Mastering The Data Normalization Cycle

By Brian LaBerge, Health Language

A healthcare provider or payer that goes through a data normalization project will want to make sure its investment in time and money continues to pay off down the road.

You’ve obtained executive buy-in, conducted a data inventory, identified key constraints, prioritized projects, established a governance structure, and embarked on your first initiative. Ideally your project resulted in more consistent and useful data, with demonstrable benefits for your organization. Your next order of business is to make sure your useful data stays that way. A healthcare organization that fails to continuously monitor its normalized model runs the risk of returning to its previous condition: data chaos.

The data normalization lifecycle doesn’t end with the completion of a project. Even with the best of intentions, institution-wide adherence to data normalization guidelines may decline over time. The inputs to be normalized will also evolve: you will continue to find new, unrecognized input data in your data feeds. Your target standards will evolve as well: industry standards are constantly being updated, and you may also need to update your local terminologies. Finally, because medical knowledge, coding guidelines, and semantics evolve, you must periodically review and update the guidelines you use to normalize input data to target standards.

Here are a few considerations for keeping your organization on track during the data normalization lifecycle:

Perform Periodic Checks

A health system should conduct periodic checks of its data to ensure the standards it has adopted are being universally applied. An organization’s data governance structure plays the leading role in this effort. Members of the project team may undertake governance duties: project managers, system owners, and data warehouse managers, for example. A higher-level governance committee may also become involved, particularly when standards need to be enforced across the enterprise. Executive-level support of the data normalization project remains important, even after it moves from development into maintenance.

When you perform your review or audit, make sure the organization has updated all of its systems and confirm that naming conventions and standards have been followed. If individual departments revert to older informal conventions, it won’t be long before data incompatibility problems resurface.

The frequency of data reviews will depend on the volatility of an organization’s data. If the data changes frequently, as with HCPCS, a health system might want to consider a monthly re-check. If the pace of change is slower, as with ICD-10-CM, a quarterly or bi-annual review may suffice.

Periodic checks also foster a spirit of continuous improvement. An audit provides an opportunity to make sure the data normalization process is working for everyone in the organization. This may uncover quick-fix workarounds that were expedient at the time but don’t meet broader organizational needs for interoperable data. It may identify opportunities to train new staff or provide refresher training for existing staff. It can also identify elements of the data normalization project that have been causing dissatisfaction. Without the attention that periodic checks provide, people who encounter obstacles in data normalization will sidestep or abandon the process. Eventually entropy takes hold and your tidy data will once again become disorderly.

A governance committee typically defines the methods and frequency of periodic checks. While it is important to raise general awareness of the benefits of data normalization for employees and managers, it is equally important to establish a formal governance process. The governance process will define how clinicians, IT managers, and coders create new values for the data they want to represent. That process helps keep everyone on the same standards page.

As the scope of your data normalization projects expands, it may be helpful to use tools to expedite and manage multiple diverse projects. Tools such as Health Language’s LEAP Map Manager increase mapping efficiency, maintain mapping consistency, and facilitate documentation of governance decisions.

Make It Count

A data normalization project represents a significant time commitment for a healthcare organization’s personnel. My goal in this series of blogs has been to help you achieve an initial successful data normalization project that lays the groundwork for broader, even enterprise-wide efforts. Certainly these projects require some investment by multiple departments of your organization. But done well, data normalization will pay great dividends in healthcare. You’ll be able to make apples-to-apples comparisons of data from different systems that employ different terminologies. You’ll find the pursuit of collaborative care models much easier. And you’ll improve your reporting and data analytics, gleaning greater insight from your data stores.

I would be interested in hearing about your own data normalization journey. Have you completed your first data normalization project? What challenges do you face in preserving the progress you have made? Leave your comments below.

About The Author

Brian LaBerge joined Health Language in September 2010 and has serviced in multiple capacities including Technical Consultant, Project Manager and Director Of Client Implementations. He brings 18+ years of implementation and software deployment experience to Health Language. He has managed several Data Normalization and ICD-9 to ICD-10 projects for Payers and Providers during his tenure at Health Language.