GI practices complete 31 prior auths weekly per physician. Learn how AI automation cuts the 2-day weekly admin burden and reclaims staff capacity.

Why Gastroenterology Practices Are Drowning in Prior Authorizations—And How AI Can Change That

The Tuesday Morning Crisis

It's 8 AM on a Tuesday morning at a three-location gastroenterology practice in the Midwest. The clinical staff is already in the procedure rooms prepping for the day's schedule—EGDs, colonoscopies, and follow-ups for patients on biologics. But before any scope hits the patient, there's a different kind of bottleneck happening in the back office.

The authorization coordinator—a competent employee who's been with the practice for four years—is staring at 47 prior authorization (PA) requests that landed in the system overnight and over the weekend. Some are urgent. A few are denials on resubmissions. One is for a routine screening colonoscopy, which would have been a 15-minute appointment two years ago but now requires insurance company approval. She's been at work for fifteen minutes and her inbox already feels like a fire hose.

This scene plays out in gastroenterology practices across the country, and it's only gotten worse. What was once an occasional administrative task has become a daily operational crisis that consumes the equivalent of 2-3 full-time employees at many practices—yet we rarely talk about it in clinical terms.

The Numbers Behind the Burden

Let's start with what we actually know about the scope of this problem. According to the 2021 National Survey of Physician Practices and Healthcare Systems, the typical gastroenterology practice with a physician completes approximately 31 prior authorizations per week per physician. For a ten-provider group, that's roughly 310 prior auth requests every single week. At an estimated cost of approximately $11 per transaction in staff time alone (factoring in data entry, payer phone calls, insurance cards verification, and follow-up), you're looking at roughly $177,000 per year in direct labor costs just to handle the volume—and that's before you account for the denials that require rework, appeals, and the claims that go unpaid while stuck in authorization limbo.

The problem has accelerated dramatically over the past decade. In 2012, prior authorization requirements applied to roughly 10% of all procedures. By 2022, that number had climbed to 35%. That's a 250% increase in the authorization burden over ten years, yet most practices haven't fundamentally changed how they manage it.

According to CAQH Index data, the prior authorization process alone generates approximately $43.8 billion in annual industry-wide spending on administrative overhead. The American Medical Association has documented that roughly 65% of all PA requests take longer than one business day to resolve, creating treatment delays and operational cascades throughout the practice.

And if you think this is a problem affecting only major procedures—think again. UnitedHealthcare recently moved to require prior authorization for routine diagnostic procedures including esophagogastroduodenoscopies (EGDs) and screening colonoscopies at average-risk patients. For a gastroenterology practice, this isn't a niche issue affecting 5% of cases. This is core business.

Why Gastroenterology Is Uniquely Vulnerable

Gastroenterology occupies a peculiar position in the specialty landscape when it comes to prior authorization burden. Unlike a surgical specialty with a few dozen procedures, or a medical specialty with a narrow treatment algorithm, GI practices manage an enormous matrix of decisions.

There are preventive screening colonoscopies (which sometimes require auth despite their proven value), diagnostic EGDs for various indications, therapeutic endoscopies for bleeding and obstruction, and an expanding formulary of biologic medications for inflammatory bowel disease. Each category has different authorization thresholds depending on the payer. UnitedHealthcare might require auth for routine screening at one coverage level but not another. Aetna might allow three colonoscopies per year for a surveillance patient with a history of polyps, while Anthem requires individual authorization for each one.

Then there are the biologics: infliximab, vedolizumab, ustekinumab, the newer anti-IL-23 agents. These medications are expensive, clinically necessary, and absolutely require prior authorization before dispensing. A single GI practice managing IBD patients might be coordinating dozens of biologic authorizations each month, each with its own documentation requirements, step-therapy protocols, and payer quirks.

According to the American Gastroenterological Association, approximately 50% of all prior authorization requests in gastroenterology relate to prescription refills and continuation therapy—not new procedures or medications, but literally continuing treatment that's already been prescribed. This creates a perverse administrative tax on disease management: the sicker and more complex the patient, the more authorization requests the practice generates.

