Kevin Vigilante M.D., MPH. Chief Medical Officer at Booz Allen Hamilton, where he also leads the Health Futures Group.
When I was a medical student in 1983, I wrote my senior thesis on healthcare costs. They had been steadily rising for over 20 years and, at about $355 billion or 11% of GDP, seemed out of control. Forty years later, despite multiple interventions—from managed competition to value-based care—healthcare costs continue to climb. In 2023, the U.S. spent almost $4.9 trillion—about 18% of the GDP—on healthcare, making the costs I previously fretted over seem quaint. And although other OECD countries spend less than we do, their cost curves look similar.
It seems like healthcare has an incurable “cost disease.”
Economist William Baumol saw it exactly this way. Studying the economics of the arts, he observed that musicians in the 1960s weren’t any more productive than musicians in the 1860s, yet because they required higher pay to survive, ticket prices had become very expensive. This contrasted with manufacturing, where increased productivity enabled higher wages but lower prices. The arts, like other labor-intensive industries less conducive to automation, had to raise prices to accommodate wage increases. They were afflicted with an endemic cost disease. Consequently, while the costs of clothes, cars and computers continually decline, the costs of healthcare, college and haircuts relentlessly rise.
There are multiple factors exacerbating the productivity-driven cost disease in healthcare. These include fee-for-service models incentivizing unnecessary care, third-party reimbursement shielding transactions from market forces, an aging population, a chronic disease epidemic and substantial waste from medical errors, redundant care and administrative costs. These comorbidities make the healthcare cost syndrome formidable and hard to cure.
A particularly acute factor is drug costs, the fastest-growing healthcare cost ailment. One of the most expensive components of drug development is clinical trials. On average, it costs $2.6 billion and takes 10 to 15 years to develop one drug, with 90% of drugs tested in phase 1 trials failing to reach pharmacy shelves. The pharmaceutical industry has its own version of the cost disease.
AI Therapies For The Cost Disease
There’s a plausible hypothesis that AI will help treat the cost disease. Healthcare is among the most cognitively intensive industries, and labor represents the most substantial cost—about 60% of hospital costs. We need tools to enhance cognitive productivity, enabling humans to focus on tasks only humans can perform. AI will transform how intellectually intensive labor is executed.
Labor in healthcare includes taking down medical histories, making diagnoses, lifting patients, removing lumps, palpating livers, writing notes, processing claims and comforting patients. In some tasks, AI will boost productivity and efficiency; in others, it won’t. There’s a productivity ceiling in healthcare we may not surpass, but it could be higher than previously thought.
Low-hanging productivity fruit lies in administrative tasks, such as automating coding, claims processing, prior authorizations, formulary requests and low-risk tasks clinicians perform daily. Ambient listening products are already capturing doctor-patient conversations and automatically writing properly formatted notes that need little revision. This has reduced pajama time—after-hours note writing—enabling physicians to see more patients while enhancing the experience.
Soon, these programs will recommend questions doctors forgot to ask, suggest diagnoses and therapies and identify medication errors in real time. They’re expected to reduce diagnostic and therapeutic errors that inflict harm and generate substantial costs. AI assistants may also make generalists more confident in their diagnoses, dissuading costly specialty consultations. In the future, robots may help nurses dispense medication and assist aides in lifting and moving patients.
AI won’t replace human practitioners; it will increase their productivity and make their work more rewarding. It will reduce low-value work and give providers more time to have meaningful connections with patients. Both patients and practitioners will have a better experience.
AI can also lower drug costs. AI and machine learning can accelerate the identification of drug targets, speed drug design through digital simulation and use high-throughput screening to test promising compounds. AI can also make clinical trials more efficient by automating patient identification and matching patients to trials. Once a trial begins, AI tools can monitor data in real time to identify adverse effects or predict outcomes, supporting decisions to continue, adjust or halt a trial, all of which could reduce drug development timelines and costs.
Navigating The Future
In 1990, healthcare expenditures accounted for about 14% of federal outlays (as shown in the Excel file); today, it’s about 30%. Healthcare is a substantial driver of the $36 trillion national debt that’s about 122% of the GDP. The healthcare cost disease has metastasized and become the government’s cost disease. The need for treatment has become urgent.
AI may not cure the cost disease, but it can treat the ailment and dampen its effects. Previous interventions heralded as “cost-saving” have barely dented the cost curve. However, AI is different. Current AI tools, especially large language models and AI agents, can meaningfully augment human cognition, improve productivity and reduce labor costs. AI represents an inflection point in the history of intellectual labor and medical work.
AI will be to cognitive labor what the steam engine was to physical labor.
However, to deliver on its promise, AI must be developed and nurtured. Too much regulation will stifle startups, hinder innovation and restrain growth. The most stringent oversight should be reserved for products directly impacting patients.
We’re at the beginning of a long AI journey and a fifth industrial revolution that will likely change how healthcare providers and administrators work. If we get it right, their work will be more rewarding, patients will get better care and we’ll better manage healthcare’s chronic cost disease.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?
