The integration of artificial intelligence (AI) into the healthcare system, particularly within the complex administrative process of prior authorization, is creating a new frontier of challenges and opportunities for physicians, patients, and insurers alike. What was once a human-driven, albeit often cumbersome, administrative hurdle is rapidly transforming into an automated exchange, raising concerns about efficiency, accuracy, and the erosion of the patient-provider relationship. This shift is prompting regulatory bodies, healthcare providers, and technology developers to re-evaluate the governance and implementation of AI in medical claims processing.
AI-Driven Denials: A New Wave of Administrative Burden
The experience of Yolanda Troublefield, MD, JD, an attending surgeon and otolaryngologist at Southcoast Physicians Group, exemplifies the growing frustration among medical professionals. Dr. Troublefield recently encountered a situation where a prior authorization request for an eight-year-old child with recurrent otitis media, a condition typically requiring adenoidectomy with ear tube placement, was denied. Despite comprehensive documentation from both primary care and otolaryngology, including an audiogram, and the parents’ own medical backgrounds, the denial was issued.
Dr. Troublefield attributes the denial, at least in part, to the AI algorithms employed by the health insurance company. She observed that the large language model seemed to focus on a single element within the medical note, overlooking the broader clinical context. Specifically, while the diagnosis was "recurrent otitis media," the assessment-and-plan section of the note, intended to address the immediate need for an antibiotic during that visit, stated "acute otitis media." This semantic discrepancy, seemingly minor to a human clinician, was sufficient for the AI to flag the request for further review, potentially necessitating a time-consuming peer-to-peer conversation that disrupts clinical schedules and delays patient care.
"We’ve been doing tonsils and adenoids and ventilation tubes for the last 50 years—seems pretty straightforward," Dr. Troublefield remarked. "Every single ‘i’ has been dotted, ‘t’ has been crossed." However, this meticulous preparation proved insufficient against the automated review process. The implication is that AI systems, in their current iteration for prior authorization, may not possess the nuanced understanding required to interpret complex medical documentation, leading to what many perceive as arbitrary denials.
The Acceleration of Denials and Physician Exasperation
The introduction of AI into prior authorization processes and retrospective claim reviews is leading to an unprecedented pace of claim denials, often in ways that physicians find perplexing. This necessitates increased administrative work for providers to secure coverage for necessary patient care, fueling a growing sense of exasperation. Health practices are increasingly adopting AI to both submit documentation and respond to denials, creating automated feedback loops with potentially unforeseen consequences for the healthcare ecosystem.
"What we’re seeing is faster denial of claims because of the use of AI," Dr. Troublefield stated. She described the AI’s denial thresholds as adjustable dials, with current settings "set high for denials." This suggests a deliberate configuration by insurers to maximize claim rejections, thereby reducing their financial outlay.
Data from a recent American Medical Association (AMA) physician survey corroborates these concerns. Sixty percent of physicians surveyed expressed apprehension that AI would either increase or had already increased prior authorization denial rates. Furthermore, 55% reported that prior authorization delays access to necessary care "always or often," and a significant 79% indicated that it sometimes leads to patients abandoning treatment altogether. Alarmingly, 26% of physicians reported that prior authorization had resulted in a serious adverse event for a patient under their care. This highlights the tangible, and potentially life-threatening, impact of these administrative hurdles on patient outcomes.
Bruce Scott, MD, a former president of the AMA and an otolaryngologist at ENT Care Centers in Louisville, Kentucky, shared his initial hopes for AI as a solution to administrative burdens. "I think physicians were very hopeful that AI was going to be a boost, a solution if you will, to reduce the administrative burdens that we all face," he said. However, he acknowledged that this optimism has not materialized.
Dr. Scott articulated a vision of escalating automated conflict: "Eventually, their bot is going to talk to my bot. Because my bot is going to tell me how to document so I get authorization, and then their AI is going to get even smarter and deny that, so I get a denial letter back. And in the meantime, the physicians and patients are stuck in the middle, and that’s the problem." This "battle of the bots" scenario underscores the potential for an arms race in automation, where efficiency gains for one party come at the expense of another, leaving the core mission of patient care compromised.
