Emergency Medicine
Prior authorization (PA) began with a seemingly logical premise: let’s have doctors request upfront approval before the insurance company pays for the healthcare services prescribed. But as physicians, we all know that PA is deeply flawed: It fundamentally alters the practice of medicine. It reduces physician autonomy, it adds to our already overwhelming administrative workload, and it leads to significant emotional and financial stress for patients. All of these can lead to negative outcomes for patients. While insurers wrap prior authorization in rhetoric about "quality of care" and "patient safety," its implementation often actively undermines these very goals.
Consider these scenarios which I've witnessed through colleagues:
A patient with aggressive cancer, prescribed a targeted therapy based on their specific tumor genetics, forced to try and fail multiple older treatments before accessing the recommended medication. Each week of delay represents not just lost time, but lost opportunity for effective intervention.
An epilepsy patient, finally stable on a medication regimen carefully calibrated by their neurologist, suddenly forced to switch to an alternative by their insurer. The resulting breakthrough seizures don't appear in any insurance company metrics, but they fundamentally disrupt a life previously brought under control (and waste money that the insurer is ironically trying to save!)
There are countless similar situations which underscore how PA has strayed from its original purpose and created a bureaucratic labyrinth that delays care, increases costs, and fundamentally damages the doctor-patient relationship. (Need more proof? Here’s over 100 replies from physicians on Bluesky when I asked the question.)
As physicians, we have the power to reform this broken system – look no further than a new bill that physicians like Dr. Tina Shah helped to enact in New Jersey that goes into effect this week. But in order to do so, we must understand the history of PA and the current technological and clinical trends shaping PA today.
Prior authorization emerged in the 1980s and 1990s. It targeted expensive, discretionary procedures like elective surgeries and advanced imaging, and aimed to prevent unnecessary interventions and protect patients from potential harm. The original intent wasn't entirely misguided — after all, not every child in the 1950s needed their tonsils removed, though physicians were happy to perform these profitable procedures. Early PA tried to rely on solid evidence to ensure appropriate care while avoiding overutilization. For example, “Maybe the patient who threw his back out but is now doing better doesn’t need an immediate MRI.”
However, this system changed over time. Entities who had decision-making authority also had:
While physicians are ethically and legally bound to prioritize patient welfare, insurance companies and their reviewers operate under a fundamentally different set of incentives.
As physicians, you probably know how the PA process works. But in case you’re early in your medical career, here’s a primer:
The doctor is not paid for this time. He or she often has to employ medical staff just to handle prior auth requests. And the patient waits in limbo for days, weeks, or months for a treatment their doctor has told them is necessary for their health.
Prior authorization has evolved into an all-encompassing mechanism, far removed from its original intent. This transformation was driven by three main factors:
Perhaps most troubling is how prior authorization exploits a fundamental asymmetry in healthcare delivery:
Physicians must think twice before ordering tests or treatments as they’re now put between a rock and a hard place: ordering the intervention forces them to do additional unpaid work in order to advocate for their patient. Not ordering the intervention breaks our ethical code to put patients’ needs always above our own.
PA steals time from patient care. The numbers are staggering: physicians spend an average of 14 hours per week wrestling with prior authorization requirements. That's nearly two full workdays not spent on direct patient care. Many practices now employ full-time staff solely to handle these requirements. We also know that physicians must navigate PA even for standard, evidence-based therapies, like generic nausea medicines for patients undergoing chemotherapy (Zofran).
What once might have been a concise, clinically focused note has become a bureaucratic exercise in anticipating insurance company demands. We're no longer just documenting for clinical care — we're writing to satisfy non-medical reviewers who may not understand something as routine as "carcinoma" meaning cancer.
PA erodes physician autonomy. The prior authorization system fundamentally challenges physician autonomy and the doctor-patient relationship. And the psychological impact on physicians cannot be overstated. We took an oath to put patients first, yet increasingly find ourselves unable to fulfill that promise without first seeking permission from insurance company algorithms. Prior authorization applied broadly assumes that the doctor is “guilty of poor medical judgment” until being proven “innocent” by the insurer. We are also the face of the healthcare system, and are ultimately the one that will be blamed for delays and denials.
While insurers frame PA as a quality control measure, the lived reality for patients often reveals a system that creates barriers to care rather than protections.
The statistics are striking: 93% of providers report that their patients experience treatment delays due to prior authorization. Oncologists report similar numbers with 80% reporting cancer progressing due to delays and 36% reporting deaths due to PA delays. Behind each percentage point lies a story of suffering, anxiety, and eroded trust in the healthcare system.
