Hospitals and other healthcare providers are scrambling to close the data governance gap.
With the rush to performance-based medical care, the importance of healthcare analytics that can help providers improve patient outcomes and substantiate that they are meeting their contractual performance goals has soared. But providers are lagging in their efforts to improve data quality and consistency on which these analytics initiatives depend.
"Information governance [IG] is not new to healthcare," says Linda Kloss, president of consulting firm Kloss Strategic Advisors and former CEO of the American Health Information Management Association (AHIMA). "But while there's good work to build on, current information governance practices tend to be fragmented, designed for a paper-based world and inadequate for a digital health system."
They also tend to be less selective than they need to be. "A robust information governance program leads to higher-quality data, which in turn yields higher value," says Thomas Janssen, manager of business intelligence and enterprise data warehousing at Memorial Health System, based in Springfield, Ill. "IS departments often think that by moving all of the data they can into the warehouse they will have everything they need, when in fact the opposite is true," he asserts. "Information governance without a data onboarding strategy is very problematic and dilutes the overall quality of data in the warehouse."
Information governance are rules, policies and procedures, and controls, to manage information in an organization. As healthcare organizations become more data-intensive, rules governing such information becomes more important to ensure data is of the highest quality and carefully maintained throughout its lifecycle.
There's no question that data analytics and corresponding data governance initiatives have become more important to the healthcare industry. "When you transition to performance-based medicine, part of your income is now dependent on how you can prove you're delivering quality and outcomes," says Jon Zimmerman, GE Healthcare's vice president and general manager of clinical business solutions. "That data isn't just apparent on a claim. You need to generate the kind of information necessary to help you meet your revenue targets."
Zimmerman argues that older sources of data, such as claims, paper medical records and scanned documents, are generally well managed, but that performance-based contracts require data and analytics of a new type. Current governance models "are going to have to pay more attention not just to the creation of the analytics, but to how information usage, transmission and tracking gets governed in this new world."
This viewpoint is becoming prevalent. In June 2014, AHIMA and Cohasset Associates, a consulting firm specializing in records management and information governance, conducted the first industry-wide survey on the status of information governance in healthcare. Of the 1,000 respondents, 65 percent acknowledged there is a need for a formal information governance process, and 43 percent indicated they have already initiated such programs.
But the findings also highlighted some key limitations. Among them:
* Formal, enterprise-wide IG programs are less prevalent and less mature than they need to be, given the importance and complexity of the information environment.
* Most organizations have not yet established a comprehensive IG strategy.
* Basic IG functions such as record retention and destruction are insufficient.
* Only 10 percent of respondents have established measures and metrics to guide IG and enterprise information management.
Based on the AHIMA survey results, providers like Memorial Health are well ahead of the game. The information governance process at the 671-bed, 6,700-employee health network has received strong support from the provider's senior management, Janssen says. "They have recognized over the years that a strong, robust information governance process is one key to our organization's success," he says.
Members of Memorial's C-suite take responsibility for the data for the functions they lead. The chief human resources officer, for instance, is responsible for all staff and salary data, while the chief financial officer oversees the health system's financial data. Memorial's CIO has responsibility for data storage and acts as the health system's chief data steward. A senior-leader governance committee meets monthly to consider project requests. The requests are vetted by two governance subcommittees, clinical and non-clinical, which review the requests and make recommendations based on resource availability, project impact and competing priorities.
As Memorial's healthcare model shifts to place greater emphasis on value over volume, so does its emphasis on data quality over quantity. To ensure that its data is accurate, clean and usable, Janssen says, the governance framework includes four core components: It begins with information security management and takes into account a variety of data types. Management of metadata-essentially, data about other data-is the next layer and sits on top of the information security foundation. The third core component is information architecture, which consists of information quality management, reference and master data management, data warehousing and business intelligence, as well as both structured and unstructured data management.
As Memorial moves from descriptive analytics, which uses data to report on what has already happened, to predictive and ultimately prescriptive analytics that can forecast outcomes, Janssen says its data governance framework provides better access to more meaningful data that can be used to lower healthcare costs and improve outcomes. Using a structured framework reduces waste and increases productivity, improving cost management. It also increases collaboration and leads to better decision-making. Similarly, he notes, the framework is conducive to better compliance with hospital and governmental regulations, which reduces risk.
