Account resolution with machine learning and revenue cycle automation

By Ken Ruiz; Ryan W. Hartman, MHSM, CHFP; and Matthew M. Szaflarski
| 8/8/2022
Account resolution with machine learning and revenue cycle automation

As healthcare revenue cycle departments continue to face resource constraints that show no signs of easing, today’s leaders need to find ways to accomplish work more efficiently. But when hiring additional staff isn’t an option, what can they do?

One solution lies in taking advantage of their organization’s data to identify opportunities within the revenue cycle to direct precious resources toward cash-generating accounts – and away from low-cash-opportunity accounts.

The issue: Too much time, too little payoff

A typical accounts receivable (AR) team within a healthcare revenue cycle department expends significant time, energy, and manual effort analyzing credit and debit accounts and identifying appropriate resolutions. For example, the typical time to resolve a single credit account is 15 minutes on average.1

Much of this time is spent resolving or pursuing accounts that have a low likelihood of returning dollars to the organization. These lower-return activities include issuing refunds, working false credits, and correcting adjustments, to name a few. And while all of these activities need to be performed, they don’t require the specialized knowledge of revenue cycle team members. Therefore, these types of activities are prime candidates for revenue cycle automation. Using data, machine learning, and analyses to automate the resolution of these nonvalue accounts can help free up staff to be more available to work on more complex, value-added patient accounts that require an employee’s knowledge and experience.

What should be automated? Using machine learning models

Understanding where the value lies in their accounts is key to revenue cycle leaders’ ability to apply their best resources to accounts that generate the most cash for their organizations and limit the human capital dedicated to low- or no-cash-generating accounts. Use of data science and machine learning can help revenue cycle leaders pinpoint which types of accounts – among their credits, debits, and denials – generate the most cash by helping to identify patterns among a revenue cycle department’s AR data. Machine learning helps illuminate these patterns and trends that individuals reviewing the data on their own might not be able to identify.

How does this work? Consider the following example of the steps a machine learning model might take to identify higher-value accounts for a typical healthcare revenue cycle department.

  1. A machine learning model gathers all the daily transaction information in the organization’s patient accounting system.

  2. The model uses that data to calculate the amount of effort it takes for staff to work and resolve an account. To do so, it assigns the account a cash value, determining the total dollars that will be generated by the account, and calculates the amount of time needed to bring the account to a zero balance.

  3. The model continues reviewing data in real time using daily feeds of transactional data and recalculates these values after every new transaction. As new information flows into the model, the revenue cycle team can adapt quickly and use updated predictions to determine new work queue priorities. Lower-value accounts can be addressed through automation tools while staff members can focus their efforts on higher-value accounts.

  4. The team can take the information derived from the machine learning model and integrate it into the revenue cycle team’s workflows. The real-time, revenue cycle intelligence generated from machine learning can create a long-lasting, positive effect on the team’s workflows moving forward.

Keeping resources focused on high-value activities

As revenue cycle leaders traverse today’s financially challenging environment, they can benefit from directing their resources toward areas of high opportunity – and away from areas of low opportunity. Applying data-driven automation solutions to account resolution processes can route and resolve accounts in a cheaper, faster manner, ultimately freeing up staff to be able to work on more value-added activities.

The model described is just one example of how organizations can use the valuable AR activity data that already exists within their organizations to identify opportunities for revenue cycle automation. To get started using data-driven automation to help minimize costs and maximize cash to net revenue as part of their overall revenue cycle goals, leaders should consider working with third-party automation specialists to learn more.

 

1 According to data from Crowe client engagements.

Contact us

Ken Ruiz
Ken Ruiz
Principal, Healthcare Consulting
Ryan Hartman
Ryan Hartman
Matt Szaflarski
Matt Szaflarski