AI partnership aims to improve unstructured data utilization

Unstructured data, such as notes from physicians, radiologists and nurses, can affect the accuracy of quality scores and, ultimately, reimbursement.
Jeff Rowe

Healthcare collaboration software vendor Apervita recently announced the addition of AI-powered data abstraction technology from Carta Healthcare to its platform’s interoperability layer, which enables providers and payers to more effectively utilize data generated in the course of patient care.

The companies say the new feature will enhance Apervita’s quality measurement and clinical intelligence solutions, and that the combination will greatly reduce the time it takes to manually gather data for quality registries, reportable quality measurements and customer quality measures. These cloud-based tools can also help make digital quality measurement more accurate, they added.

“Abstracting clinical content from unstructured fields in the EHR allows Apervita greater access to data to accurately measure and improve quality,” said Rick Howard, Apervita Chief Product Officer, in a statemen.

In an email to HealthcareITNews he added, "Provider organizations with great quality scores often have armies of chart abstractors. However, smaller and medium-sized organizations often only have the resources to do the bare minimum of chart abstraction. Carta Healthcare’s AI-assisted technology paired with Apervita’s platform will help all organizations, regardless of size, to be able to have complete patient information.” 

According to research, the companies say, 80% of data in electronic health records (EHRs) is unstructured, and often not utilized. This unstructured data, such as notes from physicians, radiologists and nurses, can affect the accuracy of quality scores and, ultimately, reimbursement. As a result, many organizations invest millions of dollars into manual chart abstraction.

To address that, Apervita's new tool will use a combination of technology, people and processes to analyze patient records for quality reporting and clinical insights.  The technology uses AI-enabled natural language processing to pull data from medical records, which are then validated by nurse data abstraction professionals – with a goal toward reducing the amount of time needed for abstraction.   

For conditions like sepsis, for example, automated analysis of words in clinical notes can detect indicators such as “yellow wound” and its variations.

The core problem, say the companies, is that data abstraction is a manual process that robs physicians and nurses of time they could otherwise spend with patients. Indeed, they note, studies estimate physicians spend an average of nearly three hours per week entering information related to quality measures, while the average staff member spends over 10 hours. Researchers projected that quality measurement documentation across only four specialties costs U.S. healthcare $15.4 billion annually. Data abstraction from charts represents a significant share of that total.

“Carta Healthcare and Apervita are aligned to empower health systems to use their data and enable better patient care,” said Matt Hollingsworth, Carta Healthcare CEO.

Photo by mikkelwilliam/Getty Images