By Josh Budman
SVP, Research & Incubation, Net Health
Recently, a joint report released by McKinsey & Company and Harvard University caught my attention as it attempts to shine light on the influence current data technologies could have in streamlining our country’s ever-expanding healthcare industry.
Specifically, the report estimates that the wider adoption of artificial intelligence (AI) use in healthcare could potentially save the U.S. healthcare system between 5% and 10% each year. In real dollars, that’s around $200 billion to $360 billion in annual spending.1
The report goes on to say that these numbers are attainable within the next five years using AI technologies that already exist. And, they would not occur at the expense of healthcare quality or access.2
Such projections are no doubt welcome news within an industry that’s ripe with inefficiencies – issues that too often impact facility operations, revenue and patient outcomes. And while I agree AI is positioned to play a significant role in the future of healthcare, I’m not quite yet ready to pop the champagne cork.
When it comes to this report’s projections, I am cautiously optimistic that the joint collaboration between vendors and consumers of AI tools can achieve these cost savings.
AI is the Future of Healthcare
Before diving into the reasons for my cautious optimism, I offer this important caveat:
As an enterprise analytics leader and technologist who strives to apply artificial intelligence to benefit healthcare practitioners and their patients, I wholeheartedly agree with the overall premise of this report.
The future of healthcare technology is automation and artificial intelligence. And at the risk of sounding cliché, the future in this case is now.
That’s because – and just as the report says – AI technologies already exist that can make healthcare more efficient, cost-effective and accessible for practitioners and their patients. And of the tools already in use, the wider adoption of these technologies would go a long way toward saving money within the field.
Administrative tools that streamline notetaking, documentation (i.e., Net Health’s own Tissue Analytics), claims and billing, scheduling (i.e., our Missed Visit Prediction Indicator), compliance-driven automation, and so on would, if used in greater numbers, net positive gains.
This report, however, considers all AI-based technologies – those currently in use as well as those still being perfected and applied to healthcare. It’s here where I question whether regulatory processes and market factors are equipped to handle the AI surge that’s coming.
Will AI Propel Healthcare to These Milestones in Five Years?
Based on the highly regulated nature of healthcare, it will take a long time for the FDA to approve, and the market to adopt, the great wave of AI technologies currently being developed.
So, while I’m not surprised by the potential for tools under the AI and machine learning umbrella have for improving healthcare, I predict it will take longer than five years for the industry to reach the financial savings outlined in the report.
Reaching those numbers is indeed possible. The arc toward reaching them, however, will likely be long and complex. Here’s why:
The Food and Drug Administration (FDA) plays a significant role in the regulation and use of AI in healthcare. The agency evaluates and approves AI medical devices to ensure they are safe and effective for use in the market, and that process takes time. Often, it takes a lot of time.
With an increasing number of AI technologies to evaluate and finite resources with which to do their jobs, the FDA certainly has its work cut out for them when reviewing AI-driven medical products. This means developers could be in for a relatively long way before being considered.
Implementation and Monetization
The actionable implementation of AI-based devices and tools is going to be a challenge, as well.
To best capitalize on the AI-based gains in healthcare, as outlined in the report, pricing models will have to match the supposed value-based trajectory of the market. In other words, for-profit companies involved in rolling out tools and devices will need to show that there are cost savings in using their technology by sharing in some of the risks.
In my experience, it’s difficult to achieve the high adoption of new technology when you expect healthcare systems to sign a license agreement before they fully buy into its potential.
The Willingness of EHRs to Integrate
When it comes to administrative and some clinical AI tools, the willingness of larger EHR companies to integrate with third-party solutions will help optimize the use of these tools as well as the viability of their developers.
While some EHRs, like Net Health, are accomplished in developing useful AI tools, others will need to remain open-minded when it comes to developing integrations with companies in the AI space.
The Impact of AI in the Medical Field
Despite these challenges, AI will indeed have an across-the-board effect on both the administrative and clinical aspects of healthcare. And, that effect will continue to grow as AI proves itself to be more reliable and accessible throughout all areas of the field.
One area of healthcare that I feel will be particularly impacted by AI in the short term – a segment that offers the quickest path toward the application of AI within their day-to-day operations – includes physician practice groups.
AI-based solutions exist today in which practices can adopt for the improvement of operational and administrative tasks. And because sales cycles are typically shorter in these groups, they can generally adopt these and other new tools and technologies more quickly.
In addition, the application of AI tools in this setting is typically a lower-risk proposition. Since most current AI tools for practices involve documentation automation, claims, billing, compliance, etc., the FDA tends to get less involved in their applications.
In general, however, I am thrilled to see more people developing a general interest in artificial technologies, whether in the healthcare field or through more mainstream applications such as ChatGPT.
Such interest evolves into understanding, and from there develops a greater amount of trust and possibility in the ways AI can positively impact healthcare efficiencies and outcomes.
In five years, will this result in a savings of 10% in U.S. healthcare spending? That remains to be seen. But, I certainly like the direction we’re going.
1 2 National Business of Economic Research, “The Potential Impact of Artificial Intelligence on Healthcare Spending,” January 2023.