The employ of artificial intelligence to improve revenue cycle management in healthcare shows promise, but executives have concerns about the technology’s accuracy and reliability.
Here are some of the results from the Inovalon study questionnaire Of the more than 400 revenue cycle and finance executives and managers surveyed, 84% said they were positive about implementing AI-powered RCM in hospitals.
However, a third of respondents said they had concerns or were sceptical about using AI in RCM, with the main sticking points being concerns about accuracy and reliability (31%), lack of familiarity/understanding (17%) and AI being too recent/untested (15%).
People are better than AI
Twenty percent of respondents said they believed human performance – at least at this stage – was superior to AI performance.
Julie Lambert, president and CEO of vendor Inovalon, said this applies across RCM, but there are certainly areas that could benefit more from AI.
“Rather than thinking about it as an either/or scenario, I’d like to challenge us to think about it more, because expertise is a key foundation for building AI/ML models that work and are continually improved,” she said. “When technology and expertise are combined, the potential for superior outcomes is there.”
From her perspective, the areas where AI can have the greatest impact on RCM are those that pose the most challenges and are the most challenging for today’s service providers.
Among those areas of denial, prior authorization and eligibility probably rank highest among all providers. She added that it’s no coincidence that all of these elements are related.
“Mistakes at the beginning of the registration process cause denials at the end,” Lambert said.
What causes denials?
Knowing which scenarios lead to denials and how to detect or predict those denials before they happen is a great opportunity to leverage AI expertise and billing performance data to build and train a model.
“Both in the processes themselves and in overall connectivity, there are opportunities to use machine learning and artificial intelligence to improve the work of suppliers,” she added.
Lambert added that an essential factor to consider is that AI is not unchanging and should never be treated that way.
“A fundamental principle of AI is to design a model that is constantly learning – models will constantly learn from data and the feedback loop that naturally emerges from the results,” she said.
External factors
It is also essential that external factors that may impact the model are known and accounted for. This could mean regulatory changes that affect the structure of the data, data elements in the responses, or other factors that may cause anomalies in the data.
“Make sure there is awareness of any changes affecting the model so that interpretation of results does not rely on false assumptions or correlations,” advises Lambert.
She added that it’s essential to make sure people understand that AI isn’t just for senior executives or just for data scientists – everyone should be involved in it, and that’s what will make AI successful.
“AI needs input from people on the ground, managing data, performing operations and controlling workflows to help build models,” she added.