Preparing Data for AI Implementation in Healthcare

Evaluate Your Current Data Maturity

The size of an organization may impact its approach to data (whether there are enough team members or financial resources to manage data, for instance), but the key factor is overall data maturity, whether you have established data governance or existing analytics capabilities. Larger organizations may be at an advantage in some cases, but smaller organizations can also excel if they have more advanced data maturity.

There are a number of frameworks that organizations can use to assess their readiness. The updated HIMSS Analytics Maturity Assessment Model aims to help providers prepare for AI adoption. Gartner also offers more general benchmark resources. Working with a partner is usually the best way to evaluate your approach to data now.

It may also be valuable to augment and channel your applications into a modern data platform once you’ve leveraged as fully as possible the data management and analytics capabilities that your core vendors have enabled within their solutions.

It’s also time to rely more on the cloud and more updated approaches to collecting, managing, storing and moving data through data platform modernization.

Focus on the People Aspect of Data Governance

Data governance lays the foundation for how data is treated as an asset in an organization, and that involves how data is managed, protected and used. It should not be a supplementary consideration; it needs to be a core part of an organization. It also allows business and technical teams to connect more on data and clarifies and redistributes responsibilities.

Organizations can usually mature data governance alongside AI governance, as AI solutions require data to be effective. AI governance includes standards and approaches to account for bias, transparency and risk associated with a tool; these align with the principles and practices of data governance.

And the human aspect of governance cannot be overstated. Organizations need to communicate with and directly involve stakeholders who will be relying on such solutions and requiring data. What training and education is needed to prepare a workforce for an AI implementation? How can a solution free up team members to shift from rote tasks to higher-level work? What is the process of evaluating a solution for specific workflows?

Part of this change includes a culture shift. While it’s normal and expected for team members to be apprehensive about new technologies, organizations need to clearly communicate expectations for AI and try specific use cases. Foster an environment that will be open to change rather than fearful of an unknown future. Articulate use cases such that you can estimate ROI, and then follow through and actually measure that ROI.

Being able to convey the meaningful connection between a technology and the business or clinical processes that it will impact is a fundamental skill that every organization will need to get better at in order to be successful with data and analytics efforts going forward. 

This article is part of HealthTech’s MonITor blog series.