A growing practice or multi-specialty organization (MSO) with multiple locations faces a coordination nightmare. The practice in location one might have one authorization coordinator who knows the local payer landscape. Location two has a different coordinator with different knowledge and relationships. Location three is trying to integrate workflows after an acquisition, with different systems entirely. Inconsistent turnaround times, duplicate work, missed opportunities to batch requests with the same payer—the inefficiencies compound.

The Real Cost of the Status Quo

Let's make this concrete with a realistic scenario. Consider a seven-location GI-focused MSO serving a mix of commercial, Medicare, and Medicaid patients across three states. The organization manages around 250 active providers when you count employed physicians and mid-level practitioners. Based on industry benchmarks, they're processing roughly 6,500 prior authorization requests per month—that's approximately 78,000 per year.

At $11 per transaction, that's approximately $858,000 in annual staff cost just to touch these requests. But the real cost multiplier comes from downstream impacts. A 2023 study in the Journal of Medical Practice Management found that 28% of prior authorization requests result in either denial or delay of >3 business days. For the MSO above, that means roughly 1,800 cases per year where care is either delayed or requires a denial appeal.

Each denial creates additional work: reviewing the rejection reason, determining whether to resubmit with additional documentation, appealing to a medical director, and following up multiple times. Appeals consume approximately $45 in additional labor per instance, and roughly 60% of appeals succeed. But even successful appeals often come too late—the patient's appointment has been rescheduled, the clinical opportunity has shifted, and the practice's throughput for that day or week is already damaged.

For the MSO in this scenario, denials and appeals represent an additional $81,000 in annual costs. Add in the productivity loss from delays (patients waiting for approvals, procedure rooms sitting empty pending authorization), and you're approaching $1.2 million in annual impact from a single operational process.

Now consider the human cost. Authorization coordinators are among the most burnout-prone roles in healthcare. They're constantly on the phone with payers who transfer them three times before reaching the right department. They're working with incomplete patient information, navigating seven different payer portals, and facing the frustration of knowing that each request they're processing could have been automated two years ago.

Why the Problem Persists

The real question isn't why prior authorizations have become burdensome—that's clear. The question is why, in an era of healthcare transformation and digital innovation, most practices are still managing this administratively.

Part of it is legacy infrastructure. Many practices use electronic health records and billing systems that were designed for a different era—systems that don't easily integrate with modern solutions, that make it hard to detect when authorization is required, and that create friction in the workflow.

Part of it is payer fragmentation. Every insurance company has different requirements, different portals, different approval timeframes. Building a solution that works across this landscape is genuinely complex.

But much of it is a simple lack of visibility. For years, prior authorization has been treated as "just part of the billing process"—inevitable, like taxes, but not something you fundamentally reimagine. The AMA's ongoing advocacy against the PA burden (documented in their annual surveys) shows the industry recognizes the problem, yet solutions have been incremental rather than transformative. CMS issued guidance in recent years encouraging payers to streamline PA processes, but enforcement has been limited and payer adoption has been slow.

The AI Approach: End-to-End Automation

What's changed in the last few years is that the underlying technology has become capable of handling the complexity of prior authorization at scale. This isn't about simple automation of routine tasks. This is about deploying intelligent systems that can understand the clinical context, navigate payer requirements, manage exceptions, and execute the entire prior authorization lifecycle without human intervention for routine cases.

An end-to-end AI approach to prior authorization works like this: The system monitors the patient's EHR and billing information. When a procedure is scheduled or a prescription is written, the AI evaluates whether that service requires authorization based on the patient's specific insurance plan. It assembles the documentation package—pulling the clinical justification, relevant labs, imaging, and previous treatment history—without manual chart review.

The system then knows where to send the request. Rather than forcing staff to navigate multiple payer portals, the AI submits through the appropriate channel (electronic when available, phone when necessary, portal when required) and begins tracking the status. For routine approvals that come back, the authorization is completed and documented in the billing system automatically. For denials, the system flags them for clinical review and can initiate appeals with additional documentation.

This approach dramatically changes the economics. Instead of 310 prior auth requests per week requiring human touch points, you might handle 250 of them through end-to-end automation, with staff only intervening on the complex cases, denials, and exceptions that require clinical or administrative judgment.

