The findings below were part of a 2022 McKinsey & Company study on the effects of applying artificial intelligence to the prior authorization (PA) process. The research revealed that leveraging AI will strategically benefit the entire healthcare ecosystem for providers, payers, and members.
“Our analysis suggests that artificial-intelligence-enabled prior authorizations can automate 50 to 75 percent of manual tasks, boosting efficiency, reducing costs, and freeing clinicians at both payers and providers to focus on complex cases and actual care delivery and coordination.”
To understand why McKinsey’s researchers were able to make such a bold claim, let’s review how most health payers receive prior authorization requests today. A 2021 AMA survey found that nearly half (47%) of providers always or nearly always submit PAs for medical services by paper-based fax machines.
That means a big chunk of the PA requests your staff has to process are received as unstructured data on fax machines. Digital fax pages are technically “images,” as well, not data. And that means your staff needs to manually review these faxes, prioritize the cases, and often re-enter the information into a database by hand.
Prior Authorizations and the Inefficiencies of Unstructured Data
A 2020 study by the CAQH found that manually reviewing and responding to PA requests — communicating with providers by phone, mail, or paper fax — costs payers an average of $3.14 per transaction. The researchers also estimated payers could reduce this average cost to just pennies per transaction — while also ensuring timely delivery of care — by automating and digitizing at least part of the process.
Why? Because the burdens of reviewing and processing unstructured data — in PAs submitted as faxes, paper or cloud — create enormous challenges for payer organizations. Just a few examples:
- Payers can’t automatically sort and prioritize traditional faxes according to member need or urgency.
- Manually reviewing and acting on fax-based PAs takes staff longer and results in slower responses to all PA requests.
- Manual data re-entry of unstructured fax content increases the chance of human mistakes — which can lead to substandard health outcomes and an increase in the total costs of member care.
Here’s another way to think about it. Let’s say you received 100 names in an Excel file, and you needed to rearrange them into alphabetical order, that task would require only a few clicks. The software can easily distinguish the letters in the names and the built-in intelligence to order them from A to Z.
Conversely, let’s say, you received those 100 names on paper. Resorting them alphabetically won’t be as easy as clicking a few buttons. Someone is going to need to physically rewrite every name on the list — or manually type them into a database.
That’s a very high-level analogy, but it underscores the difference between an AI-enabled PA process and the name-on-paper process that your organization is likely challenged with every day.
How to overcome these obstacles for both your members and your bottom line
Remove Manual Processes with Digital Cloud Faxing
Even if providers continue to submit PAs using legacy fax systems, your team doesn’t need to receive the requests as hardcopy documents.
The right digital cloud fax solution — one customized to fit your unique needs and constraints of the healthcare industry — can ingest inbound fax content into EHRs or other records systems. It can also leverage basic optical character recognition (OCR) to identify key data in your faxes, such as member names, ID numbers, names of physicians, etc.
This straightforward addition to your organization’s digital environment can remove several time-consuming, manual steps from the review process — to help your staff streamline their PA responses and ultimately save your organization money.
Sort and Organize Key Information Using Natural Language Processing (NLP)
When it comes to automating the management of PAs, NLP takes digital cloud faxing a step forward – otherwise, any inbound fax content would be manually reviewed and processed by your staff.
NLP embedded in your cloud fax platform will be able to automatically sort and organize key information in an inbound fax, and even take a next step if necessary. For example, you can instruct your NLP solution to scan every inbound fax for key terms — “urgent,” for example — and prioritize those requests by sending alerts to the right people in your organization.
And because NLP employs machine learning, it will continually grow more adept at identifying key terms and concepts as it consumes more data. It can even learn to read handwritten notes on fax pages.
Employ Artificial intelligence for Increased Efficiency
Your organization will see the greatest operational benefits when you employ artificial intelligence to your PA processes. Artificial intelligence brings together all of the capabilities listed above while adding layers of learning, thinking, and continuous improvement to your fax solution.
Returning to the McKinsey study cited in the introduction, the research team concluded that “an AI-enabled workflow requires fewer steps,” automating most of the administrative steps most payers handle manually today. AI can also benefit payers by sorting requests according to high and low priority as well as those that can be handled by RNs and those that need doctor input.
Bottom line: Because fax (both paper and digital) remains one of the most common methods for providers to submit prior authorization requests, payers need to employ new tools and technologies to make their ability to receive, review, and process these requests as efficiently and reliably as possible.
The best, easiest way to do this today is to implement the right digital cloud fax solution with built-in AI and NLP capabilities. Doing so will lessen the burden on payers’ staff, streamline their ability to respond to PAs, reduce the overall costs to the company, and improve both member and provider outcomes.