Unlocking the value of healthcare data with IBM technology

By | 3 minute read | September 27, 2018

healthcare data on tablet

Garbage in, garbage out. It may be an old saying, but it’s as true today in healthcare data management as ever.

Bad data generates erroneous conclusions

We’ve always generated a lot of data in the healthcare industry. In hospital systems, every department produces its own patient data, and, until recently, we weren’t looking at that data holistically. Now the trend is toward using and analyzing that data in areas like population health. That requires creating a unified picture of individual patients – from admissions to treatment to discharge and, finally, billing. But if we haven’t been standardizing and defining data across departments, we aren’t going to get an accurate picture of the patient, never mind an entire population of patients.

Healthcare organizations are turning to new applications for data analysis, including AI, machine learning and advanced analytics. Often the tools require training in order to find meaning in vast amounts of data. But if you train a tool with wrong or inaccurate information, you’re going to go down a very wrong path. It certainly won’t lead to the trusted, repeatable process and outcome that you look for from implementing machine learning.

At Atrium Health, we’ve decided that information governance must provide the foundation for any analysis or use of data, whether for clinical or business purposes. A governance program and appropriate tools can help us understand the context of our information and how the business uses it. We’re also building a master data management (MDM) strategy and launching a data quality program.

Data governance drives data standards

Technically, we could implement any number of tools to support information governance, but will those tools really meet the needs of the business organizations within Atrium? Can we use them to provide data in a timely fashion? And, probably most critically, can we use them to define and use data the same way across the entire organization?

Atrium had already worked with IBM to implement our enterprise data warehouse (EDW) on IBM Netezza, with the IBM Unified Data Model for Healthcare. We saw a great benefit to further implementing IBM Data Governance tools, specifically IBM InfoSphere Information Governance Catalog and IBM InfoSphere Information Analyzer, to build the foundational layer for information governance. We’re now implementing IBM Master Data Management on IBM Cloud.

Our data governance program gives us a holistic view of patient information across the system. The same, standardized data can be used for patient care, billing, analytics and population health with this holistic approach to information management.

Consistent data creates correct conclusions

The point of implementing a data governance system and standardizing data is to drive changes in decision making in your business processes, and that’s what we’re doing. Similarly, by improving data standards, we can move beyond managing data and focus more on using and analyzing that information for better patient outcomes.

For example, before we implemented the IBM solutions, each of our data analysts spent up to 40 hours per project just validating the data. They didn’t trust the data they were using, which often had never been accurately defined or was inconsistent across business units. We’ve reduced that time significantly by providing trusted sources of data from the outset of the project.

We’ve gained a clearer understanding of the context of the information we’re using, such as how it’s collected, how it’s managed and where we have duplicate data. We can even see whether we’re using particular data for purposes for which it was never intended, which could lead to incorrect conclusions or actions.

Now if we want to apply AI or natural language processing analytics to our data, we can be confident that our data is consistent and accurate. We can be ahead of the game when it comes to training a system like that. The bottom line is that the conclusions we reach from data are only as good as the data itself, and data governance is critical to avoiding the trap of “garbage in, garbage out.”

Shannon Fuller explains how Atrium Health is unlocking the value of its healthcare data with IBM InfoSphere and Master Data Management solutions: