The Time is Now for NLP/AI to Extract Intelligence from Unstructured Documents and Remove Roadblocks to Insightful Data

Launched at HIMSS 22, Clarity transforms unstructured documents into useful, analytics-ready data

I have been in healthcare technology for nearly 20 years and I have waited for us to solve the interoperability problem since that time. While working at an HIE company (we called them RHIOs), we attempted to solve the problem with semantic interoperability because back then we thought the problem was missing standards for nomenclature. We solved that with SNOWMED, LOINC and RxNorm and yet we still didn’t solve interoperability. We worked hard on Master Patient Indexing (MPI) or Record Locator Services and we still haven’t solved the problem. Why?

While the industry is pushing for interoperability frameworks, such as TEFCA, we are failing to recognize that the receiving end of interoperability may be in care settings that were not incentivized to adopt digital technology such as post-acute; skilled nursing facilities; home health; hospice, and assisted living. Many of whom still use paper fax machines. Did I really say the ‘F’ word?

At Consensus Cloud Solutions, we help providers move to the next level of digital maturity, digital faxing, without asking them to adopt FHIR standards. While one step up, (no paper, it’s digital ), cloud faxing solutions can do better to really advance interoperability and enhance workflow than paper faxing.

One of the most challenging aspects of interoperability is the need to transform digital unstructured patient documents and clinical content to structured, EHR consumable, analytics-ready data. Paper faxes and other physical documents are unstructured and can be incorporated into a patient record only as PDF or TIFF images. While providers can access and view this data, it is not searchable and cannot be incorporated with other types of individual electronic patient data or aggregate data used in population health and other care initiatives.

Moving From Documents to Intelligent Data

An optical character recognition (OCR) tool has limited value when it comes to intelligent data. While OCR can recognize and extract words and numbers, determining what to do with that data falls to the organization. In most cases, this means manually reentering information into the correct fields in another system, which blunts many of the advantages of such a system.

We have moved from OCR technology to Natural language processing (NLP) and Artificial Intelligence (AI), which represent the way forward for unstructured documents. More than one-third of hospitals were looking to adopt NLP solutions in 2021, according to an AI survey. However, there are important distinctions between the two.

With NLP, a fax solution will be smart enough not only to recognize character strings in the document (Smith, John, follow-up visit, shortness of breath) but also to know what they mean, such as John Smith is a person’s name. NLP knows where to put each detail and what action to take if the data suggests it (such as an urgent notification). In fact, NLP algorithms can recognize that a single fax transmission contains details on multiple patients, automatically segment the pages, and keep them with the correct patient record. Over time, an NLP-powered faxing solution will become capable of understanding an ever-growing range of variations of how different people present the same data.

Artificial intelligence takes technology one step further by bringing these capabilities together and adding layers of learning, thinking, and continuous improvement. An AI-powered fax solution can:

• Read, understand and extract free text in a digital fax document
• Automatically apply data (prescriptions, patient IDs, demographic information, etc.) to the right fields in a digital file
• Reduce manual data entry
• Increase timely access to data that can be used to improve care

A recent AI survey of 500 senior healthcare executives shows 96% of respondents believe AI plays an important role in their effort to reach health equity goals. In addition, 94% agreed they have a duty within the health care system to ensure AI is used responsibly.

True interoperability has the power to transform patient care at the individual level, by accelerating the time essential data is received and decreasing the time to treatment.   Consensus Clarity offers an important step towards interoperability by enabling clinicians to extract actionable information that helps them make informed, decisions faster — ensuring there is no disruption in the continuum of care.