How to Create a Healthcare Data Culture

WHAT IS A DATA CULTURE, AND WHY IS IT IMPORTANT?

To understand data culture, one must first understand data analytics.

Data analytics allow health systems to understand what's going on within their organizations, whether in terms of resource usage, employee retention, infection control, or care delivery. Using this data, systems can potentially spot problems and predict outcomes.

"The main purpose of [analytics] is to make the organization smarter, to make folks understand, really, what's going on, what has happened, and what's likely to occur through the use of the hard data, from the EHR and other sources of information that we collect. So, data culture, to me, is how the organization deals with data," Watkins explained.

Watkins broke data culture down into three key aspects: the value people within the organization place on data, literacy around analytics within the organization, and how well analytics reaches across the organization.

The value placed on data impacts how it is leveraged in decision-making processes and data governance strategies, Watkins said. If there is disagreement within an organization about the extent to which data is a strategic asset, it can be more challenging to use to generate insights that could improve health outcomes.

The level of data analytics literacy in an organization refers to peoples' ability to read and interpret analytics derived from data. Without sufficient literacy around data analytics, the insights generated wouldn't be useful because people within the organization can't interpret them or use them to develop strategies effectively.

An oft-quoted expression, 'data is the new oil,' illustrates the idea of how well analytics can reach across an organization, according to Watkins.

"Oil is very, very valuable," he said. "[Data] is very, very valuable, but also, it spreads across an organization. And if it spreads in the correct way, [the] gears of the organization don't grind. It smooths out the organization as a whole, as an engine, and enables really good decision making."

CREATING A DATA CULTURE IN HEALTHCARE

Before organizations can oil the gears of their machine, so to speak, they need to establish the key principles of their data culture strategy and delegate someone, such as a chief analytics officer, to spearhead that strategy. This person both helps guide the implementation of the strategy and advocates for it to health system executives.

As part of Stanford Children's Health's data culture strategy, Watkins and his team have focused on data governance. They employed three strategies under the data governance umbrella: high-level executive steering, which involved getting the organization's leaders on board; other governance, which included looking at the scope of the initiatives the organization is undertaking and what's on that roadmap; and standardization, which involved making sure all stakeholders are 'speaking the same language' and have similar levels of analytics literacy.

In terms of data infrastructure, Stanford Children's Health has adopted a federated approach. This "hub-and-spoke" model, as Watkins calls it, allows a central team to create the infrastructure, develop data sources, and organize the data. The 'spokes' of the wheel enable different departments within the health system to access localized analytics relevant to their specific services.

THE CHALLENGES OF CREATING A DATA CULTURE

Establishing trust, showing value, and creating the right atmosphere are critical for a successful data culture, according to Watkins. The three are closely connected, but addressing them can be a challenge for any health system.

Trust is often a hot topic in any conversation around data, and data culture is no exception, especially in the context of an organization's internal politics.

"Very often, data can be a political thing. I've seen that in other organizations where people try to hold on to their data… I think access to information is power in some ways. [These organizations] hold onto siloed information, and that really can be to the detriment of the overall big picture of analytics. So, we're trying to make it as apolitical as possible," Watkins said.

Trust can be achieved by deploying an enterprise strategy focused on making sure the organization is supported, rather than the strategy being specific to one department. Watkins noted that having this broad perspective around creating a data culture and engaging in initiatives that show this strategy in action has greater potential to establish trust.

Showing value can also be achieved by employing this whole-organization approach. Those involved in data collection and entry are not the same as those handling analytics and interpretation, but they are still crucial for a strong data culture, and thus need to see its value.

"There is a huge burden on a lot of clinicians, both physicians and nurses, in terms of data entry into the system, so you have to show value… For folks who are actually doing the data entry, it has to just mean something. Now a lot of times, the data that's presented in the chart is used for reviewing that particular patient. But from our perspective, we then look at a lot of this information and see, 'okay, what are the clinical outcomes across a broad cohort of patients? Are we doing the right things that fit them?' so to speak. That's my role in analytics, to be able to find that information," Watkins stated.

The information that clinicians gather, once analyzed, can be presented back to them to help them provide better, more effective care. It also shows them that their data collection efforts serve a broader, more meaningful purpose than just increasing their already heavy burdens as care providers, Watkins said.

Creating the right atmosphere combines trust and value to foster partnerships across the organization and data chain, ensuring that the data culture isn't just a short-lived initiative.

To establish a long-term data culture, health systems must first take on the two or three projects that provide the most value for the organization, Watkins said. In doing so, teams across the data chain forge relationships that allow future needs to be identified and addressed more easily via informal networks instead of forced collaboration.

"It's really about the people — it really is," Watkins said. "When you have good people and good people who are connected to the mission, they'll do wonderful things, and they'll help make the organization smarter. So, it's our job to make sure the conditions are right for them to be able to do that."