Using natural language processing to manage healthcare records

Опубликовано 25 июня, 2019

Principal Healthcare Program Manager, Industry Experiences

The next time you see your physician, consider the times you fill in a paper form. It may seem trivial, but the information could be crucial to making a better diagnosis. Now consider the other forms of healthcare data that permeate your life—and that of your doctor, nurses, and the clinicians working to keep patients thriving. Forms and diagnostic reports are just two examples. The volume of such information is staggering, yet fully utilizing this data is key to reducing healthcare costs, improving patient outcomes, and other healthcare priorities. Now, imagine if artificial intelligence (AI) can be used to help the situation.

The Azure platform offers a wealth of services for partners to enhance, extend, and build industry solutions. Here we describe how SyTrue, a Microsoft partner focusing on healthcare uses Azure to empower healthcare organizations to improve efficiency, reduce costs, and improve patient outcomes.

Billions of records

Valuable insights remain locked in unstructured medical records such as scanned documents in PDF format that, while human-readable, present a major obstacle to the automation and analytics required. Over four billion medical notes are created every year. The clinical and financial insights embodied within these records are needed by an average of 20+ roles and processes downstream of the record generation. Currently, healthcare providers and payors require an army of professionals to read, understand, and extract healthcare data from the flood of clinical documents generated every day. But success has been elusive.

It's not for lack of trying. In the last decade, an effort was made to accumulate and upload data into electronic health records (EHR) systems. Meaningful Use is a government-led incentive program that aims to accelerate the movement from hard-copy filing systems to electronic health records. Still, the problem is related to the volume and the lack of time and resources to assimilate masses of data.

Note: the Meaningful Use program has a number of goals. An important one is, “Ensure adequate privacy and security protection for personal health information.” Data security is a prime value for Azure services. Data services such as Azure SQL Database encrypt data at rest and in-transit.

Moving the needle on healthcare

As costly and extensive as this effort was, many believe that we have yet to see evidence of any significant impact from the digitization of healthcare data to the quality or cost of care. One way to radically improve this is using AI for natural language processing (NLP)—specifically to automate reading of the documents. That enables subsequent analytics, yielding the most relevant actionable information in near real-time from mountains of documents to the medical professional. It empowers them to deliver better quality care, more efficiently, at lower cost.

In action

A Microsoft partner, SyTrue is leading the way. In the words of their Founder and CEO, Kyle Silvestro, “At SyTrue, the next big challenge is accessing this vast pool of accumulated patient data in a serviceable way. We’ve created a platform that transforms healthcare documentation into actionable information. The focus is on three main features: speed, context, and adaptability. Our technology consumes thousand-paged medical records in sub-seconds. The innovation is built on informational models that can ingest data from multiple types of clinical and financial health care organizations. This allows diverse healthcare stakeholders to use the system. The main objective for the technology is to present key clinical and financial insights to healthcare stakeholders in order to reduce waste and improve clinical outcomes.”

Informed by natural language processing and machine learning

SyTrue relies on NLP and machine learning (ML) as the underlying technology. Using their own proprietary methods, they perform “context-driven information extraction.” In other words, they connect the dots. The graphic below shows their processes.

Diagram displaying context-driven information extraction

Improving healthcare

SyTrue’s offers the NLP OS (Operating System) for healthcare. It aids in several ways.

  • It unlocks healthcare records and enables healthcare professionals to interact with medical record data and its clinical and financial implications. Specifically, it eliminates the need for professionals to hunt for the same key observations. This enables professionals to spend more time focused on patient care.
  • NLP OS also bridges the communication between a specialist provider and a primary care physician regarding the care of a shared patient. The system extracts and highlights continuity of care recommendations generated within the patient’s care team.
  • A large healthcare organization installed SyAudit, powered by SyTrue NLP OS, at the front of their medical chart review process. Before the charts reach a nurse-reviewer, they are processed through this solution. The system interprets the documentation to determine if a nurse review is in fact needed, or if the documentation lacks actionable information. This potentially decreases the time spent by nurse reviewers.
  • A healthcare provider used SyReview, another SyTrue solution powered by the SyTrue NLP OS, for their quality capturing and reporting process. The particular process is related to an incentive program which directly ties quality to Medicare payment. Automating the quality-capturing process strengthens the feedback loop to providers that needed to show improvement. The organization also eliminated its manual quality-capture process, which was slow, expensive, and often inaccurate.

Next steps

To see more about Azure in the healthcare industry see Azure for health.

To find out more about this solution, go to the Azure Marketplace listing for NLP OS™ for Healthcare and click Contact me.