Are automated prior authorizations on the horizon?

©Halfpoint 18/stock.adobe.com

Prior authorizations are not going anywhere anytime soon, but drastic changes to how they are handled may finally be on the horizon.

An April 2018 study from the Government Accountability Office found that Medicare prior authorization programs saved Medicare between $1.1 and $1.9 billion since 2012 through a reduction in unnecessary utilization and improper payments. However, the American Medical Association (AMA) did some research of its own, estimating in a 2017 report that clinicians spent 14.6 per week on prior authorizations. 

Not only does this process take time away from patients, but 78 percent of clinicians reported in the AMA survey that the delay can result in patients abandoning a particular course of treatment. 

AMA, along with the American Hospital Association, American Pharmacists Association, America's Health Insurance Plans, Medical Group Management Association and more have petitioned for change to the prior authorization process, specifically in streamlining and standardizing the process.

Technology is paving the way to move the process along, said Robert M. Tennant, MA, director of health information technology policy and the Medical Group Management Association (MGMA). There is currently work underway to automate the prior authorization process, and these efforts are being headed up by the Council for Affordable Quality Healthcare (CAQH).

According to CAQH, 90 percent of communications between payers and providers is still done by phone or fax, which really adds up considering that there are around 77 million prior authorizations done manually each year. Automation of this process could cut the cost of performing these transactions, with CAQH estimating that standards could results in a nearly $7 per transaction savings between payers and providers.

Tennant said part of the process of automating these communications depends on when CMS will release a standard for the mandated 278 transaction. CMS has promised these standards for some time, he said, but they have yet to be released. When they are, Tennant said the prior authorization process could be automated between providers and payers with data pulled from electronic health records.

“The future of prior authorization is real-time-prior authorization should be a discussion, not a transaction,” Tennant said. “If we can automate that discussion, it can really save time for physicians and improve patient care.”

One way technology may help speed along the prior authorization process is through the Fast Healthcare Interoperability Resources (FHIR), a draft standard for electronic health information exchange created by Health Level Seven International. Tennant says HL7 is working with a number of major stakeholders in healthcare to launch the DaVinci Project.

“They are using this new standard to move clinical data, and some of the use cases are administrative. One is prior authorization,” Tennant said. “They are going to be looking at it in a very different way than what we know.”

The Centers for Medicare and Medicaid Services (CMS) has taken the DaVinci effort using FHIR and created its own Document Requirement Lookup Service Initiative (DRLS), he said.

“We are going to create a new workflow for prior authorization,” Tennant said.

The process would work directly with electronic health records. Currently, when a provider orders an intervention, there is a long back-and-forth process, Tennant said. The new approach would involved an automatic trigger in the electronic health record when a physician places an order that would send a transaction to CMS to find out if prior authorization is required. The system would offer a clinical documentation template sent by CMS outlining what is needed to justify the ordered intervention and prove medical necessity.

“So there’s no more guessing about what information is necessary,” Tennant said. “The goal is really to get all payers to agree to and support this standard. This will be a game-changer. It could be real-time, if not near real-time.”