Health Care IT Advisor Case Study Data Governance Is Key to UPMC s $100 Million Investment 2445 M Street NW Washington DC 20037 P 202.266.5600 F 202.266.5700 advisory.com June 17, 2014 14-026 2014 The Advisory Board Company 1 advisory.com
Health Care IT Advisor Meg Aranow Senior Research Director AranowM@advisory.com LEGAL CAVEAT The Advisory Board Company has made efforts to verify the accuracy of the information it provides to members. This report relies on data obtained from many sources, however, and The Advisory Board Company cannot guarantee the accuracy of the information provided or any analysis based thereon. In addition, The Advisory Board Company is not in the business of giving legal, medical, accounting, or other professional advice, and its reports should not be construed as professional advice. In particular, members should not rely on any legal commentary in this report as a basis for action, or assume that any tactics described herein would be permitted by applicable law or appropriate for a given member s situation. Members are advised to consult with appropriate professionals concerning legal, medical, tax, or accounting issues, before implementing any of these tactics. Neither The Advisory Board Company nor its officers, directors, trustees, employees and agents shall be liable for any claims, liabilities, or expenses relating to (a) any errors or omissions in this report, whether caused by The Advisory Board Company or any of its employees or agents, or sources or other third parties, (b) any recommendation or graded ranking by The Advisory Board Company, or (c) failure of member and its employees and agents to abide by the terms set forth herein. The Advisory Board is a registered trademark of The Advisory Board Company in the United States and other countries. Members are not permitted to use this trademark, or any other Advisory Board trademark, product name, service name, trade name, and logo, without the prior written consent of The Advisory Board Company. All other trademarks, product names, service names, trade names, and logos used within these pages are the property of their respective holders. Use of other company trademarks, product names, service names, trade names and logos or images of the same does not necessarily constitute (a) an endorsement by such company of The Advisory Board Company and its products and services, or (b) an endorsement of the company or its products or services by The Advisory Board Company. The Advisory Board Company is not affiliated with any such company. IMPORTANT: Please read the following. The Advisory Board Company has prepared this report for the exclusive use of its members. Each member acknowledges and agrees that this report and the information contained herein (collectively, the Report ) are confidential and proprietary to The Advisory Board Company. By accepting delivery of this Report, each member agrees to abide by the terms as stated herein, including the following: 1. The Advisory Board Company owns all right, title and interest in and to this Report. Except as stated herein, no right, license, permission or interest of any kind in this Report is intended to be given, transferred to or acquired by a member. Each member is authorized to use this Report only to the extent expressly authorized herein. 2. Each member shall not sell, license, or republish this Report. Each member shall not disseminate or permit the use of, and shall take reasonable precautions to prevent such dissemination or use of, this Report by (a) any of its employees and agents (except as stated below), or (b) any third party. 3. Each member may make this Report available solely to those of its employees and agents who (a) are registered for the workshop or membership program of which this Report is a part, (b) require access to this Report in order to learn from the information described herein, and (c) agree not to disclose this Report to other employees or agents or any third party. Each member shall use, and shall ensure that its employees and agents use, this Report for its internal use only. Each member may make a limited number of copies, solely as adequate for use by its employees and agents in accordance with the terms herein. 4. Each member shall not remove from this Report any confidential markings, copyright notices, and other similar indicia herein. 5. Each member is responsible for any breach of its obligations as stated herein by any of its employees or agents. 6. If a member is unwilling to abide by any of the foregoing obligations, then such member shall promptly return this Report and all copies thereof to The Advisory Board Company. 2014 The Advisory Board Company 2 advisory.com
Table of Contents Abstract... 4 Background... 5 The Plan... 5 The Progress... 6 Year 1 Objectives: Program Structure and Roles... 6 Year 2 Objectives: Implementation and Awareness... 7 Year 3: The Early Project... 8 Creating a Provider Master Record... 8 Managing Master Data across the Source Systems... 8 Increasing Data Transparency... 8 What Does the Future Hold?... 8 What You Can Learn From UPMC... 9 2014 The Advisory Board Company 3 advisory.com
Abstract UPMC s $100 million spend on big data and analytics made the headlines in 2012. In the background was the development of a substantial data governance initiative formed to protect that investment by ensuring accurate and actionable source data for the analytic engine. UPMC s methodical approach underscores and exemplifies the concept that data governance is a business process rather than an IT project. As UPMC entered their third year of the initial five-year data governance plan, the Health Care IT Advisor (HCITA) team checked in with Terri Mikol, Director of the Data Governance Office, to see how they were doing and what they plan to do next. 2014 The Advisory Board Company 4 advisory.com
