Developing an analytics strategy & roadmap Paula Edwards, PhD pedwards@himformatics.com Nov 15, 2012
Topics Why develop a strategic plan? Key components of an analytics strategic plan Typical planning process Key stakeholder to involve Developing an implementation roadmap 2
Why analytics in healthcare? 1. Increase speed of decision making 2. Increase confidence in the decisions 3. It is at the crux of identifying opportunities and measuring progress McKinsey, 2012: If US healthcare used data to drive efficiency and quality could see more than $300 billion in value annually. 2/3 from reducing expenditures by ~8%. Gartner, 2007: Only 7% of data is used for analysis in hospitals InformationWeek, 2012: 52% organizations have completed or are working on major BI projects over the next 24 months 3
Why analytics in healthcare? Information Week Healthcare Priorities, 2012 Survey Share Data with More Than One Provider Personalized Medicine Improve Collaboration Among Clinicians and Patients Improve Collaboration Among Clinicians Reduce Costs Increase Clinician Efficiency Improve Care Manage Digital Patient Data Meet Regulatory Requirements 0 0.5 1 1.5 2 2.5 3 3.5 4 4.5 4
Around the industry There are pioneers and leaders we can learn from Intermountain, Geisenger, Mayo, Partners, others They have shown us analytics is a journey, not a destination Their efforts have evolved over many years Their toolsets and their staff have grown and evolved with the organization s appetite for information and analytics skillsets Achieving leading, enterprise-class analytics capabilities is no small undertaking e.g., UPMC recently announced a 5-yr, $100M enterprise analytics initiative to enable personalized medicine 5
Why develop a strategic plan? Would you tell me, please, which way I ought to go from here? That depends a good deal on where you want to get to Source: Alice's Adventures in Wonderland 6
Why do you need an analytics strategic plan? Analytics, data warehouses (DW) and clinical & business intelligence (C&BI) are complex There are multiple technical components and processes They impact multiple parts of the organization Data quality is always an issue There are competing priorities It is easy for these efforts to fail - spend scarce capital and gain little value Being successful requires A shared vision Focus on the organization's strategic goals and priorities Using an incremental approach to build toward the long-term vision 7
Why do you need an analytics strategic plan? 8
A strategic plan helps answer critical questions Where do we need to go with our analytics capabilities? In the next 12-months? In the next 3-5 years? How are we currently doing? Where to we need to improve our people, processes, and technology to get where we need to be? Given limited organizational resources (both time & money), where should we start? What are the most urgent needs? What projects have the most strategic value in the future? What resources will it take to get where we need to be? How do resource constraints effect when & how we tackle specific projects/tasks? 9
Key components of an analytics strategic plan Vision & goals Gap analysis People: staff, skills, organization structure Process: project governance, data governance, support Technology: data management, data quality, information delivery High-level project needs & use cases Criteria for prioritizing projects What data is needed to support the identified projects? Cultural barriers and challenges Recommendations Cost/Benefit Analysis Roadmap 10
Typical strategic planning process Needs Assessment Stakeholder education Vision & goals session Stakeholder interviews Identify use cases & data needs Gap Analysis Current State & Strategic Information Systems Plan alignment Technical assessment Org. structure, staffing, and skills assessment Data/project governance, information management assessment Road mapping Cost estimates Benefits/ROI assessment Identify project/use case dependencies & constraints Develop roadmap Communication Develop recommendations Review & revise Roadmap with key stakeholders Assemble resources for implementing Roadmap 11
Involve key stakeholders throughout the process Inpatient Ambulatory Ancillaries Service Lines Health Services Clinical Clinical Operations Research Executives Critical Success Factor: Collaboration of Leaders & Analysts Quality Business Operations Information Technology Revenue Cycle Supply Chain H.R. Finance Data Warehouse Report Writers PMO/Architects Apps Teams 12
Progression of Analytics & BI Where are you now? Where do you want to go? Competitive advantage Optimization Predictive modeling Forecasting/extrapolation Statistical analysis Alerts What is the best that can happen? What will happen next? What if these trends continue? Why is this happening? What actions are needed? Analytics Query/drill down Ad hoc reports What exactly is the problem? How many, how often, where? Access and reporting Standard reports What happened? Degree of Intelligence Source: Davenport, Thomas H and Jeanne G. Harris, Competing on Analytics; The New Science of Winning, Harvard Business School Publishing Corporation 2007. 13
Where are you now? Where do you want to go? Data as a strategic asset, realtime alerts &integrated analytics Data quality, KPI, business performance management, scorecards Master data, enterprise data management, data stewardship, enterprise metadata Data quality, data consistency, data security, privacy Raw data, spreadsheets, databases, reports 14 Adapted from Villar & Kushner (2010). A Framework to Map & Grow Data Strategy, Information Management, Nov/Dec 2010.
