The contemporary landscape of American healthcare is increasingly defined by a high-stakes technological confrontation between providers and payers. As insurance companies deploy sophisticated artificial intelligence to scrutinize and often deny coverage requests, hospitals and physician groups are retaliating with their own AI-driven tools designed to automate the appeals process. This cyclical struggle, often referred to as the "AI arms race," has created a friction-filled environment that many industry experts warn is unsustainable. Among those calling for a ceasefire and a radical shift in strategy is Dr. Ashis Barad, the Chief Digital and Information Officer (CDIO) at the Hospital for Special Surgery (HSS).
Dr. Barad, whose career has spanned the breadth of the healthcare ecosystem, argues that the current trajectory of using AI to outmaneuver the opposing side is a "finite game" that ultimately drives up costs without improving patient health. Instead, he advocates for an "infinite game" approach—one where payers and providers move past adversarial automation and toward a collaborative model built on shared, high-fidelity clinical data. By pooling resources to create personalized care pathways, Barad believes the industry can move toward a "deflationary" use of technology that reduces administrative waste and enhances surgical outcomes.
A Career Defined by Dual Perspectives
To understand Barad’s perspective is to understand the varied roles he has occupied within the healthcare hierarchy. He began his professional journey on the front lines as a physician at Sutter Health, a major integrated delivery network in Northern California. From there, he transitioned to Baylor Scott & White Health in Texas, where he spearheaded initiatives to modernize digital tools for clinicians. This role provided him with a firsthand look at the administrative burdens that distract physicians from patient care, particularly the manual labor involved in securing prior authorizations.
In 2022, Barad’s career took a pivotal turn when he moved to the payer-provider hybrid space. He accepted the role of CDIO at Allegheny Health Network and its parent company, Highmark Health. Highmark, one of the largest Blue Cross Blue Shield affiliated insurers in the United States, offered Barad a "look under the hood" of the insurance industry. He witnessed how payers utilize actuarial data and algorithms to manage risk and control costs.
Two years ago, Barad returned to the provider side, joining the Hospital for Special Surgery in New York City. HSS is widely regarded as a global leader in orthopedics and rheumatology, performing more than 40,000 surgeries annually—a volume that dwarfs any other single-site orthopedic hospital in the country. This diverse background has given Barad a rare, 360-degree view of the systemic inefficiencies that plague the American medical billing and authorization system.
The Inflationary Trap of Automated Adversaries
The central conflict Barad highlights is the burgeoning use of AI in the prior authorization (PA) process. Historically, PA was a manual process where nurses or clerks at a doctor’s office would fax clinical notes to an insurance company, where another set of clinical reviewers would approve or deny the request based on medical necessity.
Today, that process is being rapidly digitized. Payers are employing machine learning models to scan claims and clinical documentation, often resulting in "auto-denials" for procedures that do not strictly adhere to rigid, flowchart-based criteria. In response, healthcare providers are investing in "revenue cycle AI" that can instantly generate appeal letters, cite medical literature, and resubmit claims at a volume and speed that humans could never match.
Barad views this as a losing strategy for the industry. "We’re playing a finite game when there’s an infinite game to be played," he remarked. "If we play the finite game of today, we’re going to continue on the inflationary route."
In this context, "inflationary" refers to the compounding costs of administrative overhead. When both sides spend millions on software to fight each other, those costs are eventually passed down to patients and employers through higher premiums and reduced benefits. Barad’s vision for "deflationary AI" involves using technology to remove steps from the process entirely, rather than just making those steps faster and more frequent.
The Data Gap: Claims vs. Clinical Reality
The disconnect between payers and providers is rooted in a fundamental data asymmetry. Payers typically rely on "claims data"—the information submitted on a billing form. This data tells an insurer that a patient had a surgery, whether they were readmitted within 30 days, or if they developed an infection. However, claims data is notoriously "thin" when it comes to actual functional outcomes. It cannot tell an insurer if a patient who had a hip replacement can now walk a mile without pain, or if a high-performance athlete has returned to their sport.
HSS, by contrast, sits on a mountain of "thick" clinical data. Because of its massive surgical volume, the hospital has developed a structured repository that connects advanced imaging (such as MRIs and CT scans) directly to long-term surgical outcomes. This includes Patient-Reported Outcome Measures (PROMs), which track a patient’s quality of life and physical function over months and years.
Barad recounted a recent meeting with a major national insurance carrier where he asked their actuarial teams how valuable HSS’s linked imaging-and-outcomes data would be to their models. "They were like, ‘We would salivate over it,’" Barad said.
The insurance company’s enthusiasm highlights the missed opportunity in the current system. While payers have data on the costs of complications, they lack the granular clinical insights necessary to predict which patients are the best candidates for specific interventions. If HSS can prove through data that a specific surgical technique leads to a 90% success rate in patients with a specific imaging profile, the insurer’s risk calculation changes. In such a scenario, the need for a contentious prior authorization disappears because the data already justifies the procedure.
Redefining "Gold Carding" through Personalization
One of the existing solutions to the PA crisis is a concept known as "gold carding." This practice involves insurance companies exempting high-performing physicians or hospitals from prior authorization requirements if they have a proven track record of following evidence-based guidelines and maintaining low denial rates.
However, Barad argues that the current version of gold carding is a "blunt instrument." It is usually based on cost thresholds and simple adherence to standard flowcharts. He believes the future of gold carding lies in "Gold Carding 2.0"—a model that is personalized down to the individual patient level.
In this reimagined framework, authorization would not be a separate step requested after a doctor decides on a treatment. Instead, it would be "baked into the care pathway." If a provider and a payer agree on a data-driven clinical pathway for a specific condition, any treatment that falls within that pathway is automatically authorized. This shifts the focus from "policing" individual transactions to "partnering" on clinical outcomes.
Regulatory Pressures and the Push for Transparency
Barad’s call for collaboration comes at a time of intense regulatory scrutiny regarding AI in healthcare. The Centers for Medicare & Medicaid Services (CMS) recently finalized rule CMS-0057-F, which requires many payers to streamline their PA processes and improve data exchange. Furthermore, there is growing concern among lawmakers about the "black box" nature of payer algorithms.
In late 2023 and early 2024, several class-action lawsuits were filed against major insurers, alleging that AI tools were used to prematurely cut off care for elderly patients in post-acute settings. These legal and regulatory pressures are forcing payers to seek more transparent and defensible ways to manage utilization. Barad’s proposal of using shared, real-world clinical data offers a path toward that transparency.
The Broader Impact: From Competition to Coordination
The implications of shifting from an AI arms race to a data-sharing partnership are profound. For physicians, it promises a significant reduction in burnout. According to a 2023 American Medical Association survey, 94% of physicians report that prior authorizations lead to delays in care, and 33% report that PA has led to a serious adverse event for a patient in their care. By automating trust through shared data, the time spent on administrative "back-and-forth" could be redirected to patient interaction.
For the healthcare system at large, Barad’s "infinite game" approach addresses the root cause of rising costs. Instead of adding layers of technology to manage a broken process, the industry has the opportunity to use AI to refine clinical pathways, ensuring that the right patient gets the right surgery at the right time.
"We have to figure out how to work together to figure out what the deflationary path for AI is," Barad concluded. "That’s what we owe the people, honestly."
As the technology continues to evolve, the choice for healthcare leaders is becoming clear: continue investing in tools that facilitate corporate warfare, or pivot toward a collaborative framework where data serves as the bridge between paying for care and providing it. For Ashis Barad and the team at HSS, the goal is no longer to win the battle of the algorithms, but to change the rules of the game entirely.
