellent opportunity to proactively drive more robust analytics capabilities Leverage dyads to fill healthcare knowledge gaps of EA leadership positions Weaknesses Most have no real strategic roadmap for developing a core competency in analytics Poorly understood among senior leaders Most mired in departmentally focused analytics (islands of data and capabilities) Gaps in needed skills (data architects, data scientists) and capabilities (data stewardship) Threats Competitors that develop analytics capabilities may gain competitive advantage Increasing regulatory requirements Increasing transparency mandates Reduced reimbursements/value based purchasing demand a command of management information Competing priorities for capital $s 7
The Vendor Landscape: The Gartner Hype Cycle Gartner, 2003 8
The Vendor Landscape: Gartner Hype Cycle for Analytics Less than 5% adoption achieved 2
The Vendor Landscape Category Vendor Category Vendor Healthcare EDW Platform Solutions (including tools & apps for ACO analytics) Healthcare Analytics as a Service Domain Specific Healthcare Analytics (Point Solutions) Healthcare EMR- Centric Caradigm Intelligence Platform Healthcare Data Works Health Catalyst IBM Healthcare Data Model (CitiusTech) Oracle Healthcare Data Model Recombinant (Deloitte) Explorys Humedica Lumeris Premier Alliance Truven Analytics Suite AltaSoft ABC s Crimson Suite EPSI Lawson MedeAnalytics Medventive Midas Omincell Allscripts Sunrise Cerner PowerInsight Epic Clarity & Cogito McKesson Horizon Meditech Data Repository Siemens Decision Support Cross-Industry Development Platform Cross Industry Visualization & Exploration Tools Big Data, Cross Industry Development Platforms Dimensional Insights IBM Smarter Analytics Information Builders Microsoft BI Platform Microstrategy Oracle OBIEE SAP Sybase IQ BusinessObjects Cognos QlikView SAS SPSS Tableau Cassandra Cloudera CouchDB GNS HBase MongoDB Riak Hadoop The vendors listed are some of the most popular in the category. The lists are not exhaustive. Dale Sanders, Advisory Board Company, 2013 10
Critical Success Factors (CSF) for Enterprise Analytics Analytics Technology (e.g. EDW) Data-Driven Performance Management Culture & Capability Governance Analytic Competency Center 11
CSF New Governance Structures Needed to Support Enterprise Analytics Enterprise Data Governance Sub Committee Senior Leadership Executive Committee Enterprise Analytics Steering Committee Prioritization & Oversight Enterprise Champions/Sponsors Oversight Align projects with enterprise strategic goals Drive Requirements Guide project prioritization Monitor project progress Ensure transparency Advocate technologies Measure benefits of EA Data Governance Identify Data and KPIs owners Standardize reporting Resolve data quality issues Resolve data definition conflicts Contribute to business rules Contribute to data strategies Enforce data policies Ad hoc Task Force as Needed
CSF New Organization Models Needed to Support Enterprise Analytics Centralized Consulting Functional Corporate Corporate Corporate BI Function Function BI Function Function Function Function BI Center of Excellence Decentralized Corporate Corporate BI COE Function BI Function BI Function BI Function BI BI Analytics Group/ Dept. Analytics Project Davenport, Harris & Morrison, Harvard Business School Publishing, 2010
The Rest of the Story : Necessary Critical Success Factor CSF - Enterprise Analytics Skill Gaps Enterprise Analytics Leader Data Integration Specialist Enterprise Data Architect BI Developer/Report Writer Business/Data Analyst Database Manager Project Manager Data Scientist Few have experience leading enterprise BI teams Currently being recruited from other industries Technical skills, but often lack HC content knowledge Also currently being recruited from all industries Those with healthcare content knowledge very rare Common to recruit and train from other industries Fairly common skill, but typically tied to a particular tool or application; job growth expected for this skill Individuals that have both functional healthcare content knowledge and analytic skill; increasingly important skill Commonly available skill, though growing need for DBAs in all industries will require more of this skill Commonly available skill New emerging role; requires both healthcare content knowledge & statistics/biostatistics/comp. math skills Potential new HSA major/minor focus area 1 4
CSF: Typical Analytics Staff Growth of Medium Sized HS Enterprise Analytics Leader Data Integration Specialist 1 1 1 2-3 Enterprise Data Architect BI Developer/Report Writer Business/Data Analyst Database Manager 1 1 1 3-5 1-3 1-2 Project Manager 1-2 Data Scientist 1-2 Total FTEs ~4 Begin State 1 Yr ~11-19 2 Yr 3 Yr 4 Yr ~End State 1 5
