Revolutionary new HCC Risk Adjustment tool delivers 99% accuracy when extracting HCC codes from physician notes using NLP

MARLBOROUGH, Mass., Nov. 15, 2021 /PRNewswire/ — Persivia, a provider of population health, risk management and hybrid care management solutions, today announced the release of the latest version of its AI based HCC Risk Adjustment application. A part of Persivia’s CareSpace® platform, the HCC module can be deployed as a stand-alone service or as part of the full platform.

Using AI based Natural Language Processing (NLP) techniques the module is able to extract the appropriate HCC codes from physician SOAP notes. The module processes both structured and unstructured data to extract the required codes. Persivia’s HCC coding module has achieved very high levels of accuracy:

• Sensitivity: ~ 90% (Sensitivity is the probability that a diagnosis identified as being an HCC code is indeed an HCC code)
• Specificity: ~ 95% (Specificity is the probability that a diagnosis identified as NOT being an HCC code is indeed NOT an HCC code)
• Accuracy: >90% (The percentage of diagnosis codes correctly identified as being or not being HCC codes out of the total number of HCC codes in the note.)

“This phenomenal accuracy allows us to find as many as 100% more coding improvement opportunities in a typical physician note when compared to solutions provided by the more established vendors including the large EHRs. For example, while a leading EHR, such as Epic, might identify an average of 2 HCC codes per physician note, CareSpace will find an average of 4 HCC codes in those same notes.” Said Dr. Mansoor Khan, CEO of Persivia, “To get around this inability to code as many as 50% of diagnoses organizations might use manual coding to find the opportunities missed by their current solutions. However, that introduces additional sources of errors as well as adding significant cost to the process. Typically, manual coding can take an average of 15 minutes per chart to code for HCC diagnoses. Our system can process 900 charts in those same 15 minutes. From a client’s point of view this increase in productivity results in a significantly lower cost.”.

The high accuracy and productivity of the Persivia tool also generates a very high ROI. For example, for an approximately 40,000 patient population the Persivia HCC coding tool generated a RAF improvement of about 10,000 points.

CareSpace’s single platform architecture allows any coding improvement opportunities identified by the system to be delivered directly to the provider’s desktop in real time. Combining EHR, claims, ADT and SDOH data CareSpace provides a clinical and analytics infrastructure that is unrivaled in the industry and provides significant, demonstrated ability to

• Improve operational efficiency
• Boost revenue generation and
• Enhance organizational resilience.

About Persivia Inc.
Persivia enables hospitals, practices, payers, public health agencies, and EHRs to manage multiple value-based care models across all available datasets to arrive at insights in real time, personalized at the point of care. Our first-of-its-kind single platform solution, CareSpace®, follows a patient from admission through post-acute stays and into the home. Powered by our Soliton® AI engine, CareSpace integrates disjointed legacy systems to help align incentives across multiple reimbursement models. For health systems and providers, we help improve care delivery and quality scores through AI driven workflows and pathways to create personalized care programs, optimize risk adjustment and member prioritization at the point of care, both in person and virtually. Uniquely, for the public health sector, Persivia’s SmartLab module helps state departments of public health manage the high-volume electronic test ordering and resulting requirements of large-scale epidemics like the COVID-19 novel coronavirus. Learn more at http://www.persivia.com.

View original content:https://www.prnewswire.com/news-releases/revolutionary-new-hcc-risk-adjustment-tool-delivers-99-accuracy-when-extracting-hcc-codes-from-physician-notes-using-nlp-301423267.html

SOURCE Persivia