For hospitals and other healthcare providers, data analytics just ain't what it used to be.
"Historically, we did analytics in complete silos," recounts Mark Bogen, CFO for South Nassau Communities Hospital in Oceanside, N.Y. Each department at the 455-bed facility, such as finance and quality control, would run different analyses using different data. But that began to change three years ago with the advent of health reform and the deployment of a hospital-wide electronic health record (EHR) system.
"Now," says Bogen, "the whole process is much more integrated." Instead of "dueling data," where each hospital function relied on data from different sources, "we are really sharing the data among our senior leadership and reviewing it as much as possible in a holistic way."
Likewise at South Nassau's neighboring provider, Long Island's North Shore-LIJ Health System, the way in which analytics is conducted has changed dramatically.
"For many years, we created very sophisticated analytical work," says Jeff Kraut, North Shore's senior vice president for strategy and business informatics. More recently, the 19-hospital, 6,400-bed provider's shift from a traditional fee-for-service to value-based business model "has forced us to take another look at how we're organized for analytics and what we need to do to support value-based medicine."
Previously, the hospital chain's analytics took place in what Kraut refers to as "cooperative silos." But risk-based contracting-whereby the health network gets paid for agreed-upon patient outcomes, instead of for the services rendered-has created a need to collect, collate and harness data across the North Shore system instead of department by department.
South Nassau and North Shore are representative of a transition taking place across the healthcare industry. Extracting insights from clinical and business data has become central to patient-centric healthcare delivery and value-based payment reform.
Population health and meaningful use initiatives, in particular, are dependent on the effective use of healthcare analytics to derive actionable insights from data. This, however, necessitates a less compartmentalized, more comprehensive and reliable analytics function than many healthcare providers had previously used.
"While analytics is not new to healthcare, the volume and variety of digital clinical information available today will be a critical element in improving healthcare delivery and provider performance," says Linda Kloss, a healthcare strategy consultant and former CEO of the American Health Information Management Association (AHIMA). "This is particularly true for improving outcomes within a value-based framework."
The growing importance of analytics to healthcare organizations is reflected in a survey of Healthcare Information Management Systems Society members. In the survey, conducted by Stoltenberg Consulting at a HIMSS conference earlier this year, 41 percent of the respondents designated data analytics and business intelligence as the single most important issue today in healthcare IT. But it's still new ground for many-some 51 percent of those surveyed acknowledged that they were confused about what healthcare data they should collect.
During a recent Web seminar hosted by Health Data Management, Shahid Shah, an independent technology strategy consultant for several federal agencies, tried to clarify the role and function of analytics, which he defined as "the science of examining raw data to draw conclusions.
"It is used in many industries to make better business decisions," he continued, "to increase profits, identify bad actors, improve the customer experience or satisfy open government requirements."
Shah, who helped design and deploy the American Red Cross's electronic health record solution, explained that the goal of analytics is to convert data from a number of sources into useful information that can aid in decision-making. While the process relies on math and science skills, it also requires domain or subject-matter expertise to extract meaningful insights from the data.
Breaking it down further, Shah said data analytics comprises a range of methods that include data mining, statistical analysis and predictive modeling. He emphasized that, to succeed with each of these methods, the practitioner must set clearly defined objectives and make use of the appropriate data. This, in turn, requires knowing where that data is located, how it's formatted and who has access to it-all key questions related to data management and governance.
Kloss, who also spoke during the HDM webinar, agrees with Shah that, to succeed with analytics projects, many healthcare providers need to improve their data management. Specifically, she says, they must:
* Identify, measure and correct the most critical information first.
* Prevent data inconsistencies, incompatibilities or any additional problems by changing the way data is captured, transformed and used.
* Further refine their data-gathering processes to improve data quality. These processes, she adds, can be automated and further enhanced by specialized programs and tools.
"As opposed to the traditional siloed model, where data is captured and managed department by department and function by function, new patient-centric healthcare models require an enterprise-wide strategy for managing data and ensuring quality and interoperability," the former AHIMA chief says.
From vertical to horizontal
"The role that analytics and actionable information will play in the world of value-based care continues to grow," agrees Thomas Van Gilder, an internal medicine physician and a specialist in public health and disease prevention. "Now that we're seeing payment organized around outcomes, analyzing the raw data from an EMR or a health plan and presenting it at the point of care in a way that makes sense to the practicing physician is critical, because you can't manage the health of a population longitudinally without the necessary information to support that."
Van Gilder, chief medical officer and vice president of informatics and analytics at Transcend Insights, a Campbell, Calif. -based provider of population health and analytics tools, notes that data analytics has always taken place at the hospital level. It has been used variously to track resource utilization, monitor patient safety and provide metrics on day-to-day patient care. But Van Gilder agrees with Kloss that, as healthcare providers "transition to performance-based care, what's missing is a horizontal look across those different silos and also across time."
At South Nassau, breaking down those silos to provide a more integrated view of the hospital's operations meant overcoming three major challenges. The first and most basic was coming up with unified definitions across departments for different data elements, along with common interpretations for what each definition meant.