Road Map to Becoming an Analytic Competitor Stage 1 An organization has some data and management interest in analytics Top Management Support: Full Steam Ahead Path Stage 3 Analytically Impaired Executives commit to analytics by aligning resources and setting a timetable to build a broad analytical capability Stage 4 Enterprise-wide analytics capability under development; top execs view analytic capability as a corporate priority Stage 5 CSF: Data Driven Culture Analytics as basis for competitive strategy Analytical Aspirations Analytical Companies Analytical Competitors Organization routinely reaping benefits of its enterprise-wide analytics capability & focusing on continuous analytics renewal Managerial Support: Prove it Path Stage 2 Functional management builds analytics momentum & exec s interest through application of basic analytics Competing on Analytics, Thomas Davenport & Jeanne Harris, 2007 Localized Analytics Terminal Stage: some companies analytics efforts never receive management support and stall here as a result 16
Recommended Educational Content/Program Development Areas Topic Content Target Executive Education Healthcare Enterprise Analytics Introduction of enterprise analytics as a potential basis for creating competitive advantage Overview of common problems executives encounter when moving from departmental to enterprise analytics Overview of the vendor landscape (buyer beware) Overview of total cost of ownership of an enterprise analytics program Overview of governance (how to be a good executive champion for EA) Developing: Executive Champions Better HSA Leaders Functional Leaders that want to develop a deeper grounding in healthcare analytics administration 17
Recommended Educational Content/Program Development Areas Topic Content Target Healthcare Enterprise Analytics Aligning analytics to strategy HIPAA Enterprise analytics administration Prereq: Standard HSA overview/grounding course work Co-req: Data integration concepts, data governance & stewardship concepts, data quality & validation concepts, big data concepts, etc. obtain via sister academic departments (e.g., HI, Comp Sci, Engineering, etc.) Developing: Enterprise Analytics Department Leaders Chief Data Officers (new emerging role in other industries) Chief Data Scientists 18
Recommended Educational Content/Program Development Areas Topic Content Target Healthcare Data & Information Governance Data stewardship concepts Data quality & validation concepts Data & information governance structuring alternatives Metadata management concepts (e.g., data definitions) HIPAA and secondary use of data Facilitation concepts Developing: Enterprise Analytics Department Leaders Data Enterprise Analytics Governance Leaders HSA Functional Leaders serving on data governance committees 19
Questions? 2518 Burnsed Blvd, #153 The Villages, FL 32163 614-309-3278 herb.smaltz@cioconsult.com http://www.cioconsult.com 20
Back-Up Slides 21
Recommended Educational Content/Program Development Areas Topic Content Target Healthcare Data Architecture Overview of health services administration field/industry Current topics in effective health services administration Overview of healthcare data subject areas & their nuances Overview of common problems in healthcare data standardization Prereq: Basic data architecture course from sister academic departments (e.g., HI, Comp Sci, Engineering, etc.) Developing: Healthcare Data Architects 22
An Example of a Big Data Enabled Logical EDW for Healthcare Providers User Ad hoc data sources Access Excel Sql Server End User Reporting Big Data Information Discovery Text Reports Audio/Video Social Media Sensors HR Cntrct EMR Time SIS PM Cost Cleanse Validate Standardize Transform Aggregate Geo Code Data Profile Meta Data Load Studio Endeca Server Endeca Integration Suite (Hadoop integrated) Interactive search & analysis EDW Subject Areas Structured Output Rejection Encounter Pt Sat Meds Lab Orders Data Access Portal Clinical Trial Screening Web Scorecards & Dashboards De-identified Data Multi-Dimensional Analysis & Data Mining Ad-hoc Query Text Mining, NLP Predictive Modeling SI Internal/External Systems Errors Unified data Model Transformation and Mapping Census Readmits Research Benchmarking Schooler & Smaltz, 2014 ACHE Webinar
Background: Analytics at Most Hospitals Thousands of hours of manual labor Mistrust of data in reports Ownership/control retards movement to enterprise approach Nursing Dashboard ICU Dashboard ED Dashboard CBO Dashboard Surgery Dashboard TR Dashboard Reports Dashboards Access Departmental BI Initiatives Fin DW Essbase Hyperion Diver Power Insight Manual & Semi-Manual Cognos Invision Ad hoc Signature Ad hoc Crimson Surgery Compass Lawson TSI Premier RIS BEACON eclinical Siemens Signature Cerner Various Hospital Information Systems
Background: Analytics at Most Hospitals Thousands of hours of manual labor Mistrust of data in reports Ownership/control retards movement to enterprise approach Nursing Dashboard ICU Dashboard ED Dashboard CBO Dashboard Surgery Dashboard TR Dashboard Reports Dashboards Access Departmental BI Initiatives Fin DW Essbase Hyperion Diver Power Insight Manual & Semi-Manual Cognos Invision Ad hoc Signature Ad hoc Crimson Surgery Compass In traditional companies, departments manage analytics number crunching functions select their own tools and train their own people. But that way, chaos lies. Lawson TSI Premier RIS BEACON eclinical Siemens Signature Cerner Thomas Davenport & Jeanne Harris, Competing on Analytics Various Hospital Information Systems
Background: Hype Cycle & Adoption Curves Less than 5% adoption achieved ~ 30% adoption achieved Rogers, 2003; Gartner, 2003 26