Getty Images/iStockphoto

How An HIE Approaches Continuous Patient Data Quality Improvement

An HIE in Tennessee has tapped an AI-based platform to reduce duplicate patient EHRs as part of its patient data quality improvement efforts.

The cornerstone of a health information exchange (HIE) is continuous patient data quality improvement, according to Pam Matthews, chief executive officer and executive director of East Tennessee Health Information Network (etHIN).

"While an HIE may provide a lot of different types of services, most of the time, it comes back down to the patient data," Matthews said in an interview with EHRIntelligence.

Consistency, completeness, and accuracy are key characteristics of health data quality, according to AHIMA. However, achieving quality patient data is easier said than done.

“It takes diligence to work on maintaining good, quality, accurate data,” Matthews said.

A 2019 study shows that approximately 18 percent of patient EHRs are duplicates. As a result, roughly one in five patients have incomplete health records. Duplicate patient records introduce patient safety issues, as clinicians cannot access complete patient health histories.

Additionally, duplicate EHRs lead to higher healthcare costs. Duplicate records can prompt clinicians to conduct unnecessary tests and procedures, constituting low-value care. On average, duplicate patient EHRs cost hospitals $1,950 per patient per inpatient stay and over $800 per emergency department (ED) visit, according to a Black Book survey.

etHIN works with its diverse population of data contributors, which range from small physician practices to larger healthcare organizations and local public health offices, on an ongoing basis to improve data feeds into the enterprise master patient index (EMPI).

"An EMPI is one of the main building blocks to a robust HIE," Matthews noted. "In my mind, the EMPI is just something that is an ongoing project, so of course, etHIN maintains and monitors its EMPI on an ongoing basis."

For example, etHIN's data enrichment work includes helping participants fulfill United States Core Data for Interoperability (USCDI) data requirements.

Matthews said that the HIE teams with participants to identify an action plan to achieve the desired results.

"We meet on a regular basis with each individual participant team until the effort is completed," she said. "Sometimes, a participant may include their vendor partners in meetings. While this type of work takes time, the result is well worth the effort for both the participant and etHIN."

With a background in industrial engineering, Matthews noted that it is in her nature to look at how etHIN can continue to grow and improve the data quality of its EMPI of over two million patients.

Recently, the HIE tapped an artificial intelligence (AI)-based platform from health IT vendor 4medica to help automate its data quality efforts.

The tool's master patient index (MPI) platform leverages AI and machine learning (ML) technology to reduce patient duplication rates to 1 percent.

Matthews noted that bringing down patient duplication rates is a key aspect of etHIN's role as a data steward for its participants.

"At the end of the day, it's not only a steward to say we provide value with the quality of data; it's also showing the numbers to our participants," Matthews said.

She noted that the vendor is currently analyzing the etHIN EMPI to help the HIE understand how it can continue improving data quality.

The analysis will also break down the EMPI to the source level, allowing insights into how each of etHIN's participants can enhance patient data quality.

"Once we have the analysis, we can actually go to participants and work together to see how we can improve it," Matthews said.

For instance, the analysis may show that certain participants need to spend more time at the point of patient registration and ensure that the initial data they collect and put into the system is as accurate as possible, she noted.

The vendor will also assist individual etHIN participants with data quality improvement efforts.

"It was a natural fit," Matthews said of the AI vendor partnership. "We are already working with our participants on their data quality, and this is just one more additional value add that we can help provide."

At the end of the day, cleaner data across the etHIN EMPI will benefit the patient, she emphasized.

"Fundamentally, the EMPI work should never end because that's not how you achieve data quality," she said. "As an organization, we have a valuable asset. That asset is patient data, which impacts patient care and the decisions made, as well as analytics. We have to be the data stewards. That is our responsibility."

Dig Deeper on Clinical documentation

Cloud Computing
Mobile Computing