NLP Suspecting

Not all NLP (Natural Language Processing) engines are created equally; excellent risk adjustment performance requires a purpose-built NLP.

Healthcare data is waiting. Almost 80% of a patient’s information is locked inside unstructured data. But it doesn’t have to be that way. Edifecs’ NLP engine creates a deeper, more holistic view into healthcare data around patient populations. Getting these critical insights into a community’s health status offers numerous benefits, including the keys to improving a patient’s health. Our suspect analytics allow organizations to capture the full picture of patient risk from all available documentation, with convenient delivery alternatives and customizable settings to ensure the quality, control, and probability required to engage clinicians.

Infographic - NLP Engine


Best-in-Class NLP Engine

  • Don’t settle for off the shelf products. Edifecs NLP has processed over 3.1 billion pages of clinical documentation
  • Experience the impact of transitioning from working off 20-30% of available data to >90%
  • See clinically focused, machine learning-driven development with expert oversight at work. Edifecs NLP technology is on a trajectory of continuous, industry leading performance
  • Experience what Edifecs NLP can do for you. It understands the complexities of clinical language -- including grammar, syntax, context, and intent

Convert unstructured data - like physician notes entered in an EHR - into a structured extraction of clinically meaningful intelligence that can be systematically processed by machine to derive predictive insights.

Customizable Output

  • All outputs are checked against available claims data and linked to available evidence. In prospective use-cases, if a suspect already appears on a claim, it is considered confirmed and not delivered to the user
  • Each condition or code surfaced has an associated confidence score based on the confluence of available evidence and the rate at which similar outputs are accepted or rejected, both globally and locally

Fine-tune the NLP suspecting output to align with your organization’s data input, user behaviors, best practices, and customized needs.

Configurable Delivery

  • Suspected conditions can be delivered directly into Edifecs’ SaaS-based software applications during Pre-Visit Planning or at the Point-of-Care
  • Alternatively, suspects can be routed through flat-file output, form-based output (CCV), or through EPIC BPA
  • Regardless of delivery, the associated evidence and rationale for the suspect is presented for clinical validation

Suspected conditions can be delivered directly into Edifecs’ SaaS-based software applications