The healthcare industry faces countless challenges, from staff shortages and preventing burnout to complying with government mandates. Another challenge that is sometimes overlooked or pushed to the side is unstructured data. In this article, we’ll take a look at what unstructured data is and the challenges it poses, the benefits of structured data, and how to transform unstructured data into structured data to revolutionize your ambulatory order processes.

Unstructured Data in a Nutshell

Unstructured data is information that is difficult to store and manage due to it not following conventional data models. Many analysts estimate 80-90% of data is unstructured, including text, images, audio files, and emails. One of the biggest challenges with unstructured data is that it’s not searchable. This leads to time-consuming manual processes such as reading text or listening to audio to extract data. In healthcare, this is both costly and inefficient.

A major source of unstructured data in healthcare is faxed medical records. Unstructured data in medical faxes can lead to:

  • Errors resulting in denied claims, care delays, patient frustration, and in extreme cases, adverse health outcomes
  • Resource strain from the labor costs and time it takes staff to read documents and/or manually enter data into the EHR
  • Clinical burnout

While unstructured data is a challenge in healthcare, it also provides an opportunity for healthcare providers to improve the care experience for both staff and patients…

Benefits of Structured Data

Unlike unstructured data, structured data is stored in a rigid format, allowing for consistency and searchability. An example of structured data is patient information stored in columns in a spreadsheet. Turning unstructured data into structured data has a host of benefits for providers, physicians, and patients, including:

  • Reduced labor costs for providers
  • More efficient workflows
  • Reduced claim denials and care delays
  • Reduced burnout and turnover among staff
  • Improved patient safety and satisfaction

All of these benefits can positively impact providers’ reputation and bottom line. For example, converting the unstructured data from a faxed patient order into structured data in the EHR allows clinical staff to quickly find the order and all pertinent patient information in the system. This structured data can also be sent to different systems, such as payers, for quick processing. Since the information is stored electronically in a structured format, providers save time and money by not having staff dedicated to tracking down paper documents and poring through them for the information they’re looking for. Some order management platforms can save providers even more time and money by automatically generating electronic orders and/or converting faxes into electronic files.

Now that we know the difference between unstructured and structured data, let’s look at how unstructured data is transformed into structured data.

Transforming Unstructured Data into Structured Data

As previously mentioned, one way of transforming unstructured data into structured data is manually entering unstructured data into a structured format, such as a spreadsheet or EHR. For example, clinical staff may manually type in information from a faxed patient order into the EHR. This process is very time-consuming though.

A much more efficient method involves machine learning and AI, specifically, Natural Language Processing (NLP). During HIMSS22’s Machine Learning & AI for Healthcare Forum in March, chief product officer at the National Committee for Quality Assurance, Brad Ryan said:

“I think there’s a huge opportunity to leverage natural language processing on the fly as we’re starting to collect this new data…For me, one of the most interesting applications of machine learning and artificial intelligence is moving data from unstructured to structured.”

What exactly is natural language processing, or NLP? IBM defines NLP as “the branch of computer science concerned with giving computers the ability to understand text and spoken words in much the same way human beings can.” Some of the tasks comprising NLP include:

  • Speech recognition, such as speech-to-text
  • Part of speech tagging/grammatical tagging
  • Word sense disambiguation, and
  • Named entity recognition – the ability to differentiate a patient’s name from a doctor’s name

Order Management and Structured Data

A recent study conducted by Stanford University School of Medicine highlighted the benefits of using NLP/AI to convert faxes (unstructured data) to PDF’s (structured data). For the purposes of the study, the Stanford researchers designed several AI algorithms to do the following with PDF’s converted from faxes:

  • Detect handwriting and translate it into words and digits
  • Detect and segregate COVID reporting-related pages
  • Prioritize faxes containing references to emergencies
  • Notify health officials of the most serious cases

The study found the algorithm was able to correctly identify 83% of files containing serious COVID instances, saving public health officials time manually looking through faxes. This is just one real-world example of how structured data can be a game changer in healthcare. For providers, a key area that can be improved by transforming unstructured data into structured data is ambulatory order management. Implementing solutions that utilize NLP to convert ambulatory faxes into interoperable electronic orders can:

  • Improve efficiency by reducing workloads and saving time
  • Improve order transparency by making orders easily searchable
  • Improve order accuracy and patient safety through the use of validation questions
  • Ensure compliance with CMS’s Appropriate Use Criteria (AUC) mandate requiring proof of clinical decision support consultation with Medicare claim submissions by allowing for integration with a clinical decision support mechanism (qCDSM)
  • Improve the prior authorization process by making order information interoperable, allowing it to be shared with third parties such as payers or prior authorization specialists
Benefits of transforming unstructured fax data into structured data ambulatory orders

How We Can Help with Unstructured Data

For over 24 years, we’ve been helping providers improve the care they provide patients. We offer several innovative solutions for transforming unstructured data. First, we are a licensed distributor of 3M’s speech recognition solution M*Modal, winner of the 2022 Best in KLAS award for “Speech Recognition: Front-End Imaging”. M*Modal preserves radiologists’ natural workflow and turns their dictation into accurate, electronic documents that are structured, clinically encoded, searchable and shareable.

Our award-winning ambulatory order management solution, iOrder, can revolutionize your ambulatory order process by providing:

  • Electronic order generation and fax-to-PDF conversion
  • Complete order transparency and validation to ensure order accuracy, patient safety, and minimal care delays
  • Integrated clinical decision support to ensure AUC mandate compliance
  • Patient engagement tools including text reminders, links to procedure-specific instruction videos, and facility navigation assistance
  • Prior authorization assistance to relieve staff of the burden and frustration of the prior authorization process

iOrder has built-in AI/NLP integrations that convert faxes to electronic files, transforming unstructured data into actionable structured data. Watch the video below for more information on iOrder. If you’d like to see iOrder in action, schedule a free live demo today.