How a six-step process can improve data quality

With an avalanche of patient data from external sources and the demand for real-time operational insight, effective data governance has become mission critical.

What follows are six critical steps that data stewards should be able to undertake with a quality Enterprise Master Patient Index (EMPI) in order to create and maintain the highest-quality view of the patient—leading to high-quality, comprehensive care for patients and cost-savings for hospitals.

1. Assess the quality
It’s crucial to gain a good understanding of the quality of the data that is available and to understand the source of any limitations. Organizations may, for example, have systems where the data is gathered via forms that are transcribed. Transcription means errors may creep in. If, for example, information is being gathered via phone, a name may be heard and entered incorrectly, or “15 Main Street” may sound like “50 Main Street.” It’s important to assess front-end systems and understand the quality of the data being generated.

2. Integrate data
An organization needs high quality insights, and this means integrating disparate information to build a “golden view” of the patient that brings the data together. This involves:

• Maximizing benefits, minimizing cost—Consider the priority of the data. Which information is crucial to creating the golden view and getting business done?
• Implementing data priorities—Lay out a model, or template, of how data should be rendered within the system.
• Implementing integrations in line with rate of change; real-time, overnight batch, or once a month—If the data won’t change on a daily or weekly basis, there’s little value in doing real-time work.

3. Discover a good match
Explore how the data will interact between systems, how it will blend to create the golden record of each patient, providing a more complete and accurate picture. This process also offers the opportunity to discover relationships hidden inside existing data sources.

An organization will set up that initial template, then start running data through that template to see how the data is matching—how much of the data should be in the final golden record already exists in the data? Can all the disparate information be brought together on a patient called Kelley Smith and match it to create the golden record? Can an organization differentiate the Kelley Smith on Main Street from the one on Park Avenue?

4. Manage ongoing changes
From this point forward, an organization will continually be assessing the data and stewarding it for maximum success. Data is ever in flux, with new patients, changes of address, and changes in organizational systems. It’s necessary to have the right processes and technology and to establish the business rules to maintain the quality of data in real-time.

In an ideal situation, an EMPI system now has enough confidence in the raw data that it can automatically merge all the records together into the golden view. However, oftentimes the reality is that an organization will have some confidence, but not enough for the system to merge the data automatically. In that case, it would be put to the data steward of an organization to assess the data as an appropriate match of information and then merge it into the golden record, or to investigate for more information. It is critical that an organization maximize the automatic merging in order to reduce the amount of stewardship required while maintaining confidence in the data.

5. Optimize data
If only internal data is used, the level of quality is capped. By using third-party reference sets, an organization can achieve a much richer data set.

There are three ways that the use of third-party sources can help improve a golden patient record:

• Verify Information—This may involve sending out an address (stripped of name or other sensitive patient information) and finding out whether it is accurate.
• Enrich information—In this case, an organization can provide incomplete information to a third party and get back a more complete view of the patient. This is most often used for activities such as analytics and marketing. For example, an organization could provide a third party with a name and address (no medical information) and they might provide demographic information such as household income and education level.
• Access brand new information—Many individuals in the US change address often. Some EMPI solution partners are now working with the US Postal Service to offer features that enables providers to proactively update address information monthly, whether or not the patient has reached out or made contact.

6. Share information downstream
To realize the true value of the newly developed patient golden record, it’s important to efficiently push that information back out to the source systems in an organization.

That involves setting up “business rules” for fine-grained control of this process. If there’s a high degree of confidence in a particular data field that has come from a source system (for example, if the billing system already has accurate address information), then an organization should be able to implement an automated business rule that pushes that address information back out to all the other systems in the organization. This enables each individual system to enrich and build out their golden record of each patient.

The best road to success between a solutions partner and a hospital or health system has the following landmarks:

• Transparency –actionable insights into the data that allow users to gain critical perspective and make informed business decisions
• Empowerment –an intuitive, self-service interface that allows the technology to fade into the background while the data and insights come into sharp focus
• Speed-to-value –the ability to see measurable results and return on investment quickly, without complicated onboarding or a steep learning curve

Choosing a technology partner that will provide transparency at each step along the journey is necessary to gain buy-in and extract the value hidden in existing data assets.