Health care providers are under growing pressure to better manage the quality and efficiency of care, address an often bewildering array of new payment approaches, establish a continuum of care and ensure the health of a population.

They must be able an answer questions such as:

  • Is there too much variation in our care practices?
  • What is the return on investment of a capital purchase?
  • Can we afford to sign this managed care contract?
  • Where should we focus our cost reduction initiatives?
  • Are we enhancing our ability to serve market needs?
  • Do our care processes follow established protocols?

Answering these questions requires information technology: electronic data warehouses, electronic health records, report generators and health information exchanges. But more critical than the technology are the data.

The importance of data has been noted in a wide range of industries over many decades. Airlines use data on travelers’ journeys to establish frequent flier programs and yield-management systems. Internet-based retailers use data on prior purchases and customer demographics to place advertisements and suggest purchases. The sky-high valuation of Uber is often attributed to its treasure trove of data on customer trip patterns. Equipment manufacturers use sensor data to detect equipment performance problems and resolve them before the equipment fails.

These advantages have not been lost on health care providers. Thriving under Medicare Advantage requires an accurate and comprehensive recording of hierarchical condition categories. Improving the quality of care requires coded problems, medications, laboratory results and patient demographics.

But obtaining and maintaining high-quality data is difficult. Clinician documentation is often incomplete. Claims files arrive with often startling data quality problems. Data that are received from multiple electronic health records in a health system can be conflicting and redundant.

There are no management steps that will completely improve data quality, timeliness and accuracy; data will never be perfect. However, the imperfections should not sway organization leaders to neglect their data. They need to manage it just as they manage other important assets — people, equipment, physical plants and balance sheets.

Managing the data asset

How should organizations approach managing this critical asset? Providers can, and should, take several data management steps:

Establish data management as a business imperative

Organizational managers must discuss why they should invest in an effective approach to managing data. The rationale must be more compelling than “we need better data.” It can be understanding cost structure, care quality or revenue cycle performance. Regardless of the rationale, the case should be clear to managers.

The rationale can be challenging to define, as the value of better data can be difficult to measure. It may not be obvious whether an investment of several hundred thousand dollars will improve care quality. Data investments very often rest on the strength of managers’ judgment that an investment is necessary for an organization to achieve its goals and objectives. Yet, managers need to measure improvements in the quality of data — e.g., assessment of data error rates, determination of the level of inconsistency between different types of data and data timeliness. While such improvements are measurable, leaders should understand that developing an ROI will be challenging.

Ongoing efforts will be needed to define data standards, improve data quality through changing processes and eliminate many current reporting approaches. These efforts are real work and often unglamorous and politically challenging. Managers must commit to the broad set of work that data management will require.

Create the data management vision

The organization will need to define the scope of the data management effort and the plan to achieve it. This vision often starts with improving data collection practices at the front end — the processes used to identify referring physicians in the specialty clinics — and moves all the way through definitions of the composition and organizational location of data analysis staff. The vision will describe the range of critical data, the types of analysis desired, steps to be taken to improve data quality and the class of users who will authorize analyses and receive the results.

This vision might focus on the management of data needed to perform the following analyses:

  • Population management analytics: Produce a variety of clinical indicator and quality measure dashboards along with reports to improve the health of a whole community, as well as identify and manage at-risk populations.
  • Provider profiling and physician performance analytics: Normalize (severity- and case-mix–adjusted profiling), evaluate and report the performance of individual providers (primary care providers and specialists) against established measures and goals.
  • Point-of-care health gap analytics: Identify patient-specific health care gaps and issue a specific set of recommendations and notifications to physicians at the point of care and/or to patients via a patient portal or personal health record.
  • Disease management: Define best-practice care protocols for multiple care settings, enhance the coordination of care and improve adherence to best-practice care protocols.
  • Cost modeling, performance risk management and comparative effectiveness: Manage aggregated costs and performance risk, integrating clinical information and clinical quality measures.

Establish data governance

Data governance is the most critical part of day-to-day collection and use of data. Data governance involves managing the organization structures, policies and processes needed to define, control and ensure the quality of data. Moreover, the governance must determine the focus of data management efforts such as those outlined above.

