Industry Insights

The Future of Prior Authorization: Predictions for 2027 and Beyond

Luma Team
Luma Team
|
Cover Image for The Future of Prior Authorization: Predictions for 2027 and Beyond

Prior authorization has been "about to change" for the better part of a decade. Federal rules get finalized, compliance deadlines get set, and then the industry finds ways to move slowly. But 2026 feels genuinely different. The combination of CMS enforcement muscle, AI maturity on both sides of the transaction, and real legislative momentum at the state level has created pressure that's hard to stall.

Here's an honest read of where PA is heading — what's going to land on schedule, what's going to slip, and what providers should actually be doing right now.


The CMS Interoperability Mandate Is About to Have Real Teeth

The CMS Interoperability and Prior Authorization Final Rule set January 2026 as the compliance deadline for impacted payers to implement real-time PA APIs. The rule requires decisions on urgent requests within 72 hours and standard requests within 7 days — with specific reason codes on every denial.

Most large commercial payers have made some progress on API infrastructure. The problem is smaller regional payers and MA plans that are running 12–18 months behind on technical implementation. CMS hasn't shown an appetite for enforcement leniency. Expect audits to start surfacing in Q3 2026, with meaningful penalties following in 2027 for plans that still can't demonstrate compliant API connectivity.

For providers, this matters because the interoperability rule isn't just about speed — it's about data. When payers are required to share PA requirements via FHIR APIs, providers can pull those criteria directly into their workflows rather than hunting through coverage policy PDFs. That's the unlock that makes intelligent documentation automation genuinely feasible at scale.


AI Is Arriving on Both Sides of the PA Transaction — and That's Complicated

Payers have been deploying AI-powered review systems for several years. High-profile reporting in 2023 on UnitedHealth's AI denial algorithms sparked congressional attention, but the underlying technology kept expanding. Automated review systems that flag documentation gaps in milliseconds are now standard at most large payers — not the exception.

The result has been faster denials, not fewer of them. An algorithm that processes 10,000 requests per hour doesn't pause to apply clinical judgment. It pattern-matches against criteria fields. If the documentation doesn't hit the right data points in the right format, the system flags it, and a denial goes out before a human ever looks at the request.

This is the dynamic that makes provider-side AI not a nice-to-have but a structural necessity. When the other side of the transaction is running automated review, submitting handcrafted PA documentation is a category mismatch. Tools that generate structured, criteria-mapped documentation — the kind of approach the Luma blog has been covering — exist specifically to compete on equal footing with payer automation.

By 2027, the practices that hold first-pass approval rates above 80% will almost certainly be using some form of AI documentation tooling. The ones still doing it manually will be fighting an asymmetric battle.


Gold Carding Will Expand Faster Than the Industry Expects

Texas passed HB 3459 in 2021, establishing the first formal gold carding framework in the country: providers with PA approval rates above 90% over a 6-month period can be exempted from prior authorization requirements for specific services. The implementation took time, but the model has now been evaluated by a dozen other state legislatures.

Arkansas, Virginia, and Tennessee have all passed or advanced similar legislation in the last 18 months. The National Conference of State Legislatures tracks an accelerating pace of PA reform bills — gold carding provisions are appearing in roughly 30% of them.

The prediction: by end of 2027, at least 15 states will have active gold carding statutes on the books. That's not a fringe outcome — it's the logical extension of current legislative momentum. For specialty practices in dermatology, rheumatology, and gastroenterology, where biologic PA volume is highest, this represents a genuine workload reduction if you qualify.

The catch: qualification requires demonstrating sustained high approval rates, which requires strong documentation quality now. Practices that start investing in that infrastructure today will be positioned for exemption eligibility. Ones that wait will be chasing a threshold they haven't built toward.


What Will Take Longer Than Everyone Hopes

Two things are going to slip past their expected timelines. Both are predictable.

Full EHR integration with PA workflows is still years away from being seamless. The interoperability mandates push payers to expose data via FHIR APIs, but pulling that data into EHR PA workflows in a way that actually reduces clicks requires EHR vendors to build to those APIs — and EHR development cycles are notoriously slow. Epic and Cerner have made public commitments here, but most practices won't experience frictionless EHR-to-payer PA connectivity until 2028 at the earliest.

Payer AI transparency is the bigger long-term issue. Right now, providers receive denial letters that cite criteria but don't explain what specific documentation gap triggered the algorithmic flag. There's bipartisan interest in requiring payers to disclose when AI is used in PA decisions — the HEALTH AI Act touches this directly — but federal regulatory movement here will be slow. Meaningful transparency rules are a 2028–2029 story, not a 2027 one.


What Providers Should Be Doing Right Now

The gap between where PA is today and where it lands in 2027 is wide enough to matter. Here's where the leverage is.

Audit your first-pass approval rates by payer and drug class. Most practices don't have clean visibility into this. If you don't know your baseline, you can't track improvement — and you can't make the case for gold carding eligibility.

Map your documentation against current payer LCD/NCD criteria. Payer coverage criteria for biologics update 2–4 times per year. What satisfied Humana's criteria for ustekinumab in 2024 may not satisfy the 2026 version. Manual tracking of these changes is unrealistic at volume; the practices winning here have built or adopted tools that track criteria updates automatically.

Treat the first submission as the only submission. Every element the payer's criteria requires — disease activity scores, prior treatment documentation, contraindication evidence — needs to be addressed upfront. Appeals work eventually, but they cost 4–6 additional weeks of patient wait time and 30–60 minutes of staff time per case.

The prior authorization system is going to keep getting more complex before it gets simpler. The CMS mandates help at the infrastructure layer, but payer criteria will keep expanding as new biologics and specialty drugs enter the market. The practices that invest in documentation infrastructure now are building a competitive moat that compounds over time.


Sources:
Centers for Medicare & Medicaid Services. (2024). CMS Interoperability and Prior Authorization Final Rule (CMS-0057-F). cms.gov
Siddiqui, M. & Doble, J. (2023). How UnitedHealth's Algorithm Is Driving Patients to Their Deaths. The Guardian. theguardian.com
Texas Legislature. (2021). House Bill 3459 — Gold Carding Provisions. capitol.texas.gov
National Conference of State Legislatures. (2025). Prior Authorization State Laws. ncsl.org
U.S. Congress. (2024). HEALTH AI Act, H.R. 7136, 118th Congress. congress.gov

Want to learn more about Luma?