Platforms like Honey Health take this end-to-end agent approach—detecting when auth is required, assembling the complete packet, submitting to the payer, tracking status, and following up without human intervention for routine approvals. Other vendors in this space, like Cohere Health, focus on the payer-provider collaboration angle, while solutions like AKASA and the Availity Intelligentum product line aim at broader RCM workflows. Each has a different architectural approach, but the best-in-class systems share the ability to reduce the human touch points on routine authorizations and surface only the genuinely complex cases for staff review.

The Practical Impact: What This Looks Like

For the MSO managing 6,500 prior authorizations per month, deploying a modern AI system typically reduces the time-to-authorization from an average of 1.8 business days to 0.6 business days for routine cases. That means quicker patient communication, fewer rescheduled appointments, and higher throughput in the procedure room.

The automation also reduces denials. By ensuring that requests are complete, properly documented, and submitted to the right contact at the payer, denial rates typically drop by 15-25%. For our example MSO, that represents roughly 1,170 fewer denials per year, each worth several hundred dollars in recovery value.

On headcount, a typical result is that you can shift 40-60% of prior authorization work from coordinators to the system. For a practice with 2.5 FTEs dedicated to PA management, that means you might reduce that to 1-1.5 FTEs—reclaiming capacity for other revenue cycle work like denials management, coding quality, or patient financial services.

The financial impact is direct. The MSO saves roughly $350,000 in annual labor costs from the FTE reduction, gains an estimated $400,000 in recovered revenue from fewer denials and faster collections, and improves throughput enough to add $200,000+ in annual clinical capacity. That's approximately $950,000 in tangible annual benefit from automating one process.

For the staff that remains, the job fundamentally changes. Instead of spending the day on data entry and payer phone calls, coordinators work on the complex authorizations, develop relationships with medical directors, and handle the exceptions that require clinical judgment. It's more interesting work, and burnout typically declines.

What's Required to Make This Work

Deploying an AI-driven prior authorization system isn't a simple software implementation. It requires integration with your EHR, connection to your billing system, and typically some configuration work to map your local payer landscape into the system's knowledge base.

It also requires a mindset shift. Rather than thinking of prior authorization as "something the billing team handles," the practice needs to build prior authorization into the clinical workflow. When a procedure is scheduled, the system needs to know about it. When a prescription is written, the system should flag authorization requirements. This means clinicians and schedulers need to engage with the system differently than they have historically.

Most importantly, there needs to be human oversight, especially in the early implementation phase. No AI system is perfect at understanding all the edge cases and payer quirks that exist. During the first 3-6 months of deployment, staff should still be reviewing what the system is doing, catching exceptions, and providing feedback. This isn't a "set it and forget it" solution—it's a partnership between human expertise and machine capability.

The U.S. RCM market was valued at $172.24 billion in 2024, and prior authorization represents a meaningful slice of that market. The good news is that investment in solutions is accelerating, which means the technology is improving rapidly. What seemed impossible five years ago—fully autonomous prior authorization management—is now demonstrable in real practices.

The Path Forward

For gastroenterology practices and MSOs reading this, the message is straightforward: the current prior authorization process is not a fixed feature of healthcare operations. It can be fundamentally restructured through technology.

The question isn't whether to automate prior authorization. The question is when, and with which partner. The vendors in this space have different strengths and approaches, and the best solution for a ten-provider independent practice in Ohio is likely different from the best solution for a 50-provider regional MSO. But across the board, practices that deploy these systems are reporting 40-50% reductions in PA processing time and 15-25% improvements in authorization approval rates.

The AMA's ongoing advocacy work shows that this problem is getting worse before it gets better. Payers aren't reducing authorization requirements—they're expanding them. But technology is finally catching up. The practices that act now to implement end-to-end prior authorization automation will recover hundreds of thousands of dollars, give their staff back their time, and most importantly, move patients through the authorization process in days instead of weeks.

For an authorization coordinator drowning in prior auth requests on a Tuesday morning, that's not just a business improvement. It's a fundamental change in what the job can be.

Learn more: For practices interested in exploring how AI-driven prior authorization management works, platforms like Honey Health offer detailed evaluations of how this could work for your specific patient population and payer mix.

More of our Article
CLINIC TYPE
LOCATION
INTEGRATIONS
More of our Article and Stories