"Core measures of the comprehensive information governance framework will lead to higher-quality data.," Janssen says. "Higher-quality data can lead to a safer environment with better patient outcomes, and better outcomes can reduce re-admittance rates yielding higher profitability."
Truman Medical Centers in Kansas City, Mo., also relies on a formal, well-defined data governance framework to support its analytics initiatives. About four years ago, as the two-hospital, 600-bed provider deployed its electronic medical record system and began digitizing more of its data, it quickly encountered difficulties.
One of TMC's biggest issues was the time it took to extract meaningful data out of its systems, according to Seth Katz, assistant administrator of information management and program execution. The process was erratic and overly long, and employees were unclear about who should be approached for what request. Says Katz, "There was no oversight, no support and no ownership."
This led to widespread recognition that data governance and the data request process needed to be formalized. So in 2012, Truman established a data quality standards committee and tasked it with reviewing every data request made by the hospitals' management or staff. The committee is charged with ensuring that the request is necessary and meets the organization's strategic goals. If the request is approved, the committee is also responsible for determining the best source for the data, that it's extracted in a timely fashion and whether the same data has been pulled recently for a similar report and can be reused.
The committee comprises a diverse group that includes the healthcare analysts responsible for the different databases at Truman. Some work for Truman's IT organization, while some are members of other departments such as quality resources. Also on the committee is a representative of the organization's project management office and a health information management professional who chairs the group.
The committee's work has led to a dramatic reduction in time to produce a report, and the consistency and reliability of the data has greatly improved. Throughout the organization, Katz says, "People are getting the data they need much faster, and they know it's data they can trust."
The data quality standards committee initiative was part of a broad IG effort to set policies and procedures for a variety of information management issues, such as data retention and privacy, and to address questions such as who owns data, how best to secure it and who grants access to it.
To get the undertaking off the ground, Katz says, Truman worked with IBM, following Big Blue's data governance blueprint for the first 12 to 18 months of the program. Gradually, the IBM framework was tailored by Truman's data managers to meet the health system's specific needs. The provider's current framework is a hybrid of the IBM model and Truman-devised best practices.
The UPMC model
To make the most of a $100 million investment in big data and healthcare analytics, in 2012 the University of Pittsburgh Medical Center launched a far-reaching data governance program that also initially relied on the IBM blueprint. UPMC's data governance team gradually developed its own customized plan to address the sprawling health system's unique challenges.
A key objective has been to make data governance integral to the culture of the more than 60,000 employees at UPMC, which incorporates more than 20 academic, community and specialty hospitals.
"As part of our program, we want to increase the analytic ability of people throughout the organization," says Terri Mikol, UPMC's director of data governance. "We believe that some kind of data work should become a core competency for everyone who works in healthcare."
Mikol says UPMC defines the scope of its governance program with five questions:
* Where can I find the information I need? This zeros in on the fundamental issue of safeguarding the provider's data.
* Where does the information come from? Data lineage is increasingly important because the data that feeds UPMC's analytics engine may be drawn from as many as 20 different sources.
* Is the data any good? Consistently measuring the quality of the data and address any shortcomings openly and transparently is essential.
* What does the data mean? There's a need for metadata that describes the different types of data and their appropriate usage. Says Mikol: "If we're going to succeed at elevating everyone's analytic capabilities, we need self-service access to data, and we can't provide that without extensive metadata that tells people what it is and how to use it."
* What am I permitted to do with this data? This addresses the need for clear guidelines on how any particular type of data can be used.
A key element of the program is data stewardship. No single group can master all the vast quantities of data UPMC generates, so the medical center has adopted a "divide and conquer" approach, Mikol says. Each data category is assigned a unique team of stewards, who are held accountable for that data domain across the entire organization.
Although their frameworks and organizational structures vary, the driving force behind the IG initiatives at Memorial, Truman and UPMC is similar: to better align their data with their strategic goals by managing it as one of their organization's most valuable assets.
Central to this is ensuring the quality and integrity of that data. As consultant Kloss puts it: "If the providers' information isn't accurate, the whole of our healthcare ecosystem suffers."
Register or login for access to this item and much more
All Health Data Management content is archived after seven days.
Community members receive:
- All recent and archived articles
- Conference offers and updates
- A full menu of enewsletter options
- Web seminars, white papers, ebooks
Already have an account? Log In
Don't have an account? Register for Free Unlimited Access