4 strategies for an AI-driven approach to improve revenue cycle performance

Adopting machine learning and artificial intelligence (AI) to the revenue cycle can enable healthcare organizations to get ahead of the curve on transformational innovation.

Both artificial intelligence (AI) and automation hold strong potential to drive improved revenue cycle performance, and healthcare leaders seem to agree that AI should be a priority. A recent survey indicates:

• 75% of healthcare leaders are actively implementing or planning to execute an AI strategy.
• 43% say their first area of focus will be automating business processes, such as revenue cycle management functions, to reduce costs.^[1]

For example, automating eligibility and benefits verification —  typically
the first administrative transaction in a patient encounter — could save providers $6.52 per transaction, or just over $4 billion  per year.^[2] Meanwhile, machine learning tools can prioritize work based on the likelihood
of payment or age of the account while reducing the risk for error.

Consider these 4 strategies

With use cases for healthcare AI still emerging, healthcare organizations should consider the following four strategies for applying AI and automation to revenue cycle management.

1. Replace highly manual tasks with automated processes and redirect staff efforts toward activities that provide the greatest value. Replacing “time-intensive, error-prone, manual processes” for claims management and payment reconciliation is an implementation 92% of hospitals are planning.^[3] But there is still plenty of room for improvement: Providers could save $9.8 billion by automating key revenue cycle functions.^[4] Automating claims status inquiries alone could save $9.22 per transaction, or more than $2.6 billion.

Front-end revenue cycle processes are ripe for automation. For example, real-time patient eligibility checks ensure providers have the most current information regarding patient coverage and the portion of the deductible met to date. Benefit information that is automatically pulled from the system to create patient charts helps to eliminate manual data entry — significantly reducing the risk for error. One study shows registration and eligibility errors account for 23.9% of denials.^[5]

The time savings gained from automating these tasks accelerates the patient check-in process and allows staff to spend more time on value-added activities, such as patient financial counseling.

2. Use machine learning to stop denials before they start. Each year, 9% of claims are initially denied by payers.^[6] For the average hospital, this statistic means nearly $5 million in payments are at risk of being denied each year. And while 63% of denials can be recovered, it costs about $118 per claim in administrative costs to capture the monies owed.

Machine learning positions providers to predict which claims will be denied before they are submitted. It does so by:

• Identifying the root causes of denials by payer and CPT code
• Applying this intelligence during automated reviews of claims
• Flagging areas where missing or incorrect information appears, such as missing charges or an incorrect patient identifier
• Prompting staff to follow up

This approach enables staff to correct claims prior to submission, increasing clean claim rates. It also helps revenue cycle teams more effectively manage work related to denials by focusing staff attention on high-value denials, as well as those that have a strong chance of being overturned.

3. Analyze patient demographic data to predict the right billing approach for each patient. AI can help revenue cycle teams develop highly targeted collection strategies based on a patient’s previous payment behavior, demographic data, communication preferences and preferred payment methods. It can also point to the type of messaging most likely to engage individual patients and the optimal date and time to send communications. AI can even detect early warnings that a patient may default on a hospital payment plan — and alert patient financial services to signs of trouble. Given that half of patients lack confidence in their ability to pay their medical bills, AI-enabled approaches to patient financial communications and collections could be a critical support in the years ahead.^[7]

One aspect of this approach, propensity-to-pay scoring, uses predictive analytics to determine the likelihood that patients will pay their out-of-pocket costs for care. But while propensity-to-pay solutions have been available for some time, according to one survey, just 14% of healthcare organizations use advanced modeling tools to segment accounts and predict propensity to pay.^[8] Fewer than one in four providers use a data source or external partner to support their efforts. To make a deeper impact in collection rates, more emphasis on data-driven intelligence and processes is needed.

4. Predict when payers will remit payment. With predictive analytics, providers can review payer-specific payment behavior by CPT code to determine how long it will take for a specific claim to be paid and even the day and time the payment will arrive. It’s an approach that can predict the date of remittance for claims with a high degree of accuracy.

The use of AI in healthcare is expected to grow 50.2% from 2018 to 2023, with hospitals and health systems expected to be the biggest adopters.^[9] As the push to bend the healthcarecost curve intensifies, leaders should carefully examine the business case for AI-driven revenuecycle management and explore small-scale innovations with the potential for strong return. Dipping a toe into the waters now will better position healthcare organizations to keep pace with AI advancements while strengthening financial performance. 

Footnotes

^[1]  OptumIQ Annual Survey on AI in Health Care, Optum, September 2018.

^[2] CAQH,  2018 CAQH Index, 2019.

^[3]  “Providers Driven to Implement Patient-Centric Financial Solutions as Consumer Payment Responsibility Skyrockets 29%, Black Book Survey,” PR Newswire, Oct. 24, 2017.

^[4]  CAQH, 2018 CAQH Index, 2019.

^[5]  “Report: Front-end revenue cycle processes leading cause of denials,” Revenue Cycle Advisor, Oct. 25, 2017.

^[6]  Barkholz, D., “Insurance claim denials cost hospitals $262 billion annually,” Modern Healthcare, June 27, 2017.

^[7]  Collins, S.R., Gunja, M.Z., Doty, M.M., and Bhupal, H.K., “Americans’ confidence in their ability to pay for health care is falling,” To the Point, The Commonwealth Fund, May 10, 2018.

^[8]  “HFMA and Navigant , The Future of the Revenue Cycle: A survey of provider executives about the next generation of revenue cycle management, June 18, 2017.

^[9]    MarketsandMarkets, “Artificial Intelligence in Healthcare Market by Offering, Technology, End-Use Application, End User and Geography – Global Forecast to 2025,” December 2018.