While that might seem fairly straightforward, "At the beginning of the process," recalls CFO Bogen, "there were considerable differences of opinion among [South Nassau's] senior leadership team over the meaning of the definitions, the relative importance of the data and how to interpret it in a way that becomes actionable for the entire organization."
Greater consensus was achieved by relying on the hospital's newly deployed EHR system. "The EHR allowed us to provide the required data elements without ambiguity. It either was or it wasn't," explains Dick Rosenhagen, South Nassau's assistant vice president for electronic medical records and health information management. Today, he says, "The EHR serves as the repository where we collect the information that we ultimately use in our reporting."
The second big concern was determining how best to present the data to South Nassau's governing board so it could hold the senior leadership group accountable for meeting the goals and objectives set by the group on an annual basis. "Although we talk about how the numbers speak for themselves," Bogen says, "it took some time to define the rules of the road."
The third difficulty lay in adapting the hospital's workflows and procedures to collect the data that was needed.
After the EHR was in place, the acute-care facility moved ahead quickly with certain data-gathering requirements, at Bogen's insistence. The CFO's aim was to ensure that all meaningful use dollars that had been included in the hospital's annual budget were collected in the same fiscal year. Accomplishing this provided South Nassau with an additional $11 million in federal dollars over a four-year period.
But because the data-gathering processes were implemented so rapidly, several of them weren't working to everyone's satisfaction upon completion.
For example, under the second stage of meaningful use, South Nassau needed to get 5 percent of its patients to agree to enter, share and receive medical information through the provider's new patient portal. It began the effort by recruiting volunteers to go room to room and solicit patients to make use of the portal. But this proved more difficult than anticipated, and Bogen soon realized that the volunteers lacked the necessary training. It became apparent that for the provider to reach its 5 percent target, it would need dedicated staff assigned to the effort. Ultimately, the organization hired and trained a new group of employees to work full time getting patients to view, download and transmit their medical data online.
All told, says Bogen: "We've spent the past three years continuing to refine the processes by which the EHR data is collected, how it's defined and ultimately analyzed."
North Shore-LIJ faced similar challenges-on a bigger scale.
To integrate the data generated by different hospital functions and buttress its overarching strategy of providing patient-centric care, in 2013 the sprawling health system launched an analytics initiative known as Care Solutions. The two-year-old business unit has four primary responsibilities:
* To monitor North Shore's value-based outcome performance goals, as stipulated in its contracts with private insurers and government programs, including Medicaid.
* To continuously compare those outcomes against relevant costs and resource utilization levels to ensure that the health system remains profitable.
* To provide the performance-based metrics needed to demonstrate to payers that the agreed-upon outcomes have been met.
* To provide feedback to doctors and other practitioners about their performance with respect to their patient-outcome objectives.
The core objective within all this, says informatics exec Kraut, is using insurance claims and electronic health record (EHR) data to identify and better manage those patients at greatest risk. Among the approximately 200,000 North Shore patients covered under value-based contracts, "10 percent represent 60 to 70 percent of our spend and resource utilization," he notes. So improving outcomes and lowering treatment costs among this pivotal group is the key to meeting performance goals and maintaining profitability.
In the greater Long Island area that North Shore serves, for example, 30 percent of all emergency department visits are for non-emergencies, and 30-day hospital readmission rates range from the high-teens to the low 20s in terms of percentages of patients.
By reducing both types of occurrences among the hospital network's high-risk population through close patient monitoring and timely support, Kraut expects the healthcare provider to be able to substantially reduce unnecessary resource utilization and its total cost of patient care.
To accomplish these twin goals, North Shore's Care Solutions team has developed what it refers to as "the analytics lifecycle," which draws on many millions of data points to better comprehend the disease burdens facing its patient population and the demands these place on hospital resources. Capturing patient data is at the heart of the effort.
"It's very difficult to manage what you don't have data about," says Joe Schulman, Care Solutions' executive director.
At the start of the cycle, North Shore receives claims data from any public or private insurer with which it has entered into a risk-sharing agreement. As part of the contract negotiations, an understanding is reached on what data the payer will provide, the structure of the data, how it will be delivered and the frequency of that delivery.
After the data has been received and structured, it's fed into one of two analytics engines-the web-based Optum Impact Suite from Optum and Explorys, a cloud and Hadoop-based big data platform that was developed by the Cleveland Clinic and acquired in April by IBM.
The engines synthesize claims and socioeconomic data with "transactional" or clinical data points, such as physicians seen, services received and medicines prescribed. The combined data is used to analyze or "stratify" the patient population by age, diagnosis and other demographics. These analyses are then used for planning purposes, to optimize hospital resource allocations, and to provide comprehensive pay-for-performance reporting.
So will providers like South Nassau and North Shore continue to place greater emphasis on analytics in their decision making?
Analytics should be viewed "as more of a tool than an end place," says Bogen of South Nassau. He calls it "a big mistake to think that analytics in and of itself will provide all the answers. You just can't look at statistics and automatically draw a conclusion. "Many times," he says, "the data isn't conclusive on its face and needs to be part of a larger process."
For his part, North Shore's Kraut simply puts it this way: "We're on a journey, where we're learning to harness the data we have-and preparing to make use of the data we're about to get."
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