Data management mechanisms must cover the following:

  • Identify the sources and definition of important data and the policies, procedures, technologies and tools needed to manage this data.
  • Outline the steps, communication and tools necessary for making changes to the list of strategic data elements and their definitions.
  • Define the roles and responsibilities for managing the organization’s data asset.
  • Develop means to measure the quality of data and determine if data quality improvement efforts are working.
  • Prioritize the demand for data analysis and ensure that unmet demand does not lead to the proliferation of shadow systems.

With a data management focus and vision, and using data management mechanisms, the organization takes on the work of tackling analytics. As an organization’s leaders peel apart different facets of analysis, they will formulate questions, such as these related to bundled payments:

  • How do we define the costs of a procedure so that we can analyze our performance under bundled payment?
  • What cost elements do we need?
  • Where do these data come from?
  • What do we know about the quality of the data? If it needs to be improved, what steps do we need to take?
  • Are the data consistently defined for various cost-generating sites?
  • Will there be issues linking patients and providers for these sites?
  • After we assess the above for one procedure, which procedures should be next?

The form and organizational location of the governance function will vary, largely driven by the scope of the data management effort. If the organization is focused on improving the revenue cycle, for example, then the governance structure may report through finance, and the governance steering committee will include stakeholders who manage key components of the revenue cycle. The governance body may commission changes to improve the accuracy of insurance-information data capture at time of registration and develop a single organizational definition of the data element “visit.”

If, on the other hand, organizational leaders are using data to measure care quality, the governance structure might report through the chief medical officer. In this case, the steering committee would include a large number of clinicians and managers of clinical departments. The governance body may commission changes to improve the capture of medical error events and define clinical best practices for treating patients with a chronic disease.

Regardless of organizational governance, leaders should strive to designate a “single voice” to clinicians and others about data quality needs and priorities. Clinicians, for example, often feel overwhelmed by the myriad data demands placed on them.

Those who participate in data governance should recognize that management approaches might be different from those used for large information technology implementation initiatives. Data governance will often involve the orchestration of many, at times loosely connected, projects that improve the data asset and strengthen the organization’s analysis capabilities. In contrast, a large implementation initiative will usually have a set of clearly interconnected tasks that advance a very specific goal, such as “reduce medication errors.”

Moreover, implementations end. Data management never ends.

Implement data management roles

There are four major roles:

  1. Data stewards are responsible for developing the policies and procedures that govern data. Stewards have a very challenging role. They must understand operational and business requirements and balance them with the inevitable messiness of organizational life. For example, physicians often fail to use coded problem lists because they take too much time, but coded problem lists are essential for high-caliber analyses of care quality.
  2. Data owners are the individuals and functions in the organization that generate the data. Typical data owners are those who do health information management of procedure code assignment and outpatient registration of patient demographics. Data owners are responsible for implementing the policies, procedures, training and systems — defined by data stewards — that are needed to enforce organizational definitions of data and standards for data quality. At times, the data needed are generated outside the organization — for example, from payer claims. In these cases, the data owner is someone who can work with the payers to improve the quality of the claims data.
  3. Business users employ the results of data analysis and often perform the analyses themselves. The business users are responsible for learning how to use analytics tools and understanding the strengths and limitations of the data. They should also ensure that the data to be used will address the questions the analyses are intended to address.
  4. Data managers manage the information systems that capture, edit and transform data. They need to ensure that the tools are of high quality, the ongoing operations of the IT foundation (such as backup) are performed, training is provided to users, and counsel is given to business users who have questions about the data or an analysis strategy. The data managers will develop data schemata, identify preferred source transaction systems for data and understand the limitations of the organization’s data.

Knowing your customers

For most providers, data has been secondary, less worthy of investment than equipment, buildings, and IT support of clinical, operations and finance processes.

While process improvements and investments in buildings and equipment will always be important, managing data deserves the same recognition. In some cases, the quality and timeliness of data may be the most significant source of an organization’s viability in a rapidly changing and demanding industry. For a potato chip manufacturer, knowing the demographics of the customers who consume your snack food, what else they buy when they buy your product, and where and when they buy it may be far more important than inventory management.

In an era of growing pressure on provider performance, data become critical. If you don’t know your costs and your care quality, thriving under value-based reimbursement will be challenging.

John Glaser, Ph.D., is the senior vice president of population health with Cerner in Kansas City, Mo. He is also a regular contributor to H&HN Daily.

The opinions expressed by the author do not necessarily reflect the policy of the American Hospital Association.