Promoting improved data governance in the hospital and health system sector has become a just cause of sorts for Randy Albert, who serves as vice president of finance – operations and analytics at Maine-based Northern Light Health and is a member of HFMA’s Northern New England Chapter.
“This is a topic that I’m incredibly passionate about,” Albert said.
That’s why Albert, in his role as chair of HFMA’s Financial Analytics Leadership Council, spearheaded an initiative called “Data Without Walls.” The goal is to spread best practices for data governance, with a focus on ensuring data is applied consistently and holistically throughout a hospital or health system while removing the traditional silos (i.e., walls) in competing analytics requests.
The stakes are high: Based on feedback in HFMA member surveys, analytics optimization is among the most confounding issues facing healthcare finance professionals. And in HFMA’s latest member Outlook Survey, conducted in fall 2022, 35% of respondents said they would invest in data analytics over the next 12 to 18 months — the most of any product or service (talent recruitment and retention was second, at 33.2%).
Such investments are not guaranteed to pan out.
“Most times, [organizations] implement an IT system, and operations is geared up to use it and get going,” said Caroline Gay, FHFMA, CFO of the Rheumatic Wellness Institute in Miami and an HFMA member who has worked on the Data Without Walls initiative. “But we don’t put in the upfront work to define the data elements, to all speak the same language.”
She added, “I would put data governance in that category with cybersecurity of things that we don’t really want to spend time on, but they’re mission critical to run our organizations.”
Principles of a better approach
The Data Without Walls concept hinges on precise data governance. Decision-makers in every segment of an organization must use the same data definitions and the same approaches to converting data into analytics.
“You can do Data Without Walls — data governance — without a data warehousing platform,” Albert said. “You just need to establish the rules.”
At Northern Light Health, one key is a data governance council that brings together expertise in “the many facets of delivering data in an organization,” Albert said, including areas such as information security, infrastructure, programming and data warehousing.
More than any technological solution, the council is “the number one tool in my tool set,” Albert added.
Establishing the role of chief analytics officer also is important. Before moving into that position (albeit with a different title), Albert lacked the influence to institute systemwide standards for the 10-hospital organization.
Once in his current role, he had the authority to facilitate definitions for basic concepts such as patient panels.
“It no longer is a single person or group deciding, in a silo, this is the definition of patient panels,” he said. “Prior to being in this role, it was difficult going to the various hospital leaders and saying, ‘You’re counting patients this way, but our system definition is that way.’ There was no data governance policy to fall back on.”
When bringing together the perspectives of different departments or analytic functions, Gay said, the guiding principle should be how to best serve patients.
“What do we want to know about the patient? What questions do we have that we need to answer? And what data do we need to collect to answer those questions? If you come at it with that framework, you actually notice there’s a lot of data being collected that we don’t ever do anything with,” she said.
In a Syntellis Performance Solutions white paper that was based on the Data Without Walls initiative, one recommendation is to consider the following factors regarding the patient journey:
- Total opportunity cost (as calculated by comparing budgeted cost to actual cost)
- Cost variability
- Process variability
- Patient outcomes
- Patient experience
The foundation of optimal business intelligence
Intuitively, good data governance sets the stage for the caliber of analytics that can guide an organization’s strategic decision-making.
“When you have good data, it allows for the analytics divisions to do their job,” said Gay, who previously was vice president and chief analytics officer with Lakeland Regional Health in Florida. “When you don’t have good data, then your very highly paid and skilled analysts become kind of data monkeys, cleaning up the data, doing manual things.”
Yet successful data governance also allows analytics to be decentralized.
“If we’re all speaking the same language and have a high degree of trust in the data elements coming out of the system, then you can have more user-enabled analytics closer to the front lines,” Gay said. “You don’t have to have a highly skilled analyst if you have really clean data.”
Northern Light Health views analytics through the prism of transformation, Albert said.
“I’m bringing together people, process and technology to deploy effective and sustainable performance improvement,” he said. “And performance improvement is what has the ROI.”
An ongoing initiative to study care variation is a prime example. The output from Albert’s analytics team helps the study team pinpoint differences in quality outcomes stemming from care variation at different provider sites. The investigators will use those insights to continually align care protocols and are expecting an ROI in the millions, Albert said.
Organizations, he added, should think about the ROI of data governance and analytics programs as a return on transformation.
Implications for the big picture
The Data Without Walls concept has ramifications for industrywide interoperability.
When Gay was with Lakeland Regional, she sent data to the state of Florida as part of an initiative to maximize capacity during the pandemic. She was struck that “in a very high-stakes situation, we weren’t really even speaking the same language when it came to how to count the number of available beds in a hospital.”
Albert thinks the situation will improve through “shared thinking, through a different type of management structure that would allow for that interoperability to happen in a more collaborative way.”
For such a model to materialize, common concepts and definitions of data will be critical.
Said Albert, “Evolving the U.S. healthcare industry into a long-term sustainable model will require drastic standardization across all nonclinical spaces to give stakeholders a chance to remove some of the considerable excess cost from care delivery at scale.”
Editor’s note: For more information on “Data Without Walls” best practices, contact Randy Albert.
3 pillars of data governance in healthcare
A white paper authored by Syntellis Performance Solutions, formerly a sponsor of HFMA’s Financial Analytics Leadership Council, lists three elements of a Data Without Walls initiative:
1. Adopting a leadership structure and best practices for data and analysis
A key early step is introducing the role of chief analytics officer to oversee a dedicated data governance team and ensure data-related initiatives have visibility in the C-suite.
Organizations also should consider adding an analytics consultant, who can support operational and clinical leaders “as a partner and problem solver,” according to the white paper. Finally, executive sponsorship bolsters any data governance program. The white paper recommends “creating a steering committee that meets quarterly and brings together executives from financial, operational and clinical domains with IT.”
2. Collecting the right data and getting it to the right users
Foundational approaches include engaging in collaborative conversations, identifying valuable data, addressing data quality and standardization, and combining analytics and action (e.g., incorporating analytics in quality-related initiatives).
Core resources and processes that support data governance include a data warehouse, data scientists and engineers, robust reporting and role-specific dashboards that “empower individuals in a variety of roles to easily view” meaningful information.
3. Measuring the return on data and analytics investment
Key questions to consider when gauging ROI include the impact on specific financial and performance improvement goals and as a strategic decision-making tool.
The end result conceivably is a progression from descriptive to prescriptive analytics — i.e., from examining why something happened in the past to applying insights proactively.
“With prescriptive analytics, healthcare organizations stand to make significant strides in patient care, quality outcomes and operational performance by more accurately predicting and planning for what’s ahead,” the paper states.