Will artificial intelligence (AI) make money or damage hospital finance? Excitement about AI commitments has attracted billions of dollars in investment, but hospitals still have difficulty predicting the value they can see in these technologies in the future.
A recent McKinsey survey found that while about half of health system leaders expect a return on investment (ROI) for AI, only 17% are currently able to measure positive returns. With so many hospitals operating negative or thin margins, they do not have the flexibility to gamble on tools that may or may not support long-term growth.
Historically, hospitals have passed the innovation passively, rather than actively driving it – in cases such as mandatory EHR implementation and consumer-driven digital front doors. To avoid a repeat of history, healthcare leaders need to seize the opportunities AI can offer, including operational efficiency and profitability.
AI ROI Challenge in Hospitals
Historically, new technologies have a lot of records in delivering on their promises – raising health skepticism among hospital executives. Traditional healthcare technology business models improve the bottom line for software companies, but the health system, from radiological imaging to everything patients require billing (they rely on these technologies – and don’t always see the same benefits. As they continue to operate in competitive markets and median margins drop below one percent, productivity and responsible AI investment are crucial to their economic survival and ability to serve patients. AI must work for health systems, but policymakers need to be confident that they make the right choice.
A combination of tight profit margins and the struggle to make ends meet for even the most resource-rich hospitals can force leaders into hasty AI investments. Anxiety, some of the first solutions that seem to improve efficiency, usually don’t have enough ROI plans. Others wait for the magic box to appear and solve their problems. Both can threaten the hospital’s future solvency and ability to provide care, which is also not a strategy.
To make AI work for its hospitals, leaders need a plan – a clear way to measure value, select tools, and scale effective methods.
Start with the benefits
Thinking about ROI should start by appreciating all the potential benefits. Many people only consider AI-assisted automation and cost savings, but that's just part of the picture: AI can also help teams do better – improve accuracy, expand scope and enable smarter decisions.
Take clinical document integrity (CDI) as an example. Even the most experienced CDI teams usually only capture the most common 40% of missed codes, such as sepsis and respiratory failure. But what about long tail – 60% other? AI can help discover less frequent but high-impact diagnostic codes, thereby significantly improving capture and revenue without changing human teams. Bottom line: AI not only spends less time doing the same job. It's also about doing better work – smarter, faster, more comprehensive.
Even if you have huge ambitions, narrow down and start with victory. Hospitals should start using AI for a relatively small number of people – like a department – once they are satisfied with how to choose and operate AI. When the person in charge of AI strategies starts at a very young age and then shows a positive ROI, they earn political capital that can take advantage of further investment.
Align success metrics with the analysis team
Many hospitals lack the internal analytical capabilities to measure ROI on their own, so they rely on AI vendors to achieve ROI. As a major hospital transformational director for overseeing AI II, and in my current position as a clinical AI company, I have seen this challenge in hospitals and table vendors. Before hospitals choose suppliers, they should align with the analytical team and suppliers to attribute value and the effectiveness of success.
This is a crucial step: You don't want poorly conceived metrics to resist your AI investment. If a vendor reports 80% efficiency improvements by measuring costs only, but you still have 100 people completing 20 tasks, you may feel like you aren't actually improving efficiency. Define how to attribute value in advance and be as specific as possible. Make sure your supplier is on board and assumes responsibility.
Help AI vendors help you
Vendors want to know about their optimized Polaris metrics, but they may need help getting there. This is a common problem for hospitals. To address this problem, hospitals and suppliers should step by step to solve the steps they jointly solve the challenges. If a supplier cannot clearly explain how their products provide value in your context, that is a red flag. If possible, give them the data and context of success. ROI is a shared responsibility.
Build the right AI success team
AI success depends not only on technology. It depends on the people who choose, drive and support it. Supported by a strong internal analytics team, only a few highly competent leaders’ AI task groups can help hospital systems make informed bets by working closely with technology partners and internal stakeholders.
You don't need a huge committee. Only a few powerful, curious, and analytical people can make a big difference.
Learn about the future through experience
With confidence in its performance, AI adoption in hospitals is on an exponential curve. What once seemed like a futuristic capability (like human understanding of clinical documentation) is now commercially viable, with powerful large language models. If hospitals don’t start learning from AI now, they are not only likely to fall behind. They may also miss the upside that AI can provide.
AI has proven its value in the healthcare backend. Conducting Revenue Cycle Management: AI can now make a second-level review of each patient chart before billing – the app can increase efficiency while generating a 5:1 ROI. That's not the potential for the future. Today, this is a real expression.
Artificial intelligence does not need to be a black box – hospitals do not need to invest based on blind beliefs. With the right structure, problems and metrics, healthcare leaders can cut the hype and make decisions that really drive value.
In a financial landscape where every investment is calculated, AI can’t just have hope – it has to be productive. For prospective hospitals, it is already.
Image: Woch, Getty Images
Michael Gao, MD, is a physician, data scientist and innovator in healthcare technology. As CEO and co-founder of SmarterDX, he led the development of clinical AI, which helped hospitals recover millions of revenue and optimize the quality of care. Dr. Gao previously led the AI program at Newyork-Presbyterian, served as assistant professor of medicine at Weill Cornell and served as director of transformational health care. He holds degrees from the University of Michigan and completed training at Newyork-Presbyterian/Weill Cornell, where he also completed Silverman Scholarships from Silverman Healthcare Innovation.
This article passed Mixed Influencer program. Anyone can post a view on MedCity News' healthcare business and innovation through MedCity Remacence. Click here to learn how.
