Creating actionable intelligence

Hospitals and health systems have created a wealth of data by adopting electronic health record systems that can integrate information from a variety of sources. For many provider organizations, however, the ability to capitalize on the value of this information has lagged behind the ability to assemble and accumulate it. Hence, they have an opportunity to apply the lessons of other information-enabled sectors that have already learned how to extract real value from their data stores.

As demonstrated in a recent American Hospital Association study (, providers spend tens of billions of dollars a year to meet federal compliance requirements without consistent, clear evidence of corresponding benefit. The resulting focus on compliance with regulatory mandates has influenced how health care organizations in general view their data assets.

Thankfully, sector leaders in some provider organizations have learned how to mine “the gold in them thar hills” and are establishing a new approach to reaping value from the data they now possess. They are providing proof that data and information technologies can drive organizational improvements, support enhanced decision-making, and enhance patient experience and clinical outcomes. Their progress has been coupled with new approaches to data collection, analysis and organization.

Trustee talking points

  • Hospitals and health systems have access to reams of data, often at a high cost.
  • Finding value in the data can require performance improvement in health information technology operations.
  • Leaders can devise strategies and procedures to acquire this "actionable intelligence."
  • While they don't have to be HIT experts, trustees need to ensure that data systems are creating value for their organization.

As part of this shift in strategic thinking, many trustees have a unique opportunity to refocus their organization's information resources on performance improvement by creating "actionable intelligence." This article identifies strategic perspectives and policies that leaders can adopt in the process of accelerating patient-focused improvements in efficiency and effectiveness — delivering services of expected quality as inexpensively as possible.

Trustees do not need to be experts in health data and information systems to guide this transformation. They can begin by asking their facility managers a basic question: How is our organization using data in real time to deliver actionable intelligence that optimizes the productivity of resources and reallocates them as necessary to meet strategic goals? Because technologies and strategies evolve constantly, leaders should keep asking the question to focus health IT managers on improving daily operations — something very different from compliance with government regulations.

Curing dysphoria

HIT professionals know there’s a problem. Many call it information dysphoria (a profound state of feeling unhappy, uneasy or dissatisfied). The condition encompasses several daunting challenges they face on a daily basis:

  • Collecting relevant data from disparate sources and information systems.
  • Integrating data from different systems into a “single source of truth.”
  • Organizing data into accurate and meaningful information.
  • Making appropriate information available to decision-makers on a timely basis.
  • Providing decision-makers with knowledge and tools to use data intelligently.

Trustees need to be aware that HIT “has issues.” More money will not make a health system’s information resources more productive if it does not target the underlying causes of dysphoria. Investments should be directed specifically toward providing information that improves productivity, such as systems that organize information from several sources to facilitate caregivers’ decision-making and free them from data-gathering tasks that can be automated. (Data also must be meaningful and accurate, but that’s a topic for another article. Good decisions simply cannot be made with bad data.)

HIT investments that eliminate information dysphoria create more time for caregivers to interact with patients — not computers — consistent with findings of the AHA’s recent study and many other reports questioning the benefit of HITECH. According to AHA President and CEO Rick Pollack, “There is a growing frustration for those on the front lines providing care in a system that often forces them to spend more time pushing paper than treating patients. Too often, these regulatory requirements seem detached from good and efficient patient care.” 

To improve care, HIT systems must provide the right information to program directors and unit managers who make the operational decisions (e.g., staffing, space allocation, scheduling or purchasing) that give clinicians more time for patient interaction.

Serving its purpose

The wealth of information held in a typical hospital’s HIT systems is in a collection of discrete data silos, each with a primary purpose that has value on its own. Shortcomings in integration across the silos, however, stymie providers’ ability to obtain full value from their data. The use of comprehensive EHR systems has helped, but EHRs only include a portion of the data created and maintained by the organization. The new focus on viewing data as a strategic asset for the entire organization — not merely a necessary cost of doing business for its individual departments — has enabled leading providers to answer the vexing questions about how to optimize quality (including consumer experience), outcomes (including population health) and efficiency (including fixed reimbursement).

Trustees are responsible for strategy, so it’s their role to make sure managers’ attention is focused on deriving more value from HIT by addressing the inefficiencies and missed opportunities for patient care arising from widely distributed, uncoordinated data. For example, managers commonly assume that the organization has a single location for all patient data because it has an EHR. This is incorrect; an average hospital has more than 400 separate systems interfaced with the EHR. Many of these interfaced systems only share a subset of their information with the EHR, even though the data would be useful in other applications.

The gap between clinical care and operations (e.g., EHR and personnel management systems) can be even larger. Data warehouses are often created to store everything in one place, but a data dump does not yield useful information on its own. Information tends to remain locked in incompatible structures that act as a barrier to coordinating data from different sources (e.g., admissions information, surgical schedules, patient acuity levels, staff capabilities or relative costs of staffing alternatives).

The complex nature of data warehousing frequently requires an HIT organization to hire consultants to deal with backlogs, which are often managed on a first-in/first-out basis rather than prioritized according to strategic value. Trustees don’t need to be knowledgeable about day-to-day management of data warehouses or other technical issues. But they should ascertain that appropriate governance processes are in place to prioritize projects (including data integration), enhance decision-making and attain strategic objectives. Without sound governance, the HIT organization might allocate resources based upon perceived urgency, the title of the person requesting assistance or the loudest voice — which could result in doing well a project that is not worth doing.