Background UPMC sees the future of care in personalized medicine. (The National Institute of Medicine defines personalized medicine as...an emerging practice of medicine that uses an individual's genetic profile to guide decisions made in regard to the prevention, diagnosis, and treatment of disease ). With the help of big-name technology partners Oracle, Informatica, and IBM, UPMC is creating a data warehouse intended to power the medical effort, bringing together vast amounts of data for analysis. At the inception, the project was envisioned to target source data totaling 3.2 petabytes in over 200 source systems. Primary sources of data include the multiple registration systems and EMRs, including Cerner, Epic, and Varian. Additional data is sourced from ancillary systems, genetic research databases, and claims from a UPMC-owned health plan. Early on in the project, leadership realized the need to more effectively protect and manage the data that would provide the basis for this huge endeavor. During the startup phase of the project, the vendor partners encouraged UPMC leaders to see the full value of their data asset and made a compelling case that an investment in analytics required formalized and coordinated data governance to protect the asset. While data governance concepts were not new to UPMC, the effort to create a centralized approach was a departure from what had existed previously. Thus, the Data Governance Office (DGO) was created as an enterprise-wide steward of good data management practices. DGO (a director and 8 full-time equivalents [FTEs]) is situated within the Enterprise Systems and Data Management (ESDM) division of IT, reporting to the VP of Enterprise Analytics. DGO describes its work as answering five basic questions to ensure optimum value from the UPMC data: Where can I find the information I need? Where did this information come from? Is this information any good? What does this information mean? What am I permitted to do with this information? The Plan An initial three-year plan was developed in 2011 to structure and guide the DGO efforts and achieve the long-range plan of sustainable processes to produce accurate and complete source data. After studying the IBM model of data governance, 1 the UPMC team developed their own customized plan to address the challenges and culture unique to their organization. The plan was initially vetted and approved by the Data Governance Council (see Figure 1). The council reviews the plan annually. The goal of the plan was to shift ownership of the data process from ESDM to the council by the end of a threeyear period. More specifically, the first year would be spent defining the program structure and policies, defining the roles, and acquiring the technology. The second year would focus on tool implementation and creating institutional awareness of the efforts. The third year was targeted to initiate significant projects, placing responsibility and authority with the relevant operations teams. The years that follow would be for mature adoption, including measurements of compliance and success, specifically in the areas of data quality and metadata utilization and effectiveness. The UPMC approach is markedly different than others we have seen in that it allocates three full years to firmly establish the new data governance structures and processes and allocates considerable 1) 1 IBM Corporation, The IBM Data Governance Blueprint: Leveraging best practices and proven technologies, May 2007. 2014 The Advisory Board Company 5 advisory.com
effort for corporate communication intended to transform the corporate culture by changing the attitude toward data ownership and stewardship. The centralized function is being developed on a track that is parallel to the existing grassroots efforts that had already existed in various places throughout the organization, with a goal to eventually supplant the latter. The Progress As of April 2014, UPMC was on schedule, having achieved the goals targeted for the first three years. Year 1 Objectives: Program Structure and Roles The program structure and roles are defined in Figure 1. Figure 1: UPMC Data Governance Program 2 The process for identifying and naming data stewards, identified as a first-year goal, closely follows the targeting of data for the enterprise analytics platform: as new data is identified for inclusion in the warehouse, the process for identifying the most appropriate data steward begins. Often there is not an obvious, single choice to serve in the role; 2) Reprinted from Mikol, T., UPMC Enterprise Analytics and Data Governance, February, 2014. Used by permission of UPMC. 2014 The Advisory Board Company 6 advisory.com
rather, it takes study and, often, negotiation. To date, more than 200 data stewards have been named and granted decision rights. The stewards role has had an immediate impact in driving decisions to manage and select from duplicate data sources, such as selecting the best system and elements to a track patient s location and identify care teams within the Emergency Department. Although today data governance is narrowly focused on support of the enterprise analytics priorities, it may be extended to the transaction systems more generally in the future. While UPMC agrees that best practice is to change or correct the data at its source, they acknowledge that resource and time constraints have led them to instead incorporate some changes in the downstream processes. Figure 2: UPMC Role Definition 3 Data Steward Create and maintain data/business definitions Assist with defining data access rules, data mappings, and data aggregation methods Assist with Master Data management trust rule definitions