Where are you now? Where do you want to go? Healthcare Analytics Adoption Model Level 7 Level 6 Level 5 Level 4 Level 3 Level 2 Level 1 Level 0 Personalized medicine: Integration of genomic, familial, text, and patient self-reported data used for predictive modeling, preventive care and wellness management. Waste elimination: The focus is in maximizing quality and minimizing cost of production. Complex modeling and forecasting is readily available. Data from ACO partners and claims is integrated with patient specific costing and claims data and used for identification and elimination of variability & waste in the complete, endto-end care process. Cultural data literacy: Permanent technical and clinical improvement teams in-place for top 10 conditions; at least 60% of employees have access to KPIs actionable to their role. Analytics are embedded in the EMR to affect clinical & financial improvements at the point of care. Evidenced-based population management: Patient registries for at least the top 10 patient conditions within the organization, supporting acute & chronic condition mgmt; measurement of clinical guideline usage; and clinical research Automated external reporting: Regulatory and other reports such as Value-based Purchasing, PQRS, MU; accreditation/regulatory such as JCAHO, ACC, STS, HEDIS. Adherence to industry standard vocabularies are required at this Level. Automated internal reporting: Key performance indicators, highly interactive dashboards and reports that allow for effective hospital and clinic management and business modeling are available. Vocabulary, metadata, & data governance: Searchable metadata repository, core data elements linked with standardized naming and data types. Data governance & stewardship processes in place. Core data integration: As a minimum EMR Level 3 data, Revenue Cycle, Financial, Costing, Supply Chain, and Patient Experience integrated into a single data warehouse. Source: Sanders, D. (2012) A model for measuring industry-wide adoption and capability of healthcare analytics and data warehousing in the USA, ElectronicHealthcare, Vol. 11 No. 2, p. e5-e6. 15
Gap Analysis: How do your capabilities compare? Dimension Strategic & Operating Plan Culture of Analytics Data Governance Data Quality Data Capture Data Accessibility Information Delivery Support Services Technical Architecture Achievement Level Poor Average Leader Typical Challenges No integrated plan; project by project funding & design Analytics skills concentrated in a small set of people, limited senior leader sponsorship, disparate acceptance of data-driven decision-making and planning Limited standards, no data dictionary, data linage unknown Conflicting numbers, incomplete data, significant delays in data availability, varied levels of standardized terminology Little standardization in key master data and underlying terminologies, disparate edits Silos of data, manual data integration required, unstructured data 'trapped' Highly manual; Requires 'expert' users, inconsistent amd siloed tools Uncoordinated and inconsistent support resources, tools and SLA's across silos Lack of standard platforms and maintaining version and patch levels 16
Gap analysis food for thought Executive Sponsor Business/ Clinical Data Stewards Super Users Subject Matter Experts Executive Oversight Analytics Operations Subject Workgroups IS Data Managers Project Managers Infrastructure Support DBAs Do your current org structure and processes facilitate collaboration between IS & the business/clinical personnel? Are you staffed to enable you to appropriately support business/clinical users (now & in the future)? 17
Gap analysis food for thought Do you have defined processes for key areas of data governance? Are they used? Are they effective? Data Governance Change Management Data Integration Data Quality Master Data Management Data Distribution Metadata Management Data Stewardship Security/ Privacy 18
Needs assessment tools For each use case identified, which user groups do they support? What is the benefit? How do they support business & clinical priorities? Use cases Users Benefit Needs Matrix Data source Use Case Data Quality Source Matrix For each use case identified, which what data is needed to support it? For each data source, how is the data quality? How hard will it be to clean up? 19
