Are you still trying to address some of the biggest pain points in your revenue cycle? AI technology may soon be our answer.

AI and machine learning applications are not a new phenomenon in healthcare, but their use is accelerating quickly as new use cases and apps develop quickly. Some of the best processes to apply AI to are those that are heavily manual, time consuming, repetitive, and expensive. Many parts of the revenue cycle fit this description perfectly, introducing new opportunities to automate and optimize processes.

Fifteen cents of every US healthcare dollar go towards revenue cycle inefficiencies. That’s over $400 billion spent on claims processing, payments, billing, RCM, and bad debt. Much of this cost is associated with manual methods, such as phone calls and mailings.

Tackling Prior Authorization Headaches

In an AMA survey of 1,000 practicing physicians, 86% of doctors described the burden of prior authorizations (PAs) as high or extremely high. Prior authorizations often entail a lot of back-and-forth between the payer and the provider, adding up to a lot of time spent. In fact, the AMA survey found that physicians and their staff spent almost two full business days each week on prior authorizations. Also, most physicians had staff that were hired to work exclusively on PAs.

In spite of these statistics, adoption of electronic prior authorization processes remains fairly low. Barriers to implementing electronic processes include lack of operating rules and standards, infrastructure, and organizational readiness. In spite of these challenges, some organizations are finding a great deal of success using emerging technology – artificial intelligence.

Prior authorization is listed as one of the BEST use cases for artificial intelligence in healthcare right now, due to the transactional nature of the tasks. AI can leverage real-time analytics and machine learning to easily identify those cases that need prior authorization, submitting the requests to payers, and then checking statuses.

Claims Management Automations

Revenue cycle processes already use a lot of codified data – from CPT and ICD-10 codes to denial codes. These data points are relatively easy to “teach” an AI application to work with, making many manual tasks able to be automated with the right technology in place.

By training the AI with algorithms that imitate intelligent human behavior, the AI can find patterns and plan future actions in order to produce the right outcomes. This differs from previous machine learning apps that follow an algorithm to present an answer. AI actually “learns” from the processes and is able to plan future action based on historical performance.

So what tasks are best suited for this technology? Some of the most important and troubling parts of the cycle – denials management, claim status checks, out-of-pocket cost estimates, and of course, prior authorizations. Better yet – the information can be had in real time, as the care episode is scheduled and while the patient is present. For example, providers need to know in real time if the patient is cleared financially for a certain treatment so that the patient can make the most inf9rmed decisions at the time.

MidLantic Urology leveraged AI through a financial clearance automation solution that offered visual cues to practice executives so that they know where each patient is in the financial clearance process, in real time. This solution helped the practice improve gross revenue by 18%, with an average of 85% fewer claims lingering around to be worked on at any given time.

This workflow in turn equaled a 50% reduction in anticipated staffing requirements for its billing department. In a time when revenue cycle employees are increasingly hard to find, this opportunity to optimize staffing is exciting.

Not everyone is on board yet but results like these are promising indeed. About half of healthcare organizations use some form of AI, but 88% of them expect more widespread implementation in the next five years.

Clinical Documentation Improvement

CDI programs act as another layer of review to make sure that the clinical documentation reflects all of the patient’s diagnoses, comorbidities, and complications – both for accurate charge capture and accurate quality reporting. Nurses who are adept at reading charts and have some coding knowledge as well look for opportunities to increase the accuracy of the medical record by querying physicians for needed clinical data and details.

So how is AI used to enhance CDI processes? Again, the process leans heavily on codified data with already defined tags. New platforms embed AI in both the clinician and CDI workflow, analyzing EHR notes and clinical data to find gaps and deficiencies before the notes are even saved to the EHR.

Black Book Research surveyed 128 hospitals with between 150-400 beds after AI implementation for CDI programs. In four quarters, they accomplished an average of $1.6 million from case mix improvements.

Doug Brown, President of Black Book Research states that:

“AI-enhanced CDI also allows providers to explore new opportunities with payers because hospitals have the data to show where quality outcomes are performing best.”

The Rise family of companies work together to provide unmatched revenue cycle support and transformation, with a multitude of tools – including AI, analytics, dashboards, EHR optimization, productivity tracking, and more. We can help customize a pathway to improvement for your organization. Simply contact us to get started.