Over the past several years, health care has seen the beginnings of a significant transformation in the nature of analytics. This transformation is being driven by the advent of a new world in which providers hold increased accountability for the efficiency, quality, and safety of the care that their organizations provide—and it is occurring across several dimensions.
Historically, analyses have been retrospective, with systems assessing the organization’s performance over the preceding months or years. Sometimes gaps of weeks, if not months, occur before the data needed for analyses become available. Retrospective analyses will remain important in the years ahead, but data analytics increasingly will be used to provide a real-time view of a healthcare organization’s financial and clinical performance.
Under payment models that reward efficiency and high-quality care, if a hospital or health system is losing money due to inadequate clinical performance, it cannot afford to wait one or more months to find out about the problem. Healthcare leaders should understand how their organizations are performing today so they can take corrective action before revenue loss becomes a hemorrhage.
Use of Analytics to Better Understand Cost
The need for greater understanding of a healthcare organization’s cost structure is one driver for the increased use of both retrospective and concurrent analyses in hospitals and health systems.
Traditionally, providers have focused on understanding costs of care in a particular setting. For example, leaders may examine costs of an inpatient admission for acute myocardial infarction or the costs of a specific surgical procedure. In the future, understanding costs around “units” of care will remain important. However, the advent of “holistic payment” requires that costs be understood across organizations and, often, over long periods of time. For example, payments for procedure bundles require that the provider understand not only the costs of delivering the procedure, but also upstream and downstream costs (e.g., providing rehabilitation and outpatient testing before the procedure).
Holistic payment necessitates understanding costs across care settings. These settings may be organizations that have affiliation arrangements with the provider, so cost data may need to be obtained from a wide range of IT platforms. Aggregating cost data has often required bringing together data from multiple systems. This challenge becomes more pronounced with the need to perform broad cost analyses across several organizations for holistic payment methods.
Processes for costing also are changing. Leaders traditionally have relied on diverse approaches to estimate and allocate costs when assessing the organization’s costs. These approaches often result in analyses that have been “good enough” for general purposes. However, with the advent of holistic payment and the increasing materiality of payment rewards and penalties, reliance on “good enough” estimations of cost becomes increasingly perilous.
When estimated costs for a bundle of care rely on the summation of a broad set of estimated costs, then the confidence interval for the total costs of care becomes wider. Accuracy of the cost estimates becomes impaired once the range for an aggregation of a larger number of estimated cost variables is too broad. For example, a hospital will find it particularly challenging to identify costs of diabetes care accurately if it is trying to base its calculation on estimates of all costs associated with care provided for a patient with poorly managed diabetes.
Moreover, in a payment environment that pays less for care when it is managed poorly, the loss of costing accuracy becomes dangerous. For example, a provider may think the organization has achieved positive margin on a procedure bundle when it actually has recorded a negative margin because the cost estimates were inaccurate—e.g., there was a failure to fully allocate costs or capture the additional costs incurred with poor performance.
Narrow Versus Broad Diversity of Performance Measures
An increased diversity of performance measures is perhaps the most signifcant effect of health care’s new era of accountability. Historically, providers have had to assess fairly narrow performance categories, such as financial performance of a service line, productivity of nursing staff, and costs to provide care under specific DRGs given
a particular payer mix. Although such analyses will remain important in the years to come, leaders now need to assess much broader aspects of performance, including the following.
Waste. In 2013, the Institute of Medicine issued the report Best Care at Lower Cost, which examined waste as a healthcare cost and quality driver. Although no one disputes that waste is a massive problem, defining waste remains fuzzy. Going forward, providers will need to be able identify and assess performance in relation to waste in its many forms to optimize value of care delivered.
Readmissions. In October 2014, penalties for high readmission rates will increase up to 3 percent. Also, conditions included in the readmission cohorts will be expanded to include hip and knee arthroplasty and chronic obstructive pulmonary disease. As such, readmissions performance will more significantly impact financial performance.