Other questions to consider Do your core vendors have analytics solutions? Should you standardize on one as the Enterprise analytics solution? How do we avoid creating new data silos? What information architecture is needed to integrate data across the enterprise? 20
Other questions to consider What data governance processes and information management tools are needed to Improve data quality? Standardize on common metrics, definitions, and master data? What initial use case(s) are the best place to start? Data is available, lower technical complexity, high value to the organization What are your big data needs now & in the future? 21
Developing an implementation roadmap
Best practices for implementation Use an incremental, project-based approach to build toward the long-term vision Address analytics foundational needs in parallel with initial projects Data architecture Standardization, data governance Processes, roles, & responsibilities Data quality is an on-going process, not a one-time project Include initiatives to grow analytics knowledge & skills across the organization Remove cultural barriers to success Don t forget to engage the infrastructure team early 23
Rome wasn t built in a day Achieving the vision will take time The Roadmap will be a valuable tool for communicating with stakeholders The implementation roadmap should be built to enable demonstrating tangible progress while you build toward the vision You should have deliverables every 3-5 months in order to demonstrate progress and maintain buy-in Start with simpler projects Integrate initial, high-value data sources Gain lessons learned Lay the foundation for tackling more complex areas 24
Recommended Practices For information delivery projects Define a reasonable scope including 1-3 data sources and a core set of dashboards/reports to build Address data quality and standards for the selected subject area Initial data cleanup as part of the project Identify a data steward(s) for on-going cleanup & support Identify project dependencies Some projects cannot be completed until foundational issues are addressed Information delivery projects are unlikely to be successful until operational processes are defined and implemented Data quality, standards, and operational processes are critical to user adoption and satisfaction 25
Example Project Prioritization Criteria Integrates a high demand data source(s) High demand = supports lots of use cases or used by many groups Ease of integrating data source Technical complexity Data quality Existing standards User readiness and buy-in Benefit/Impact Business/clinical impact of supported use cases Progress toward achieving analytics goals/objectives 26
Example Roadmap One Size Doesn t Fit All Year 1 Year 2 Year 3 Governance (Continuing) Leadership Prioritize Projects Roadmap approval Establish Accountability Data Quality (Continuing) Communication Plan (Revise for new BI Tool and ACO focus) Architecture Design Ongoing Design, Build and Deployment Early Projects Categorize and prioritize Quick wins Long Term Projects (Focused on ACO Continuum of Care) 2
Example Roadmap One Size Doesn t Fit All Governance Operating Plan Define roles, procedures, policies Implement operating plan Ongoing evaluation & optimization Communication Training & Education Data Architecture Arch planning & design Select/purchase gap tools Analytics project-based implementation Ongoing support & tuning Data Quality & Source Data Projects Provider Master MPI Pt Registration (address, PCP/referring provider) Cost Accounting Clinical documentation initiatives Analytics Projects Project 1 Scope, Design, Build, Rollout Project 2 Scope, Design, Build, Rollout Subsequent projects Y1 Y2 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4
Where to start for your strategy & roadmap Organize for success Who owns the Strategic planning process? Who is the executive sponsor? What key stakeholders need to be involved? Educate for success What analytics is and why it is important to YOUR organization Analytics is an on-going initiative, not a short-term project Analytics best practices start setting expectations 29
Questions & Discussion Paula Edwards, PhD pedwards@himformatics.com