AI health

US hospitals and the health care industry are still struggling with the aftereffects of the pandemic even as many people have put the COVID-19 virus and its resulting damage in the review mirror.

The pandemic and other challenges – including worker shortages and inflation – are conspiring to hobble the financial picture of many health care facilities, according to global management consulting firm Bain and Company. In fact, more than half of the hospitals in the country exited 2022 with a negative margin, making it the most difficult financial year since the pandemic started in 2020.

According to a new report from Bain, 60% of health system executives surveyed said rising costs were their primary concern.

“Payers and providers are now on the hunt for margin improvements,” Bain partners Eric Berger and Margaret Dries wrote in the report. “In our experience, the most successful companies won’t merely reduce costs, but also ramp up productivity. When done right, modest technology investments can accomplish both.”

AI technology can help. Berger and Dries said the cost of training an AI system has dropped 1,000-fold since 2017 and emerging segments like large-language models (LLMs) and generative AI – which thrust itself into the public consciousness with chatbots like ChatGPT and Google’s Bard – are AI tools that can boost productivity for a relatively low cost.

Lot of Hope, But Little Planning

The problem? While 75% of health system executives believe generative AI can help remake the industry, only 6% have a generative AI strategy in place.

“Many recognize the potential AI offers to boost productivity, yet they are acutely aware of the uncertainties around evolving technology,” Berger said in a statement. “This uncertainty cuts both ways. While there is hype, there is also opportunity.”

He told Techstrong.ai that “the reluctance [to take the first steps into the AI pool] is driven by a range of factors,” which, as listed in the report, include the rapidly expanding options of generative AI applications and the worry about overinvesting in the wrong technologies and underinvesting in the right ones.

There’s also the newness of the technologies, the lack of internal expertise, and budgets that already are tight.

“A wait-and-see approach is a tempting prospect,” Berger and Dries wrote.

Questions and Promise

Bob O’Donnell, principal analyst with TECHnalysis Research, told Techstrong.ai that the temptation is understandable. According to a recent TECHnalysis report about AI in the enterprise, there are ongoing concerns about the technology, from security and data protection to tool immaturity, inaccuracies and ethics.

Corporations also have a lot of questions about AI but little expertise.

“It’s a super-confusing time for all industries, health care included,” O’Donnell said. “There’s absolutely a lack of expertise. There’s also a huge amount of things that are different this time around.”

That includes the nature of generative AI itself. It’s both an engine – like ChatGPT, Bard, and Meta’s Llama-2 – and an application, such as data search, content creation and code development for other applications. It includes both foundation models and AI apps.

That said, 88% of companies are already using generative AI, though – mirroring what Bain found in the healthcare field – only 7% have a formal policy in place, according to TECHnalysis’ study. In addition, 95% of survey respondents believe generative AI will profoundly impact work.

That 95% is driving the 88% and likely will help increase the 7%. Companies believe AI can help them and know that their competitors are jumping into it. They don’t want to get left behind and at a disadvantage. In the health care field (which for the study was combined with education), the percentage of organizations that aren’t using AI at all is only slightly higher than in other industries, O’Donnell said.

Start Small, Then Grow

In their report, the Bain partners said health care companies should start using generative AI focused, low-risk use cases that improve productivity and cost efficiency. Margins would improve over three to nine months, and they would understand how to implement a generative AI strategy while growing the needed funds and experience.

“Despite the uncertainty, generative AI already has the power to alleviate some of providers’ biggest woes, which include rising costs and high inflation, clinician shortages and physician burnout,” they wrote.

Health system executives surveyed for the report see improving clinical documentation, structuring and analyzing patient data and automation workflows as near-term benefits. Long-term goals include predictive analytics and risk stratification, clinical decision support tools, and diagnostic and treatment recommendations.

Bain noted a number of areas where AI tools are making a difference, from generating documents to improving call center responses with conversational AI to integrating ChatGPT with electronic health records to create response messages to patients.

Principles to Follow

That said, there are challenges, according to executives, including budget restraints, the lack of expertise, and legal and regulatory considerations.

“Health care has some unique considerations, which include data security, privacy, being highly regulated, [and] patient care,” Berger said.

Organizations should embrace guiding principles as they move forward with generative AI, according to Bain. These include first piloting low-risk applications with a narrow focus, deciding whether to build or partner on the needed technology, and funneling the cost savings and experience gained through earlier use cases into more transformative projects later on.

Lastly, they should remember that generative AI isn’t a strategy.

“To build a true competitive advantage, top CEOs and CFOs are selective and discerning, ensuring that every generative AI initiative reinforces and enables their overarching goals,” the partners wrote.