Conformance to best practices. Value-based purchasing requires organizations to measure adherence to best practices across a wide array of patient conditions. As such, healthcare leaders will need to understand clinical performance on a much broader and deeper level.
Safety. The regulatory focus on hospital-acquired conditions (HACs) has increased to new levels, leaving providers with a greater need to understand the factors that are contributing to a high or low incidence of HACs in their organizations. Healthcare organizations that are seeing a high incidence of HACs, including pressure ulcers and healthcare-associated infections, face a penalty of a 1 percent reduction on Medicare payments.
“Small Data” Versus “Big Data”
Organizations are under increased pressure to make better use of big data, or “high-volume, high-velocity, and high-variety information assets that demand cost-effective, innovative forms of information processing for enhanced insight and decision making” (“IT Glossary,”, Gartner).
IBM estimates that 90 percent of existing data is less than two years old, and the amount of data will grow by at least 40 percent over the coming years (“Harnessing the Value of Your Data on a Smaller Planet,”). Growth in quantity of information coupled with greater numbers of devices to access this information creates new challenges. Organizations will need to manage data in new ways, whether in collecting, in storing, or in using data. Providers also will need new business intelligence (BI) tools to report and mine this “big data” set.
Big data will be important—perhaps even revolutionary. However, its potential is still a matter of speculation rather than fact in health care. And in the pursuit of big data opportunities, providers should not forget the importance of small data.
Current-Generation Versus Next-Generation Technology
IT continues to experience relentless and impressive advances and innovations, and the area of analytics is no exception. New BI tools have evolved to help providers do a better job of assessing and reporting financial performance. BI tools can help track productivity across cost centers, departments, and care settings to ensure effective use of resources. In addition, more precise methods of estimating cost—including activity-based costing and exact costing—are becoming more readily available in health care.
Providers can monitor key clinical analytics to gain the insight needed to make continuous workflow improvements. Heightened levels of performance monitoring promote better recognition of processes that are working well and those in need of modification.
The next generation of BI technologies brings many advantages:
- They are more effective at importing data from diverse sources, including image and free text.
- They are becoming tightly integrated with core systems, such as electronic health records, allowing analytics to be more concurrent.
- They offer a broader array of analytics capabilities, such as predictive analytics.
- They have become more intuitive to users and support facile modeling.
The nature of analytics use in health care also is changing. Healthcare analytics traditionally has focused on reporting the status of the organization’s performance. Although this role will continue to be critical, the use of data analytics in health care increasingly has become a critical contributor to the organizational efforts to excel. Consider the following examples.
More efficient and diverse data access. Analytics can support automated abstraction of clinical information from structured and unstructured sources, such as clinical narrative text in observation notes, discharge summaries, and consultation notes. Through the use of natural language processing, the analytics can reduce reliance on manual abstraction.
Better-informed clinical decisions. Workflow logic and clinical decision support can be used to monitor and critique specific clinical decisions and clinical and operational processes. These types of analytics will become important in efforts to manage population health; the analytics can better inform decisions—for example, alerting care providers during a patient’s visit when a patient has had a previous visit to the emergency department or has missed several follow-up appointments with specialists.
Timely understanding of changes in health risk. Dynamic risk stratification algorithms can shift a patient between risk levels based on evidence of whether the patient’s condition is improving. Such understanding can help healthcare providers better assess the care interventions needed.
Preparing for the New World
Analytics and BI tools always have been important to providers. In the years ahead, their significance will become even greater, given shifts toward performance-based payment.
As a result, providers should proactively devote resources to developing a BI strategy. Their success depends on their ability to maintain financial viability while providing high-quality clinical outcomes and patient satisfaction. In this new world, organizations require the ability to analyze the myriad contributors to performance to better manage their processes and outcomes to support these goals.
John Glaser, PhD, is CEO, Siemens Healthcare Health Services, Malvern, Pa., and a member of HFMA’s Massachusetts-Rhode Island Chapter.
Publication Date: Thursday, May 01, 2014