Significant Change Coming in Prior Auth and Inpatient Determinations

Jon: I am an actuary by background, and I was at Milliman for 25 years, running a consulting practice in Denver and sponsoring other new developments. Milliman Care Guidelines was launched, and I became Chairman and then CEO. In 2012, we moved to Hearst and became MCG Health. Over the years, Milliman Care Guidelines/MCG transformed from a content company to a product company and now to a medical technology company.

Jon: MCG has long been the source for evidence-based clinical guidelines to give providers and payers confidence in their decisions. We are currently working with multiple organizations to automate these decisions using information in the EMR and codification/AI/NLP and soon LLMs. This is both in the space of medical inpatient admissions and prior authorizations.

As an organization, we focus on further automating the process to take the overhead out of these key decisions, on measuring the value that these changes produce, and on using and collecting data that allows us to further automate the decision-making processes.

Jon: The historic processes for both Prior Auth and Inpatient determinations have been very manual. The Inpatient side has been after the fact. The trends I see are :

1.  Significant pressure on payers to simplify the processes and speed up response times. This takes the form of PR and federal/state legislation.

2. There is more legislation for gold carding, which is a blunt instrument – it is hard to get a heterogenous group of services/providers to run the right analysis on.

3. Payers are working hard to satisfy the demands, both in current constructs (fax/portal) and in future interoperable approaches. It is hard for payers to trust automated processes, so they are working on what gives them more confidence.

Jon: Yes, and AI approaches need to be integrated with evidence-based decision making (as opposed to historic pattern recognition). Our data science group feels that future generative AI (large language models) will commoditize current AI, because AI will be much more accessible. AI will be very good at reading EMRs and applying guidelines.

Jon: This rule has not yet been finalized, and our sources suggest CMS will finalize it in December. We expect the implementation date to be pushed as a result.

Nonetheless, health plans are starting to work now to make that date, as there are many moving pieces. This includes finalizing and codifying policies and guidelines and working with providers on interoperability. The biggest step may be working together with providers in what has been a challenging relationship.

Jon: If you map out a project plan, complying with the CMS rule, even by January 2027, requires many steps. It is important to start working now.

Jon: With all of the overhead, it is easy to forget that both payers and providers are trying to do what is best for their members/patients. Making evidence-based decisions allows all of us to align. The administrative improvements should help fix the biggest problems in documentation, so that this becomes less of a battlefield.