Big data challenge: Junk in means junk out

#5in5_Blog_BannerAccess to good, quality data is one of the most pressing challenges related to the use of analytics in health care.

If you input incomplete or “dirty” data, outputs won’t be reliable. Even the best algorithms in the world won’t work when applied to a distorted source.

To ensure you are crunching numbers in a way that will provide insights you can use to improve care and drive down costs, you need to start with clean data.

Clean data is data that is normalized. Normalization is the process of organizing data to remove redundancies and dependency. You identify what data should be included and its relationship to other data. This way you can compare values.

Providers also need to consider data linkage. That is the ability to track or link records that refer to the same subject across different sources, such as clinics, hospitals, pharmacies and laboratories.

Cleansing data remains a largely manual and thus expensive process. It is also an expense that doesn’t go away, because electronic health record data tend to migrate as users change the way they document information in the user interface over time.

Better analytics platforms address this up front and apply automation where they can to monitor and manage this process to ensure consistency at the lowest possible cost.

Read more about other big data challenges and opportunities to overcome them in the #5in5 episode Big Data: Big Opportunities, Big Challenges. You’ll read about data literacy and blending data into workflow.

The #5in5 series features five questions answered in five minutes on issues relating to the evolving health care industry.


About the Author:

Reti-AlejandroAlejandro Reti, MD, MBA
Chief Medical Officer, Optum Analytics

With responsibility for the Office of the CMO, Alejandro is accountable for the clinical integrity and relevance of Optum Analytics’ provider solutions and contributes to thought leadership and clinical product innovation for the organization. Alejandro came to Optum from Premier, where he served as Vice President, Population Health Products with general management responsibility for Premier’s organically developed population health suite. Prior to Premier, Alejandro served as Senior Vice President, Clinical Informatics at Verisk Health, where he led development of a provider analytics solution that achieved top 4 in market share nationally. Previously, Alejandro held positions of increasing responsibility at Avalere Health and The Advisory Board Company. Alejandro received a bachelor’s degree in Psychology from Amherst College, magna cum laude and his MD and MBA degrees from Yale University.

Leave a Reply