Regulatory Intervention and the Shifting Landscape of AI Oversight
In response to the growing concerns, a wave of new state laws has been enacted or is on the horizon to regulate the use of AI in claim denials. These legislative efforts aim to introduce a human element and patient-centric considerations back into the automated decision-making process. For instance, an Alabama law mandates that insurers base determinations on a patient’s unique circumstances rather than relying solely on aggregated datasets. Similarly, an Indiana law prohibits insurers from using AI as the exclusive basis for downcoding claims, ensuring that clinical judgment remains paramount. A Washington state law requires that claim determinations be made only by licensed and qualified health professionals, reinforcing the principle of expert oversight.
Beyond state-level interventions, federal initiatives are also emerging. The Centers for Medicare and Medicaid Services (CMS) is beginning to leverage AI for its own health insurance programs. The WISeR (Wasteful and Inappropriate Service Reduction) program, launched in January in six states—New Jersey, Ohio, Oklahoma, Texas, Arizona, and Washington—employs AI alongside human review to "reduce clinically unsupported care by working with companies experienced in using enhanced technologies to expedite and improve the review process for a pre-selected set of services that are vulnerable to fraud, waste, and abuse." While this program aims to combat fraud and abuse, its reliance on AI for pre-selected services raises questions about potential overreach and its impact on legitimate claims.
Combating Automation with Advanced Automation: A Physician’s Response
The pervasive use of AI by insurers has compelled healthcare providers to adopt similar technologies to navigate the evolving administrative landscape. Bradford Bichey, MD, a rhinologist at Indiana Sinus Centers, has developed an AI-powered product designed to streamline physician office operations, including the crucial task of generating documentation for prior authorization submissions.
Dr. Bichey recounted a period several years ago when his practice began experiencing a significant increase in "clawback attempts"—requests for documentation for cases that insurers deemed should not have been covered. He observed a dramatic surge from approximately five such requests to fifty per period, noting that the language used in these requests often suggested "early experimentation" with AI by the insurers. The verbiage was "slightly off," with one particularly jarring letter described as "almost hateful." Over time, these communications have become more professionally couched, but the underlying intent to scrutinize and deny claims has persisted.
In response to this trend, Dr. Bichey developed "Blue," an AI product under his company, Nemedic. Blue generates not only clinical notes based on patient encounters but also tailored prior authorization documents specific to each insurer’s requirements for a given indication. The system is designed to recognize keywords—such as "United" to trigger a template for United Healthcare—and then access publicly available insurer documents to generate a submission that meets their precise criteria. These submissions often require extensive justification, far beyond the scope of a standard clinical note. For example, a prior authorization for a nasal endoscopy might include detailed explanations about the limitations of less invasive examinations and the medical necessity of direct visualization for accurate diagnosis and management. The AI also meticulously notes exclusions, clarifying that the procedure was not for routine screening but to inform treatment decisions.
"You have to know [the insurer’s] inclusion and exclusion," Dr. Bichey emphasized, highlighting the complexity of keeping up with hundreds of different insurance plans. His AI solution aims to alleviate this burden by automating the process of understanding and adhering to these intricate requirements. He reported that by leveraging this advanced documentation, appeals have become simpler, often involving merely pointing to overlooked information in the initial submission. Dr. Bichey has not needed to engage in a peer-to-peer call in two years due to the improved quality of upfront documentation.
AI: A Symptom of Governance Deficits, Not the Root Cause
Matthew Crowson, MD, assistant professor of otolaryngology—head and neck surgery at Harvard Medical School and director of clinical informatics and artificial intelligence at Massachusetts Eye and Ear, offers a nuanced perspective on the role of AI in prior authorization. He argues that the AI itself is not inherently the problem, but rather the "governance around this stuff, or lack thereof."
Dr. Crowson contends that AI has not resolved the fundamental tension between providers and payers. Instead, it has amplified existing decision-making workflows, leading to what he describes as a "battle of the bots." The application of AI by both providers and payers is seen as a double-edged sword, intensifying the inherent conflict over medical necessity.
