Artificial intelligence (AI) in healthcare has moved from interesting idea to potential game-changer in just a few short years. AI’s clinical applicability, for example, has been explored at length and has found toe-holds in everything from launching sophisticated population health management initiatives to reducing unwarranted surgical variation.
However, there is a powerful argument to be made that both provider organizations and health plans can reap immediate, transformative results with AI by breaking ground on the business and administrative side of healthcare. Optimizing the prior authorization process is perhaps the most compelling example.
Prior authorizations have been a necessary evil for years. ‘Necessary’ because they were designed to insure appropriate care is being provided and to mitigate overutilization of healthcare services; ‘evil’ because they have proved to be a sand trap of administrative inefficiencies and can introduce delays in patient care.
According to a 2017 survey by the American Medical Association (AMA), 84 percent of providers consider prior auths to be overly burdensome, and 91 percent say they negatively impact patient outcomes. Seventy-four percent of providers say they wait up to three days for responses to prior auth requests.
The Making of a Quagmire
The pain point isn’t primarily the prior auths themselves, but the manual or partially automated processes by which they are requested and obtained. According the 2017 CAQH Index, an annual report of the adoption of electronic business transactions, 82 percent of prior auths are submitted manually. The report goes on to say that adoption of fully automated electronic prior auths among health plans hovers around 8 percent.
Multiple stakeholders are usually required to submit and approve requests for healthcare services, and much of their work is redundant or duplicative or both. Request and review requirements vary by health plan and provider, further complicating the process. Incomplete or excessive patient data often sparks a furious storm of phone calls and faxes. Suddenly, a simple request becomes a chronic source of healthcare waste, patient dissatisfaction and payer-provider abrasion.
The problem is so severe that earlier this year the American Hospital Association, America’s Health Insurance Plans, the AMA, American Pharmacists Association, Blue Cross Blue Shield Association and Medical Group Management Association released a Consensus Statement outlining their shared commitment to industry-wide improvements to prior auth processes and patient-centered care.
Unfortunately, most solutions on the market today—automation of certain components of prior auths or utilizing health plan portals—at best only scratch the surface of the most significant pain points, but more often lead to additional administrative burden and complexity.
AI-driven solutions, on the other hand, can cut through the fog of payer requirements, minimize provider effort and enable a process for prior-auth submission, verification and status that is not only optimized, but can learn and improve in real-time, achieving faster and more accurate responses with every subsequent request.
Why AI is the Future of Prior Auths
According to Accenture, health plans and providers can save billions of dollars by leveraging AI-driven technology for prior auth processes. However, despite this prognostication, AI in healthcare retains an aura of mystery that can be intimidating to all but the most devout health IT evangelists. So let’s peel back the curtain a bit.
At its essence, AI is automation with intelligence. A task is completed, the outcome is evaluated and refinements are made based on past experience. This is exactly how a human learns, except that AI accomplishes this on a scale far beyond what is possible for the human brain.
Prior-auth processes contain a number of components that are routine and repetitive, but also include high levels of variation and complexity. For example, prior auths cover thousands of medical services—from prescriptions and CT scans to elective surgery and clinical trials. Additionally, rules-based architecture and data exchange standards for obtaining prior auths can vary widely, depending on the practice, services and health plan.
These repetitive and complex tasks are ideally suited to AI’s ability to monitor and refine processes and make predictions to achieve even greater efficiencies. AI has the potential to transform prior auths into an entirely patient-driven process. For example, AI can mine data from lab, medication and claims data to recommend appropriate treatments and evaluate the patient outcomes.
Getting Started
Though proper governance and infrastructure is required to break ground on a net-new AI investment, this technology isn’t necessarily the heavy-lift many assume it would be. Healthcare has spent the past 20 years investing in and refining solutions to digitally capture and store billions of data elements of every patient.
The next great frontier is putting that data to work for better patient outcomes, improved efficiencies and seamless and secure data exchange among healthcare’s myriad stakeholders. By implementing technology that automates and improves the prior auth process, health systems and health plans can decrease denials, improve yield by increasing operational efficiency and scheduling capacity and enhance patient, physician and staff satisfaction.
About the Author
Ron Wince is the Founder/CEO of Myndshft, a healthcare company working at the intersection of artificial intelligence, automation and blockchain.
Sign up for the free insideBIGDATA newsletter.
Speak Your Mind