The current process of obtaining prior authorization (PA) for healthcare services is inefficient and burdensome for everyone involved, from patients to physicians and health plans. PA functions as a cost-control strategy for health plans, who require providers to obtain advance approval for a wide range of services in order to qualify for reimbursement.
From the physicians’ point of view, the manual work of submitting PA requests (41 per week, on average) is purely bureaucratic, as they are already prescribing what they believe to be medically necessary care. An overwhelming 91% of surveyed physicians also believe PA has a negative impact on clinical outcomes, as care delays can lead patients to abandon treatment or suffer from serious adverse events.
While health plans realize that the vast majority of authorization requests are eventually approved, they are reluctant to eliminate PA entirely. When requirements are relaxed, deviations from standard practice and evidence-based medicine tend to increase, leading to both higher costs and unnecessary care.
While the Centers for Medicare and Medicaid Services (CMS) has proposed a new rule designed to streamline PA processes and drive greater interoperability between providers and payers, progress has been slow. Despite advancements in technology and a shift to value-based care, the current approach to PA continues to focus on each proposed service in a vacuum, divorced from the larger clinical context.
Without a more collaborative process, lower-value care will persist.
By using artificial intelligence (AI) and machine learning (ML) capabilities to automate many of the manual processes associated with PA, health plans can incorporate critical patient context and forge a more efficient system built on a longitudinal view of the patient. AI and ML can ease the administrative burden of PAs while encouraging physicians to make the most clinically appropriate, high-value care decisions for their patients.
Using Evidence-Based Clinical Recommendations to Improve Outcomes
An intelligent authorization platform leverages interoperability and ML to survey the patient’s unique clinical history and surface recommendations on high-value care decisions that are a good fit for that particular patient. As a PA request is entered, ML models can produce probability estimates—which rank how completely the information matches both what is expected and what is considered best practice—that trigger automated prompts.
For example, the ML model might detect that a physician requesting an injection for a patient has not supplied evidence of advanced imaging within the clinical notes; this omission can immediately prompt the platform to request imaging documentation before proceeding.
In addition to ensuring a complete request, intelligent authorization platforms can also offer automated, evidence-based clinical recommendations to promote best practices. For example, the platform may suggest a provider switch from an inpatient to an outpatient setting when the evidence warrants it, creating a valuable opportunity to improve patient outcomes, reduce inappropriate utilization, and decrease unnecessary medical expenses.
Identifying Episode-Based Care Paths to Anticipate Needs in Advance
An AI-driven authorization process can also identify episode-based care paths based on the patient’s diagnosis, which can be used both to accelerate PA approval and support evidence-based care choices.
As physicians enter a CPT code for a particular service, an intelligent authorization platform can automatically suggest additional services that might be appropriate for a bundled authorization. For example, a patient undergoing knee surgery might require outpatient physical therapy as well as additional rehabilitative services. By consolidating multiple authorizations for a single episode of care, intelligent platforms can save physicians the time and expense of completing several PA requests in a row, while also improving healthcare access for the patient.
Physicians are incentivized to accept evidence-based recommendations on such areas as the site of service, therapy usage, and postoperative care planning because they know they are building a more clinically appropriate case, which is far more likely to be approved. With full transparency into the authorization process, physicians can see not only the guidelines that inform the health plan’s medical necessity criteria but also the health plan’s approval policies. Automatically generated recommendations help accelerate the onset of care while reducing overtreatment.
Rewarding High-Performing Physicians
Although most providers include adequate data and information with their requests for treatment, even physicians with high PA approval rates are required to submit PAs, which results in unnecessary provider friction.
By integrating AI and ML into UM processes, physicians can be tiered according to trust levels for those who consistently make evidence-based, high-value care decisions.
Rewarding high-performing physicians can improve overall satisfaction with the health plan’s UM program while simultaneously incentivizing evidence-based care choices.
By automatically expediting PA requests from physicians in the highest tiers, physicians who frequently fail to provide the necessary documentation or whose requests often deviate from accepted clinical standards can be encouraged to change their behavior. An intelligent authorization platform can suggest more solid, evidence-based treatment plans that better suit the individual patient.
Using AI and ML to reimagine and augment our existing PA processes will help reduce the administrative burden physicians face, eliminating unnecessary friction. The combination of intelligent automation and transparency also helps drive PA submissions that result in real-time approvals, giving patients faster access to the high-value treatments they need.
Share:
About the Author: Cohere Health
