As healthcare trends hit their highest level in two decades and annual domestic spending on healthcare services exceeds $5 trillion, some have suggested that artificial intelligence (AI) could lead to efficiencies and eventual cost reductions. In the commercial market, every new vendor is touting their use of AI and suggesting large returns on investment. For employers, who pay the most for commercial plans, the question isn't whether AI will change healthcare delivery, but whether it will significantly lower benefits spend. There are good reasons to believe that AI will transform healthcare delivery without substantially lowering total medical costs.
While many new technologies have promised to lower medical costs, they have generally fallen short. For example, electronic health records (EHRs) were thought to be revolutionary, with optimistic estimates they would lead to $78 billion in annual savings. Those savings never materialized, and EHRs often led to higher costs as they facilitated more intensive coding. And even if providers can provide care with fewer labor inputs, these savings may well not be transmitted to the healthcare purchaser.
There's no question that AI can transform healthcare delivery. AI will likely lead to improved diagnostics, quicker development of new pharmaceuticals and more insights from medical images. But this might not lead to lower costs. For example, if high-value drugs can be brought to market more quickly, these drugs will be priced based on their value, not based on the cost of development. New drugs cost a lot because they are protected by patents. So, speeding up the pipeline for new medications is likely to lead to more clinical benefit but higher costs.
AI could also encourage patients to receive care from providers offering lower costs or higher quality. An AI healthbot could help patients find the best provider based on current quality and cost information from many different sources. Still, this will require patient adoption. Right now, patients continue to seek care from their existing providers, and AI navigation systems will need to build up trust over time. When members do rely on chatbots, it often is for transactional questions, leaving less opportunity for steerage to optimal care. Further, the top 10% of providers don’t have the capacity to take care of everyone.
AI could also lead to higher utilization. For example, expanded predictive capabilities could lead to early detection of cancer, which would often not lower costs. All recommended cancer screenings are currently cost-effective (they cost less than $100,000 per quality-adjusted life year saved), but they aren't cost-saving.
Ultimately, AI investments in healthcare will lead to real changes in how care is delivered and could lead to meaningful improvements to the quality of care. AI could help empower patients and improve the convenience of care. And it can help employers drive internal efficiency and improve program oversight. But AI is less likely to lead to lower costs of care. Its introduction may very well follow the path of EHRs, which fundamentally changed healthcare delivery but didn't reduce total medical expenditures.
Enhancements to parental, bereavement and caregiver leave continue to rise
Our video series boosts employees' mental health knowledge, equipping them to support themselves, their loved ones and foster a supportive workplace community.