Maintain reference tables for codes, descriptions, mappings, and groupings Participate in data integrity improvement initiatives Year 2 Objectives: Implementation and Awareness Informal roundtable discussions are organized by the DGO and attended by data consumers and members of the data governance community. These roundtables are part of the Year 2 awareness campaign, which also included the creation of a LinkedIn group and Twitter feed. Roundtables provide a forum for business owners to share information about their content domains. Recent roundtable topics included the data processes for collecting and maintaining genetic data and another that focused on finance data. DGO has, as planned, begun to turn over the agenda-setting reins to the council. Data quality issues are viewed as production support issues and as such they are reviewed and prioritized by the Information Owners in the business and clinical units. For example, a recent data quality initiative surfaced a need to standardize lab codes for specimen subtypes. The issue was brought to the attention of the Information Owner and changes to the source systems were prioritized. The technology has been acquired from and implemented with the original vendor partners. The tools to support the work include Informatica for managing master, reference, and metadata and for basic data quality checks. Microsoft SharePoint stores educational material such as training documents for employees newly appointed to the data governance team. Finalized policies are stored along with all enterprise policies on UPMC s intranet. IBM s portal will be deployed as the formal enterprise analytics portal and the data governance results such as metadata statistics, data quality scores, and glossaries will be made available on this site. Overseeing all of this, a 29-member council has been established to drive the work forward. The council is intentionally reticent to approve policies, not wanting to overburden the processes that are a part of the organization learning a new way of conducting its operations. They have approved only three policies: a policy for 3) Ibid. 2014 The Advisory Board Company 7 advisory.com
information ownership and decision rights, a policy for glossary 4 change management, and a policy structuring the allowable testing in production systems. Year 3: The Early Project Creating a Provider Master Record Three earlier attempts, over 15 years, had been made to develop a standard provider master identification record to be used by all systems within the enterprise. Now, within the framework of the DGO there is significant progress. The Provider Gold Record was created and verified against the many existing fragmented data sources. The new structure and processes involved assigning the role of data owners (see Information Owners and Data Stewards in Figure 1), rather than relying on volunteer stewards, and formalizing the responsibilities with decision rights and accountability. Today, the new Gold Record is used within the enterprise analytics framework. There are future plans to position it to: Underpin the accountable care and pay-for-performance reporting that is done by physician and service line Serve as the basis for the external facing web pages directory functions Managing Master Data across the Source Systems The various transaction systems within the UPMC enterprise were implemented with disparate master identifiers for employees and different code sets for patient types, discharges, insurance plans, and facilities (e.g., rooms, beds, and physicians offices). The inconsistencies necessitate that the reporting and analytics functions must cross map the codes when combining data for horizontal accounting and care continuum reporting. Although the ultimate goal is to standardize and eliminate the need for mapping, the early initiatives are focused on creating a single mapping source of truth. The first initiative is reconciling employee identification information. Increasing Data Transparency All output from the data warehouse has an accompanying metadata fact sheet that includes glossaries, data lineage, and data quality information. The fact sheet is intended to aid with the interpretation of the output in its original form and the further manipulation or analysis of the data by the end users. This work is seen as complementary to and supportive of the work of the Information Security Department. Information round tables (mentioned above) also contribute to data transparency by sharing data collection and usage information among the relevant stakeholders. What Does the Future Hold? UPMC feels that it is on the right path, so the future plan includes maintaining and extending successful processes. Notably, while today UPMC s data governance program is tightly coupled with its enterprise analytics efforts, in the future its role and scope will expand to include the transaction systems more directly. The Provider Gold Record and other mappings and sources of truth will be made available to the transactions systems to correct data collection at its source. 4) The glossary currently includes more than 300 terms. 2014 The Advisory Board Company 8 advisory.com
What You Can Learn From UPMC Lessons learned and shared by UPMC include: Data are among a health enterprise s most valuable assets. Its value is protected through formal data management practices. New, dedicated data governance roles are necessary as a catalyst for change. All data governance roles should be explicitly defined to include authority and responsibility. Data governance is not a project; it is an ongoing business process. Master data are a good early focus. Education and awareness are key elements of a successful data governance program. 2014 The Advisory Board Company 9 advisory.com