Common Mistakes in Healthcare Data Governance – Part 1

Jul 7, 2021

INTRODUCTION

Healthcare data governance is difficult. And frankly not very exciting. Decision-makers often express appreciation of those who develop a flashy new report or deliver a new insight based on analytics, but they rarely see the amount of work it takes to ensure the underlying data is valid: resolving conflicting definitions, tracking down data integrity issues, and monitoring data quality.

While it might not get the glory, data governance is foundational to all other data analytics success in an organization. And avoiding a few common mistakes that we’ve seen in healthcare organizations we’ve worked with will help increase your chances of success significantly.

Here are three common mistakes found in healthcare data governance:

Mistake #1: Making Enterprise Data Governance an IT Project

Based on name alone, the information technology (IT) or information services (IS) teams seem like they would be a good fit for anything data and analytics-related. It’s bits and bytes, right? However, as data becomes more and more of a key strategic asset, organizations are realizing that data management and analysis functions and capabilities must be more closely tied to the clinical and operational processes they analyze. It’s no longer enough for IT to gather requirements from the users and then go build. Instead, IT needs to enable and support the users to apply their clinical and business knowledge while they fish for themselves.

In the same way, data governance has to be a business-led, IT-supported endeavor to be effective. Those closest to the clinical or operational data subject matter are best equipped to create definitions, validate data, identify correct and incorrect results, and determine the priority by which data issues should be addressed.

In addition, if clinical and operational stakeholders don’t have specific data governance responsibilities, they are unlikely to be meaningfully engaged and even less likely to be proactive. Having a cross-functional data governance steering committee isn’t enough—there must be cross-functional data governance execution.

Mistake #2: The “Everyone But No One” Problem

The natural inclination of data governance is toward disorder. While some organizations do a great job of developing and communicating a compelling vision for shared data governance responsibilities, they can eventually drop the ball with day-to-day operations. Over time the attention of the organization moves to other issues. Things like filling in descriptions for reports and calculations, ensuring healthcare metrics stay in sync, and data validation fall by the wayside.

When everybody has responsibility for data governance, ultimately nobody has responsibility for data governance. That’s why there needs to be at least one person in your organization who has data governance facilitation in their job description. Ideally there is a team of data governance facilitators, but even having a partial FTE devoted to “minding the store” of data governance is a good start to help the program continue to make progress.

If you’re interested in deploying a data governance facilitator, their daily tasks include monitoring the data catalog to ensure definitions are populated, interfacing with data stewards to keep their responsibilities top-of-mind for them, and reporting issues, challenges, and overall progress to the data governance steering committee.

Mistake #3: Assuming “If You Buy It, They Will Come”

Apologies for mangling the famous movie line, but some organizations believe that if they purchase a market-leading data catalog and metadata management tool, people will finally take enterprise data governance seriously. They hope the costly investment will force the organization to take notice and take action. Not surprisingly, the desired results do not come about simply due to throwing money at the problem, and the organization is left with a large bill and an even steeper hill to climb when they take another run at data governance the next time.

Instead, a much more holistic and pragmatic approach is needed. Definitions, calculations, and lists of available reports, dashboards, and metrics need to go somewhere; otherwise, the organization can’t leverage them, making a data catalog essential. However, overspending on a product in which only 50% to 80% of the functionality will be used any time in the near future is not a good way to way to get the organization on board. A better approach is to acquire a more affordable hospital data catalog in line with realistic needs and spend the additional money on data governance facilitators, data literacy education/support, and internal marketing.

And of course it’s not just about money. Data governance is challenging and it’s a project that is never done. So while you can make some progress by setting up a data catalog, making definitions visible and business logic transparent to the organization, and then using these things to help “sell” the vision, ultimately without strong leadership support the data governance program will not progress very far. Like most endeavors, you must move forward on all fronts of people, process, and technology to see data governance take hold in your organization.

___________________________________________________________________________________

Learn more about how the Compendium Data Catalog can increase both the trust in and the efficient use of data in your organization.