Health systems increasingly recognize data as one of their top strategic assets, but how many organizations have the processes and frameworks in place to protect their data? Without effective data governance, organizations risk losing trust in their data and its value in process and outcomes improvement; a 2018 survey indicated less than half of healthcare CIOs have strong trust in their data.
By following five steps towards data governance, health systems can effectively steward data and grow and maintain trust in it as a critical asset:
1. Identify the organizational priorities.
2. Identify the data governance priorities.
3. Identify and recruit the early adopters.
4. Identify the scope of the opportunity appropriately.
5. Enable early adopters to become enterprise data governance leaders and mentors.
Download
As data gains more prominence in healthcare as a top strategic asset, health systems are increasingly understanding the importance of a data governance strategy. Healthcare analytics leaders are learning, however, that there’s a big leap from knowing they need data governance to implementing a strategy. Even expert data analysts benefit from guidelines as they move towards data governance.
This article discusses the importance of data governance in healthcare and how organizations can transition from simply doing analytics to doing analytics under a strategy that safeguards and maximizes one of their top assets—their data.
According to the HealthIT.gov, by 2015 more than 80 percent of U.S. hospitals had figured out how to collect their clinical data. But, just as health system data has increased, so has mistrust in their data. According to a 2018 survey, less than half of healthcare CIOs have strong trust in their data. They often don’t know how their analytics are derived or on what data they are based, and they rarely have any easy way to alleviate their concerns.
Balanced data governance builds trust and effectiveness. In years past, analytics teams emphasized data security. Protecting data was paramount, and it continues to be important today. But what good is data if only a few people can use it? To paraphrase Henry David Thoreau, that governance is best which governs least. Effective organizations secure their data from misuse. Beyond that, their job is to liberate data and facilitate its best use by everyone.
As a subject with many facets—from data security to data quality to data stewardship—it’s beyond the scope of this article to define the details of data governance. Data governance is more than a little overwhelming, so much so that many organizations struggle to get started; or they do start, but the scope is too broad, making it difficult to gain traction.
According to data management expert Will Bryant, “Surveys report that as many as two-thirds of all initial data governance efforts die on the vine.” Bryant explains that these efforts “mushroom into such complexity during the planning and design phases that they are abandoned before they have the chance to deliver any value.”
Healthcare organizations can navigate the complexity of data governance set-up by following five practical, manageable steps:
For analysts, report writers, and other data practitioners, it’s often easy to identify the data problems. But deciding where to focus data governance based on data awareness alone is usually a recipe for failure. Governance done for the sake of governance is a hollow endeavor that rarely finds enough leadership or grassroots support to gain or sustain momentum. If it does, it often becomes a law unto itself with little regard for the analytic needs of the organization. Such governance may be worse than no governance at all.
The purpose of data and analytics is to serve the strategic goals of the organization. So, the first step in developing a new data governance program is to identify the organizational priorities. This requires an understanding, for example, of the top five targets the executive team wants everyone in the organization to help deliver over the next year. These could include lofty goals, such as “increase patient engagement” or “improve performance-management analytics.” They might also include more specific goals, such as “decrease adverse drug events by 5 percent” or “increase patient use of telehealth services by 9 percent.”
A simple list of the organizational goals, however, is likely insufficient. Effort should be made to understand the rationale behind the goals, the discussions that led to them, who was involved, and the nuance of their positions. (Understanding these last two points will be important for step 3.)
Step 2 in developing a new data governance program is to identify the data governance opportunities that overlap with the organizational priorities step 1 identifies. These organizational priorities may even have been developed into improvement initiatives already.
All analytics depend on timely and reliable data, so it’s likely that finding overlapping data governance opportunities will be easy, especially in organizations whose data governance is still maturing. For example, if decreasing adverse drug events were one of the organization’s priorities, proposing the creation of an accompanying data validation process to ensure the accuracy of the analytics would be an easy sell. Look for an implicit data governance prerequisite such as this and call it out
Keep in mind that enterprise data governance is the end goal of these efforts. If possible, choose data governance opportunities that are relevant beyond the initial, narrow focus. It may be appropriate to scope narrowly at first. The new data quality validation process in the above example could be focused on adverse drug event data initially, but this should only be the first draft of what would eventually become a shared process to effectively validate data quality anywhere in the organization. Otherwise, it may become just one of many redundant related processes throughout the organization that gets reinvented every time another team needs something similar. Such variability is not only very wasteful—it’s also virtually impossible to govern.
The next step is to survey the list of intersecting priorities identified in step 2, note the accountable leaders for each and identify which of these leaders are likely early adopters of data governance. Energetic leaders who truly understand the importance of data governance should be top of mind. Once each group establishes its governance objectives, these leaders often naturally become the champions of the organization’s new data governance program.
Creating a list of desired characteristics for a data governance leader launches the search for early adopters. Geoffrey
Moore’s book Crossing the Chasm offers valuable insights into important characteristics of effective early adopters:
As the list of candidates grows, some questions that probe each person’s understanding of the benefits of data governance will help narrow down the list to those most likely to carry the data governance program across the finish line. Consider questions such as the following:
Many organizations tackling data governance try to do too much too soon and get bogged down. It’s important to scope the opportunity narrowly enough to get traction with a small pilot team that is nimble enough to improvise and redesign processes on the fly. Focus on a specific data governance opportunity (e.g., data quality, data usage, etc.) within a specific area (e.g., clinical domain, patient population, geographic region, etc.) to target the effort on a manageable subset of the organization’s data governance—versus trying to fix all of it at once.
If the data governance team chooses one of the loftier organizational goals, such as “increase patient engagement,” it will need to refine this into something more precise, such as, “increase usage of the patient portal by 10 percent of patients in the Tri-Cities region.”
This is psychologically and physically pragmatic. It renders the problem more familiar by focusing on the team’s immediate domain (e.g., patient-portal data for the Tri-Cities region). It also lowers the amount of data, oversight, and approval that might otherwise be needed and could impede the team’s progress
As the first couple of teams achieve their data governance objectives, it may be time for them to generalize their processes for broader adoption. This shift in perspective from local to organizational naturally coincides with each early adopter’s transition from team leader to enterprise data governance leader. With new success under their belts, they are well-positioned to champion data governance generally and to recruit and mentor others; their message is not just conceptual—it’s based in their own hard-won experience, which gives them instant credibility and a host of concrete examples to drive home the how and why of data governance.
By following these five steps for doing data governance, a small team can kick off a data governance program, marry it with an organizational priority, possibly in the form of an existing improvement initiative, and make progress toward enterprise data governance. By recruiting enthusiastic early adopters to champion the program, starting small, and focusing on the overlapping organizational and data priorities, data governance will be widely recognized as a vital tool for advancing the whole organization.
Would you like to learn more about this topic? Here are some articles we suggest:
Would you like to use or share these concepts? Download the presentation highlighting the key main points.
