Healthcare spending in the U.S. is rising on an unsustainable trajectory, but artificial intelligence could be the antidote.
The application of AI in drug discovery, hospitals, value-based care and in other areas could create savings of $400 billion to $1.5 trillion, according to Morgan Stanley Research.
“We believe the diffusion of AI through select segments of the U.S. healthcare ecosystem will drive disruption and cost savings that can help to counter rising costs due to demographic trends,” says Terence Flynn, who leads coverage of U.S. Pharma and Biotech at Morgan Stanley Research.
The scope of the issue can’t be overstated. U.S. healthcare spending has been rising for decades, outpacing increases in other developed economies. It reached about 18% of GDP in 2023, compared to an average of 11% for a group of the country’s peers, and the trend is expected to continue as the population ages. Healthcare is projected to be about 20% of the U.S. economy in the early 2030s and, if the linear growth path continues, such spending could equal 25% of GDP in 2050.
The annual cost savings that are needed run into the trillions of dollars to keep healthcare costs to 20% of GDP in 2050 rather than 25% or 30%. If one assumes 2050 U.S. GDP of $46 trillion, policymakers need to find gross annual savings on the order of $2.3 trillion to $4.6 trillion.
To understand the most likely impacts of AI in healthcare, policymakers and investors can look to two primary areas: speed to market in drug development and efficiencies in healthcare facilities.
AI’s Impact on Drug Development
The cost of medicines represents a modest slice of U.S. healthcare spending, around 9% of the total, even if drug costs come up often in policy debates. But medicines can cure disease and keep people from needing visits to the doctor or ending up in the hospital. They can also help people leave the hospital sooner, with academic research showing that newly launched drugs have contributed to an 11% to 16% reduction in hospital stays.
“AI has the potential to improve drug design by more precisely predicting the molecular structures and the interactions between drug candidates and the target,” Flynn explains.
If AI can drive an increase in the number of approved medicines, the resulting savings in hospital care and physician or clinical services will be meaningful. The potential savings in 2050 could be in the range of $100 billion to $600 billion, assuming an increase of 10% to 40% compared to the recent pace of newly approved medicines.
Many biopharma companies, both public and private, are leveraging AI for drug discovery, and the range of views about the potential impact of the technology is broad.
“To judge the potential impact, it’s important to take the long view, for several reasons,” Flynn says. “It will likely take at least five years for AI to start having a significant impact on drug discovery, and research shows that an additional period of time passes before the impact of new drugs can be seen in hospital and clinical care costs.”
Savings on Hospital and Physician Care
AI can improve hospital efficiency by optimizing staffing and patient scheduling, and improving supply chains and drug management, to name a few examples. Hospital operators already are reporting savings in these areas, and more is clearly possible.
AI-related cost savings of 10% to 20% should be achievable for hospitals, the single largest category in U.S. healthcare spending, for cost savings in 2050 in the range of $300 billion to $900 billion.
“It will take years to fully implement AI solutions across these various areas, but when we look out to 2050 and estimate the gross cost savings that could be possible, the numbers are really meaningful,” says Erin Wright, who covers Healthcare Services for Morgan Stanley Research.