After waiting weeks for an open appointment, John finally gets in to see the orthopedic surgeon to evaluate his acute knee pain. He’s already been off work for over two months and has barely been able to leave the house, popping anti-inflammatory and pain medicine to manage the pain every day.
Suspecting a torn meniscus, the doctor orders an MRI to verify the diagnosis and plan his treatment approach. The nurse begins the process by calling John’s insurance carrier to verify coverage, determine the need for prior authorization, and submit the request. But after three weeks of back-and-forth, faxing additional documentation, and waiting for clinical review, John still hasn’t received approval for the MRI.
The doctor and his staff have spent an exorbitant amount of time on the approval process, and John is increasingly frustrated by the wasted time. The doctor wants to get John scheduled for surgery ASAP, but this bottleneck is preventing him from delivering the care his patient needs. Meanwhile, the insurance company is spending dozens of hours and hundreds of dollars on the clinical review process for this straightforward and routine case—resources that could be better spent on more complicated cases.
Frustrated and still in pain, John can’t afford to be off work any longer, so he takes it upon himself to double his dose of pain medication and heads back to the job site. Every day is torture, and he spends every evening on the couch, icing his knee. He can’t play with his kids or mow his lawn and now has to add a daily proton-pump inhibitor to his drug regimen to deal with the gastric side effects from the anti-inflammatory medications. He’s depressed and gaining weight, which compounds his already genetically high risk of heart disease.
Originally designed as a cost-control measure, prior authorization (PA) is widely recognized as one of the most time-consuming processes for health plans. In addition to creating an overwhelming burden on physicians who are already strapped with labor shortages and pandemic burnout, it also severely impacts patient care, with nearly four in five physicians reporting that PA causes patients like John to abandon a recommended course of treatment. Worse yet, 34% of providers say PA has resulted in a serious adverse event for a patient, including 24% having a patient hospitalized and 26% reporting a life-threatening situation, including permanent disability or death.
PA Reforms Overlook the Core Problem
Recent efforts to reform PA through proposed legislation and mandates have focused on standardizing the documentation and submission workflows and enforcing a fixed turnaround time for payer decisions. While both measures would help, two major obstacles remain:
– Once enacted, implementing new workflows would likely take at least 12-18 months as providers and payers work through the process of data standardization and API integration. That could mean years before these measures would make a difference; meanwhile, millions of patients continue to suffer.
– They still don’t address the root cause of the problem—the decision-making process. Already, as many as 13% of prior authorizations are often wrongfully denied, and streamlining submission doesn’t improve or accelerate payers’ internal processes. As a result, most payers would likely suffer the penalties of not meeting decision turnaround mandates, which only adds to the cost and doesn’t help patients at all.
While digitizing PA transactions would help—currently only 21% are electronic, the lowest of all claims transactions—and provider “gold card” systems could reduce some of the burden, these measures don’t go far enough because they ignore the root cause. That means payers who invest heavily in temporary solutions today will likely find themselves forced to invest more down the road.
Instead, we need a comprehensive approach that addresses bottlenecks at the primary source: the decision-making process.
AI Decision-Making Offers a Viable Solution
AI-powered predictive decision-making can offer a future-ready solution with the potential to enable real-time PA decisions, drastically reduce turnaround times, accelerate patient care, and improve outcomes.
Unlike conventional rules engines that look at whether PA submissions and clinical documentation meet specified criteria, the AI approach to PA relies on payer administrative data to conduct instant reviews based on historical decisions. By examining several years’ worth of decision history and building a predictive model based on that data (e.g., diagnosis, procedure requested, provider source, etc.), algorithms can automatically approve PA requests and route only the truly complex cases for manual review.
In John’s case above, if a payer historically approves 95% of MRI requests coming from John’s orthopedic surgeon, why spend the time and money on a clinical review? Rather than controlling utilization costs, John’s situation actually drives up administration cost, causes frustration for the provider and the member, and delays care, potentially forcing additional, more costly interventions. AI can optimize the request reviews, creating a streamlined process that gets patients like John the care they need, when they need it.
AI: An Immediate Solution that Serves Payers, Providers & Patients
In addition to addressing the root cause—the decision-making process—to substantially lower PA friction, burden, and cost, the AI approach also offers a more immediate solution. Where data standardization and new submission processes would take months to implement across the thousands of providers and hundreds of payers, AI solutions can deliver early benefits to organizations by first auditing their decisions to find opportunities for efficiency. With full implementation in as little as 60-90 days, payers can begin optimizing PA requests this quarter, instead of sometime in the next couple of years.
It’s important to note that, when deployed as a supplement to the typical PA process, AI solutions don’t force a complete rip-and-replace of current systems. In fact, providers and payers don’t have to change anything—the AI review “bolts on” to their existing workflow, performing an initial review of requests and recommending manual reviews as needed. Payers can leverage the speed and scalability of the AI audit and review before committing to more expensive systemic change.
According to McKinsey, AI-enabled PA could automate up to 75% of manual tasks and contribute toward some $35 billion in savings across the healthcare continuum. In the drive toward value-based care, where efficiency and patient outcomes are paramount, the impact could be exponential, substantially reducing manual reviews to eliminate bottlenecks and accelerate care.
While legislative and regulatory pressure may be exactly what’s needed to push the industry to reform the PA approach, AI-based PA solutions present an immediate opportunity that addresses the root cause, rather than scrambling to address emergent mandates down the road. In the move toward value-based care, adopting innovative AI solutions now creates a solid foundation for future success and offers a key competitive advantage for payers.
About Matthew Plack
Matthew Plack is vice president of innovation and new solution strategy at Apixio, where he focuses on improving healthcare operations and outcomes through artificial intelligence and advanced technology.