Trustees’ role

To ensure that data systems justify their high costs by creating strategic value, trustees and CEOs must articulate a clear expectation of performance improvement across their organizations. (Pervasive and accountable performance improvement is almost certainly a precondition for the essential regulatory relief being sought by the AHA and other sector leaders, but like the quality of data, it is a topic for another article.) Getting a better return on HITECH-driven investments in data systems is the next step toward efficiency and effectiveness in health care, which means using information resources to create actionable intelligence — consistent, current, correct and complete information that leads to better use of resources.

This new strategic goal may be easy to state, but it’s hard to achieve because health reform for the past decade focused on how to collect and store data rather than how to create valuable information as efficiently as possible. Trustees are, therefore, essential to the process of creating the expectation that HIT will provide actionable intelligence.

Trustees should not seek to manage the transformation, even if they are IT experts, but they must take steps to make it happen within their organization to support enhanced outcomes and operational efficiency. Like health care in general, performance improvement through actionable intelligence is local. As the ultimate voice for their local communities, trustees have an obligation to expect that their organization’s HIT produces actionable intelligence.

Gerard M. Nussbaum, J.D. (, is a strategic health care adviser and principal with Zarach Associates in Chicago. Jeffrey C. Bauer, Ph.D. (, is an independent health futurist and medical economist based in Madison, Wis.

Health care information systems

The complexity of demands on today’s health information technology is illustrated by this partial list of databases and information systems needed to operate a health care delivery system:

Clinical data

  • Electronic health records.
  • Ancillary department information (e.g., laboratory, radiology).
  • Chronic disease management.
  • Disease registries.
  • Patient portals.
  • Population health management systems (e.g., demographics).
  • Research (e.g., clinical trials, pilots, investigational studies).
  • Health insurance interfaces.

Business data

  • Revenue cycle (e.g., billing, collections, receivables).
  • Claims denial and audit results.
  • General ledger.
  • Supply chain (e.g., purchasing, inventory).
  • Human resources.

External sources

  • Payer systems (e.g., Medicare/Medicaid, private).
  • Government records (e.g., licensure, certifications).
  • Sector benchmarks (e.g., National Committee for Quality Assurance, Leapfrog, American Hospital Association).
  • Health information exchanges.
  • Sector reports (e.g., Healthcare Effectiveness Data and Information Set, Press-Ganey).
  • Social media (e.g., report cards, consumer ratings).

This list demonstrates that the challenge is not a shortage of data; rather, it is organizing an abundance of data to identify needs and deploy resources for operational improvement. One of the most important steps in this process is relating all these data sources to one another with a single patient identifier or algorithms that integrate information for single individuals who are registered under multiple names.

Another step toward data efficiency is eliminating identical data collected in more than one system (a real annoyance to patients who are asked the same questions over and over again). Regulations addressing interoperability have not necessarily solved the problem. However, even organizations that have achieved higher levels of interoperability still face barriers to using data for decision-making due to challenges like semantic conflicts and inconsistent data structures among connected systems. Trustees must help managers to keep the focus on integrating data to support decision-making; they cannot assume that value is created just because information systems are HITECH-compliant or interoperable. — Gerard M. Nussbaum and Jeffrey C. Bauer

Creating value from existing data with 'actionable intelligence'

Vital signs are among the most important data in patient care. Unfortunately, the traditional manual process of collecting and reporting these data can lead to measurement errors and delays in making data available for decision-making. Automation at the interface between measurement devices and the electronic health record can solve these problems, but a more robust solution is needed to create strategic value.

A system focused on "actionable intelligence" will leverage collected information to support analysis of the workflow — such as the caregiver’s time on task and compliance with protocols for each vital sign-measurement event. Nursing supervisors can use this information in real time to ensure that appropriate care is being delivered, to identify caregivers whose performance needs to be improved, to compare performance across nursing units for identifying differences that suggest changes in protocols and workflows, and to assess the costs and value of different equipment. — Gerard M. Nussbaum and Jeffrey C. Bauer

Trustee takeaways: Steps toward 'actionable intelligence'

Trustees can help their organization to create value through "actionable intelligence" by asking that evidence of success be included in project status reports. Here are examples of outcomes that demonstrate real progress:

  • Providing information in real time: To ensure employees’ successful compliance with performance improvement programs, line managers need data to identify performance deviations and to take corrective action on a continual basis, using current data.
  • Eliminating unproductive alerts: Due to Centers for Medicare & Medicaid Services requirements, caregivers are inundated with warnings about data values falling outside normal ranges. Alert fatigue creates stress, reduces productivity and causes job dissatisfaction. A system focused on actionable intelligence, such as automated vital signs management, uses caregiver feedback to customize alerts to each patient, not to one-size-fits-all standards.
  • Enabling scenario analysis of alternatives: Health care is changing at an accelerating rate and heading in new directions. Scenario analysis considers a variety of possible outcomes and related assumptions, including some that may differ greatly from historical experience. It allows decision-makers to anticipate unprecedented forces, prepare appropriate approaches and monitor evolving outcomes in consideration of organizational strategies. Actionable intelligence is needed to answer the unprecedented what-if questions.