However, Dr. Crowson also acknowledges AI’s potential to enhance administrative efficiency in ways that benefit patients. For instance, when a prescribed medication is not on an insurance plan’s formulary, a manual appeal process could take days. AI can now draft a contextualized appeal rapidly, potentially expediting medication access for patients. "What would take a human process maybe seven business days to do, now you can do it in one," he noted, suggesting a significant acceleration in administrative response times.
Efficiency Gains or Escalating Activity? The Unclear Impact of AI
A recent report by the Peterson Health Technology Institute, based on workshops involving a diverse group of healthcare stakeholders, paints a complex and somewhat ambiguous picture of AI’s impact on the U.S. medical landscape. While AI may reduce the cost of prior authorization for individual organizations, the report indicates that it has not yet led to a systemic reduction in overall healthcare costs.
Furthermore, the deployment of AI by providers is reportedly increasing "billing intensity," characterized by the identification of more severe diagnoses and advanced treatments, which inherently carry higher associated costs. Workshop participants expressed concern that the optimized use of AI by both insurers and providers risks making the entire process "more activity-intensive" without necessarily achieving genuine efficiency. One healthcare provider quoted in the report lamented, "Bots don’t get tired of asking questions, so my review queue keeps growing," illustrating the potential for AI to create an endless cycle of automated administrative tasks.
Patients Caught in the Crossfire of the Automated Arena
The ongoing struggle with prior authorization, now increasingly mediated by AI, continues to strain the physician-patient relationship. Dr. Troublefield highlighted that complex discussions with patients about insurance coverage hurdles are a daily reality. "It takes more time—I don’t think any of us runs on schedule at all," she stated, contrasting the current situation with five years prior when punctuality was more achievable. "We’re finding that we’re constantly having to apologize, apologize, apologize. And that’s not the way people want to be treated." This erosion of trust and the need for constant apologies underscores the human cost of administrative inefficiencies.
Dr. Scott recounted a poignant example of a patient with a maxillary sinus tumor who had finally accepted the necessity of surgery. The patient’s confidence was shaken when the insurer denied the procedure, citing a preposterous rationale that the patient had not yet undergone antibiotic treatment. This denial led the patient to question the surgical recommendation, prompting Dr. Scott to engage in a delicate process of rebuilding trust. "Now I’ve got to go explain to this patient, and I’m thinking about the trust relationship that I had with this patient and having to re-establish that trust," he explained.
Dr. Scott advocates for physicians to appeal denials more frequently, noting that his practice appeals every denial. However, he recognizes that physician burnout can make this an untenable burden for many. He and the AMA are actively pushing for greater physician involvement in the development of regulations and processes surrounding AI in healthcare, aiming to ensure that the technology serves rather than hinders patient care.
A Critical Juncture for Prior Authorization: Navigating the Path Forward
Dr. Crowson posits that the field is at a critical juncture, facing two potential paths. The first involves both providers and payers utilizing AI to process routine paperwork and basic tasks more efficiently, freeing up human resources for complex cases. The alternative path is an "escalating compliance and gaming war," where each side continuously refines its AI to outmaneuver the other.
He expresses hope for a "pragmatic, middle-of-the-road approach," advocating for the use of AI in its strengths: automating administrative monotony and other simple tasks that can be safely and efficiently scaled. Dr. Troublefield voiced concerns that AI might simply introduce "another hoop" for physicians to jump through, requiring adherence to specific "buzzwords." However, she finds the concept of software that automatically generates insurer-compliant templates an attractive prospect, provided it is practical to implement.
Ultimately, Dr. Troublefield believes that a combination of technological solutions and smart legislation, such as regulating AI’s role in insurance decisions, could offer a path forward. She also stresses the importance of amplifying patient voices, and even celebrity endorsements, to bolster legislative efforts. In this rapidly evolving landscape, physicians must adapt. "The system is changing, and either you have to be part of the change, or you’re going to get left behind," Dr. Troublefield concluded. "It’s the same thing as when the automobile happened, right? You’re not stopping it, so you have to figure out ways within the system to make sure that what you want, and what your goals are, are actually achieved." This proactive engagement is essential to ensure that the integration of AI in healthcare aligns with the core principles of patient well-being and medical necessity.