The ripple effects also extend far beyond immediate care delays:
Even if insurance executives proclaim that prior authorization is not intended to make them money, it is impossible to deny that it does. Here’s how:
Direct Cost Avoidance: Beyond simple denial of services, the mere existence of PA requirements creates a deterrent effect. Physicians, knowing the burden of the authorization process, may hesitate to prescribe certain treatments—even when clinically appropriate. A 2021 study in Health Affairs estimated that prior authorization from insurers costs everyone else in the healthcare system $90B a year.
The Power of Delay: Even when insurers ultimately approve procedures, delays in authorization translate directly to financial gains. Every week of delayed payment represents additional time that premium dollars can generate investment returns for insurance companies.
Algorithmic Efficiency: The rise of AI-driven denial systems has made it increasingly cost-effective for insurers to implement broad PA requirements. The human cost of these automated decisions remains largely unexamined, but ProPublica’s reporting on UnitedHealthcare’s AI tools is quite damning. STAT News also reported on an AI tool that auto-denied 90% of claims for UHC, and NY Mag also has a very nice overview of prior authorization.
After years of mounting frustration, we're finally seeing meaningful pushes for reform. Perhaps most promising are state-level reforms, where laboratories of democracy are experimenting with different approaches to taming the prior authorization beast. These efforts reflect a growing recognition that the status quo has become untenable for both patients and providers.
Take New Jersey's bold new law, effective January 1:
This represents a fundamental shift in approach—rather than accepting insurer delays as inevitable, it treats time as a precious clinical resource.
Texas has taken a different but equally innovative path with its "gold card" system. By exempting high-performing providers from routine PA requirements, it acknowledges what we've long known: the vast majority of physician requests are ultimately approved once appealed (if they have the time and energy to fight the denial). This approach rewards clinical expertise while preserving resources for cases that truly warrant review.
Other states have attempted smaller prior authorization reform, for example Colorado has changed PA requirements for some medications and Oklahoma has proposed mandating the publication of prior authorization standards on insurer websites for transparency, while New York has banned prior authorization for the first 14 days of mental health hospitalizations. There’s a great PA state-level tracker from the ACP available as well.
While Congress remains gridlocked on many healthcare issues, even here we are beginning to see signs of movement. The Improving Seniors' Timely Access to Care Act, despite its current holding pattern in the House, represents a bipartisan acknowledgment of PA's systemic failures.
More concretely, CMS's Interoperability and Prior Authorization Final Rule signals a shift in regulatory thinking. Though its implementation timeline stretches years into the future, its requirements for electronic processes and transparent denial reasoning suggest a recognition that the current system's opacity serves no one but insurers.
The emergence of AI in this space presents a fascinating dichotomy. While startups promise to streamline authorization processes, there are clear potential pitfalls. I can't help but imagine a future where AI authorization bots engage in endless loops with AI denial algorithms, creating a high-tech version of our current gridlock, and I’m not alone.
Yet innovative approaches are emerging. Mark Cuban's initiative to eliminate PA requirements for his companies' self-insured plans represents a radical rethinking of the employer-insurer relationship. By trusting physician judgment while using analytics to identify true outliers, this model might chart a path toward more efficient, trust-based systems.
And don’t count out the EHRs; this year at UGM, Epic’s Judy Faulkner announced new tools with payor partners to streamline the PA system as well.
While recent reform efforts offer hope, they also highlight the scale of change needed.
If you look at any chart of American healthcare spending, more prior authorization has not slowed healthcare spending. By contrast, numerous other countries have far more effective strategies to contain costs: Taiwan has global budgets, Japan uses a fee schedule, the UK’s NICE performs cost-effectiveness reviews, and Germany does price negotiation. But in America, we have uniquely adopted the least effective option that incurs the biggest costs to our patients physical, mental, and financial well-being.
Here are two other options that I think might solve the prior authorization pain:
For those of us on the front lines of medicine, the path forward is clear: We must continue advocating for our patients while working toward systemic reforms that restore clinical judgment to its proper place.
We must also restore trust in physician judgment. While oversight has its place, the current system's presumption of physician error unless proven otherwise fundamentally misaligns incentives and undermines care. (And physicians and health systems must hold up their end of the bargain as well.)
As we push for change, let's remember what brought us to medicine in the first place— the desire to heal, to comfort, and to serve. Prior authorization too often stands between us and these fundamental duties. It's time for a system that supports rather than impedes these essential goals.
The cost of inaction is paid daily in patient suffering, physician burnout, and wasteful bureaucracy. We can — and must — do better.