AI can improve the revenue cycle but increases risk, Moody’s says

Artificial intelligence has the potential to improve revenue cycle operations, optimize labor through administrative efficiency and lower operating costs, but it comes with the risk of cybersecurity attacks and patient data breaches, according to Moody’s Investors Service.

AI has the potential to improve health systems’ bottom lines by bringing accounts receivable down and minimizing write-offs. Accounts receivables days at health systems rose to a median 48 days in fiscal 2022 from 46 days in 2020, according to fiscal 2022 median data, Moody’s said. 

AI can summarize data quickly to free up clinician time, resulting in cost savings through reduced personnel expenses, Moody’s said. It also enables clinicians to see more patients and focus on more important tasks, which will help prevent physician burnout. 

The overall credit impact of AI on the industry will depend on the extent to which healthcare systems can also manage the risk. The amount of data processed with AI algorithms and technology is evolving rapidly, Moody’s said during a webinar held Tuesday.

In addition, hospitals will need to meet heightened regulatory and compliance challenges that will arise from AI use. The U.S. Food and Drug Administration recently announced the establishment of a Digital Health Advisory Committee to provide advice on scientific and technical issues related to digital health technologies, including AI and machine learning, digital therapeutics, wearables and remote patient monitoring. 

These risks will increase investment in IT, or may shift spending for other IT priorities, according to Raj Joshi, vice president, senior credit officer for Corporate Finance, Moody’s.

“We know that GenAI investments will be large,” Joshi said. “We think some of the GenAI investments will come from other priorities that will impact traditional IT spend, such as servers and storage. It could potentially push out upgrades.”

WHY THIS MATTERS

AI also has the potential to change the industry’s competitive landscape, Moody’s said. 

But unlike generative AI, most AI models are not plug-and-play, and their integration into business models are expected to take several years.

Overall, AI brings potential benefits to hospitals in the form of clinical documents, decision support, prescriptive intelligence, appointment management, billing and collections, health monitoring via wearables, and natural language processing.

There will be an expected race for hospitals to keep up with AI investments.

Successful implementation of AI could widen the gap between the best and worst performers by disrupting revenue generation for those that do not adapt to the new technology. It could also empower startups to enter established markets and take away market share. 

While it offers opportunities, use of AI in a clinical setting still requires significant human oversight, since AI products are still developing and sometimes produce incorrect responses, Moody’s said. 

Bias in the data used by AI that reflects racial, gender or socioeconomic discrimination can result in AI models perpetuating or amplifying that discrimination, which could present legal and healthcare equity risks, Moody’s said. 

Incorporating AI technologies into healthcare systems can also create an outsized reliance on particular vendors and necessitate substantial investment. Given that AI technology is still in a nascent stage, some healthcare systems will overinvest in opportunities that do not fulfill expectations for them.

Certain AI could become obsolete quickly, after a hospital has invested a large amount of money.

THE LARGER TREND

Moody’s on Tuesday predicted the credit quality future of artificial intelligence in a webinar entitled, “AI and credit quality: Opportunities and challenges for non-financial corporate issuers.”

At a high level, the overall credit impact is positive. 

In the healthcare sector, AI is expected to have little credit quality impact over the next two years, and then have a low impact from 2026 to 2030, said Francesco Bozzano, vice president and senior analyst for the corporate finance group at Moody’s. 

The same is true for telecommunications, Moody’s said.

AI is expected to have a greater effect and positive impact on data-rich sectors such as the software and semiconductor industry.

Execution needs dedicated teams. Small firms lag behind in the use of big data, and AI and will suffer from lower scale, Moody’s said.

Rapid adoption of AI is underway, Joshi said. According to an IDC survey, 21% of the respondents are already making significant investments in Generative AI, and about 60% of companies are still in the exploration stage.

Moody’s examples of AI in use at health systems include:

• BayCare Health System announced a partnership earlier this year to implement technology that could reduce nurses’ documentation burden.
• Mercy Health in Missouri and Microsoft Corp. recently announced a collaboration that includes implementing generative AI-assisted communication to help patients better understand lab results. It will also help schedule patient follow-up appointments and provide a chatbot that Mercy employees can use to find answers to human resources questions. 
• Mass General Brigham is collaborating with GE Healthcare on an AI algorithm to predict missed care.
• NYU Langone is using homegrown Large Language Model AI to help reduce hospital readmissions and hospital-acquired conditions (HACs) through analyzing patterns in patient data. 
• HCA has developed an algorithm that collects and analyzes clinical data such as patient location, vital signs, pharmacy and laboratory information to signal caregivers in real time to initiate early sepsis care. In partnership with technology company Augmedix Inc., HCA is also piloting AI-enabled medical dictation software designed to convert clinician-patient conversations into medical notes that physicians and nurses can review before they are transferred to the electronic health record system. 

Twitter: @SusanJMorse
Email the writer: SMorse@himss.org