Making the right decisions around health care AI: Insights for driving results

Make the right decisions about AI: ©Aliaksandr Marko - stock.adobe.com

There’s a willingness to adopt artificial intelligence (AI) in health care unlike anything seen in recent years, driven largely by the enormous strain on health care resources. But driving optimal value from AI investments depends on the ability to translate AI into quantifiable results that strengthen efficiency, quality of care and the bottom line.

Moody’s Investors Service recently highlighted seven areas where AI is poised to make an impact in health care in 2024. While AI-powered clinical documentation and care decision-making support made the list, so did areas like natural language processing (NLP) to improve patient access or AI for prior authorization and billing and collections.

Key to evaluating AI investments for value: balancing the promise of AI with the need for practical return on investment.

Low-risk, high-reward AI investment

Too often, health leaders have been burned by the hype of new technology—and AI is no exception. Although two out of three health care leaders believe AI could reduce the cost of care delivery, one out of four believe the pace of adoption is occurring too rapidly, one recent survey indicates.

And while AI and machine learning are a top focus for one out of three hospital and health system CIOs, “a heavy operational burden is placed on IT to implement, support and optimize [AI and other] emerging tools,” an August 2023 survey found.

Just as there is pressure to leverage AI in ways that set their organizations apart from the fact, health leaders also are mindful of the consequences of making the wrong decisions. Implementing AI-powered tools could take pressure off overloaded clinicians; however, they lose their value when the pressure to optimize and maintain these tools falls to IT teams that lack the manpower or knowledge base to take this role on.

That’s why a deliberate approach to AI implementation is essential. Here are three considerations for assessing AI tools for investment.

Prioritize AI tools with the potential for immediate impact. What are the pain points your organization faces, and where could an AI-powered solution be helpful right now? Once you identify these pain points, start with low-effort opportunities that could have a big impact for your organization.

For instance, one emerging area of importance lies in connecting with health care’s “late adopters”—organizations that weren’t eligible for EHR implementation incentives, like post-acute facilities, and struggle to exchange and receive information with hospitals and health systems. No matter where your organization lies in the health care ecosystem, at some point, you will need to share data with one of health care’s digital have-nots, whether to facilitate a patient admission or a referral. This requires a pragmatic solution for extracting and exchanging data.

This is an area where AI and NLP can play a transformative role in a cost-effective way. For example, by combining a commonly used tool—digital fax—with AI and NLP software, organizations can extract the information they need from faxes or scans, even with handwritten notes. At a time when 80% of health care data is unstructured, this type of technology gives health systems and digital have-nots the power to make critical connections that improve health by moving patient care information to the people who need it most, faster. It’s also a highly affordable approach—and one that doesn’t require organizations to adopt new technology that could potentially disrupt their established workflow.

Look for AI solutions that are easy to implement. Even as prominent EHR platforms look for ways to incorporate AI, including generative AI, into their offerings, IT talent strain remains a barrier to AI optimization and support, the health care CIO survey found. As a result, ease of implementation should drive AI investment decisions. Prioritize solutions that meet your organization where it is today, rather than requiring additional training or support from staff. Talk with other health leaders to assess the level of lift required to adopt a new solution. Just as important: Ask how staff have responded to the new tool and whether it meets their expectations.

Emphasize low risk with a solid return. Eighty-five percent of health system leaders believe AI’s ability to automate administrative tasks will be a top driver of value, according to the CIO survey, but determining where to focus for optimal return remains a challenge. One low-risk, high-return area involves the use of AI and NLP to address challenges related to document processing.

For instance, AI and NLP can recognize a faxed or scanned document and quickly extract relevant data points irrespective of format—even when the data is handwritten—regardless of where the data appears in the document. From there, AI and NLP can structure the data and use pre-existing interfaces and processes to reduce the need for manual intervention and accelerate processing. This speeds referrals to specialists and access to treatment. It also provides a simple ROI for providers and patients.

Closing thoughts

AI continues to generate tremendous enthusiasm in health care, but it also comes with a healthy dose of skepticism, especially when it comes to generating ROI. It is a tool that can be helpful in many respects. However, it should not be implemented for the sake of implementing AI. Taking a value-driven approach to AI adoption involves focusing on where AI can most effectively address pain points without a great deal of effort or expense. By concentrating on low-lift, high-impact opportunities—such as by pairing AI and NLP with existing technologies—health care leaders can get the most bang for their investment while improving quality of care and driving efficiency as well as satisfaction among patients and staff.

Johnny Hecker is chief revenue officer and executive vice president, operations, Consensus Cloud Solutions.