Technology

Health Fidelity helps you see the whole picture.

In order to perfect the risk adjustment cycle, it is necessary to uncover risk factors hidden within clinical charts that cannot be easily searched or analyzed.

Efforts to manually extract meaningful insights from clinical charts are costly, labor-intensive, time-consuming, and error-prone – unless you have the power of the industry’s most advanced natural language processing (NLP) engine supporting you.

Our technology platform offers health plans an automated method to utilize 100% of members’ data regardless of format to accurately analyze and identify risk across multiple populations.

Technology Platform

Discern Inference

  • Machine Learning
  • Association Mining
  • Coding Guidelines

DISCERN takes structured output from REVEAL and uses proprietary algorithms to provide highly relevant information to Health Fidelity applications. The algorithms improve continuously using machine learning and association mining, and as a result provide insightful and accurate coding suggestions.

id_revealnlp

  • All clinical document types
  • High precision and recall
  • Semantic parser
  • Disambiguation

HF REVEAL is the most advanced healthcare NLP engine in the industry; it drives Health Fidelity’s risk adjustment applications and allows clients to use unstructured data that was previously impossible to use. Developed in an exclusive R&D collaboration with Columbia University’s technology ventures group, HF REVEAL offers the highest accuracy over the broadest range of use cases compared to any other NLP engine in the market. HF REVEAL’s precision and recall metrics are in the 90% – 95% range.

id_terminologyengine

  • Industry standard terminologies
  • Robust taxonomy

Health Fidelity’s robust terminology library consists of more than 5 million word forms and relationships that have been refined by 20+ years of research at Columbia University. Our system can support standard clinical taxonomies such as ICD-9, ICD-10, SNOMEd, RxNorm, LOINC, etc.

id_data_acquisition

  • HL7/CCDA
  • OCR
  • Virtual remote print

Our data acquisition engine is an extremely versatile integration suite that can handle highly complex and disparate data sources. It supports PDF/image/scan data sources through its advanced OCR (optical character recognition) capabilities. It also supports HL7/CCDA interfaces for direct data integration from EHR systems, as well as direct medical chart ingestion via secure file transfer. Additionally, our data acquisition engine consumes all administrative data required for risk adjustment.

90%+

Accuracy

100+

Peer-reviewed Publications

150M+

Clinical Documents Processed

columbia_technology_ventures