US Health Spending Is Entering a Decade Defined by Consolidation, Intensity, and AI
Health care spending in the United States is not moving along a single track. It is being pulled by corporate consolidation, by the intensity with which care is delivered, and by the expanding use of artificial intelligence across clinical and administrative work. A JAMA Perspective argues that these forces will determine not only how much the country spends, but also who ultimately bears those costs and whether higher spending translates into real health gains.
The central problem is familiar to every clinician and health system leader: more spending does not automatically mean better care. In practice, the relationship between cost and outcome is often unstable. Some expenditures finance truly beneficial interventions, while others reflect fragmented delivery, duplicated services, market power, or a clinical culture that favors more testing and more treatment without clear evidence of incremental benefit. The next decade will likely be shaped by how well the system can separate those categories.
Consolidation sits at the center of that tension. When hospitals, physician groups, insurers, or other health care organizations become more concentrated, market leverage tends to rise. Prices can shift upward even when the underlying clinical service has not become more effective. For patients, that can mean higher premiums, higher out-of-pocket costs, and narrower room to navigate between competing delivery models. For employers and public programs, it can mean a larger share of national income being diverted into medical spending rather than wages, investment, or other social needs.
One sentence captures the danger plainly: market concentration can increase spending without improving health.
Intensity of care adds a different pressure. In many settings, clinicians face a system that rewards volume, speed, and procedural throughput more reliably than prevention, continuity, or careful watchful waiting. High-intensity delivery can be lifesaving when it matches disease severity, but it can also drive unnecessary imaging, interventions, admissions, and follow-up care. The biologic appeal of doing more is easy to understand at the bedside. The economic consequence is equally clear at scale: once a service is normalized, it becomes difficult to reverse even when the marginal value is low.
This is where the Perspective’s broader message becomes especially relevant. Health spending is not only a budgetary issue; it is a signal of how the system allocates attention. Intensive care patterns can emerge from clinical uncertainty, defensive medicine, fragmented referral pathways, or financial incentives that reward activity more than restraint. Over time, those patterns can harden into norms. Changing them requires more than exhortation. It requires measurement, payment reform, and clinical discipline that preserves benefit while reducing waste.
The role of artificial intelligence complicates the picture further. AI could lower some administrative and clinical costs by automating repetitive tasks, supporting triage, improving documentation, and helping clinicians identify risk earlier. It could also create new expense categories if adoption is layered on top of existing workflows without replacing older inefficiencies. In other words, AI may reduce friction, but it may also become one more cost center if its deployment is additive rather than substitutive.
A second sentence belongs here, because the risk is easy to miss: digital efficiency does not guarantee financial restraint.
The financial effects of AI will depend on where it is inserted into care. In administrative systems, it may accelerate authorization, coding, scheduling, and claims processing. In clinical practice, it may affect diagnostic support, note generation, surveillance, and patient communication. Each of those uses carries a different spending profile. If the technology reduces labor and administrative overhead, savings may accumulate. If it encourages more testing, more follow-up, or more low-yield alerts, overall spending could rise despite the appearance of modernization.
The Perspective also points toward a more fundamental policy question: who bears the costs when health spending rises? In the United States, the burden is rarely distributed evenly. Patients may face deductibles, co-insurance, and surprise bills. Employers may absorb higher premiums. Public payers may shoulder greater budget pressure. Providers may face reimbursement constraints that influence staffing and service availability. When spending rises because care is more intensive or because markets are more concentrated, the cost often shifts rather than disappears. That redistribution matters because it shapes access, affordability, and trust in the system.
Biology rarely provides a neat answer to cost problems, but clinical behavior sometimes does. Better prevention, earlier intervention, and more selective use of services can avert downstream expense while improving outcomes. Yet those gains are often slow, diffuse, and harder to measure than the immediate revenue generated by a test or procedure. Health systems therefore face a structural imbalance: the most valuable care is not always the most visible, and the most expensive care is not always the most effective. The next decade will reward organizations that can make that distinction operational, not merely rhetorical.
The article’s larger warning is that spending trajectories will not be shaped by a single innovation or policy lever. Consolidation can raise prices. Intensity can inflate utilization. AI can either reduce waste or accelerate it, depending on implementation. Each force will interact with payer design, regulatory oversight, and local market structure. None of those interactions is abstract. They will affect whether patients pay more at the point of care, whether public programs remain financially sustainable, and whether clinical time is spent on meaningful care or on systems that generate more work without more value.
If the next decade of US health spending is to look different from the last one, the system will need to measure value more honestly, constrain concentration more effectively, and deploy AI with a sharper eye toward substitution rather than addition. The unresolved issue is not whether spending will continue to rise; it is whether the growth will buy better health, or merely a more expensive version of the status quo.