Keywords Data Governance and Technology

Date posted: Tuesday, September 07, 2021

Category: Security

For those of us working in healthcare data and analytics, if the COVID-19 pandemic showed us one thing, it was that we need to be able to react more quickly to pressing needs. Decision makers have been complaining about the amount of time it takes to get a new report or data mart for years, but COVID-19 was the straw that broke the camel's back and finally brought out into the open the ineffectiveness of the ways in which we currently produce data and analytics in our organizations.

Information that was not a high priority before the pandemic became a high priority almost overnight. For example, executive leadership at Mille Lacs Health System in Onamia, Minnesota had been satisfied with volume statistics from across inpatient, outpatient, and long-term care services that were delayed by several weeks. But that changed fast when COVID-19 came along. Now, much more current information on volumes was needed so that the organization could quickly adapt their staffing levels.

"We didn't have a way of getting volume information without physically looking at logs," Amy Ninham, Health Information Manager for Mille Lacs Health, said. 

It wasn't because the data didn't exist in their data systems-that particular set of data just hadn't been prioritized for automated reporting yet. The manual effort of reviewing logs and compiling the data in a spreadsheet no longer met the urgent needs of the organization. 

Fortunately, due to their smaller size and more nimble processes, Mille Lacs Health was able to pivot their priorities quickly and automate much of the data and analytics they needed. However, this is not the case at many, if not most, healthcare provider organizations. When an emergency directive comes down from on high, it causes no shortage of chaos in the data and analytics teams, and this needs to change. 

For some of us, this means we need to rethink the entire way we operate. The old way of having everyone submit a ticket to a large central team, which then prioritizes those tickets and works through them as their resources allow, doesn't work anymore. Actually, it never worked for the people whose requests never moved up the prioritization list. Often, those requests only get closed when they are deemed old enough that they must not be needed anymore. The truth is that the data was needed by someone at some point, and was probably still needed, but they just didn't have the political capital in the organization to ever get their request moved up.

The answer to the one-team traffic jam isn't self-service analytics. While that is an important component of the overall data and analytics strategy, traditional self-service mechanisms still require a central team to develop the data model and elements for users to choose from, and that is still a bottleneck.

So, what can be done to increase the "speed to market" of data and analytics in order to respond to not only emergencies like COVID-19 but the quickly evolving needs and priorities of healthcare providers in general?

What is needed is to stop treating report writing, data analysis, data mart/data warehouse development, and other data and analytics disciplines as separate activities performed by a small number of siloed teams. Instead, we need to move toward an interconnected, collaborative effort that cuts across developers, teams, and departments. Remove the bottlenecks by opening more avenues for more people from across your organization to participate in building your collective information wealth.

This isn't just pie-in-the-sky dreaming. This is the goal of a discipline called data governance. Enterprise data governance is the discipline of defining, monitoring, and managing data from initial capture all the way through to use of the data to make decisions, across the entire organization. 

Done right, data governance provides the structure to allow the kind of cumulative, cross-organization effort that advances data and analytics capabilities much more quickly than old methods. A key tool in the data governance toolbox is the enterprise data catalog, which is a central repository of definitions, calculations, and lists of all the reports, dashboards, and other information assets from across the organization. The data catalog provides ways to share codes among disparate teams, standardize metrics, and monitor for discrepancies to help prevent the analytics landscape from becoming the Wild West.

As a user-friendly atlas of your organization's data and analytics world, the data catalog also encourages less-technical and non-data-savvy stakeholders to join the cause for more timely and better trusted data. Ostensibly, you don't have to worry so much about data governance when everything is funneled through a single central team, which works closely together to solve problems, and ensures their definitions and calculations are aligned. However, even if it was true that close proximity ensured good data governance practices, the central team would still be missing a key perspective: that of the subject matter experts!

In the context of data governance, these subject matter experts are referred to as data owners or data stewards. A data analyst might be able to tell that data is being pulled from the database properly, and that numbers are calculating as expected, but it takes a data steward to know if the numbers accurately reflect workflow or clinical quality or operations. Involving the right people in the process of definition and validation drastically increases the integrity, efficiency, and speed of turning data into needed insights.

However, not all hospital systems have the right people in place. Think of it like a traditional furniture store vs. IKEA. In the traditional store, there are a limited number of workers who can help customers, and customers normally have to come back to pick up their purchase in the future. At IKEA, customers are given all the information they need on the show floor: samples of the furniture, examples of how it can be used, the price, etc. Since customers can retrieve their furniture themselves on site, the cost savings and speed are significant compared to processes at a traditional store.

What if the IKEA method had no information on where to find the furniture in the warehouse, or worse, if there were no showrooms at all and the warehouse was full of unmarked boxes—that would be a disaster. That's essentially what most organizations are dealing with in their data and analytics. A "customer" can submit a request that gives an idea of what they are looking for and hopes their request gets routed to someone who can find the item in the warehouse for them.

Proper data governance makes information available so the whole organization can find what they need, when they need it, and leverage the data in appropriate, safe (aka, "well-governed") ways. The turnaround time for reports and analytics is quicker because existing assets are optimized, and more people-each leveraging their unique skills-are involved in the process of developing new information assets. This is the only way we are going to able to speed up data and analytics and more quickly respond to the next COVID-19.

Kevin Campbell is the CEO of DTA Healthcare Solutions, a healthcare data and analytics consulting firm. He has more than 17 years of experience in healthcare data warehousing, analytics, and data governance, both inside a large healthcare system and as a consultant. Ready to Improve Your Outcomes?