Yesterday, the Centers for Medicare and Medicaid Services released the latest data on national health expenditures (NHE). The headline number, 7.2 percent growth in 2024, is concerning but hardly a surprise. It follows 7.4 percent growth in 2023. This rate of NHE growth is not sustainable. It exceeds general inflation and growth in the gross domestic product (GDP), pushing the share if GDP devoted to health care spending to 18 percent in 2024; the share of GDP devoted to health care is projected to rise to 20.3 percent by 2033. In fact, these figures may be an underestimate of the fiscal burden of the health care system because spending on some things, such as employer administrative costs, are not captured.
Government policies that shield employers, their workers, those seeking individual coverage and participants in public insurance programs from the financial burden of the health care system can mitigate access and affordability problems from the perspective of those groups. But shifting the financing burden from employers and individuals to taxpayers broadly does not solve the affordability problem and will exacerbate already challenging federal and state fiscal situations. Long term fiscal stability of the system requires addressing the underlying growth in spending, not simply who pays.
What Is Not Driving Spending Growth
Given all the attention to prices and insurer profits, it is important to note that those factors are not the main drivers of spending growth—this time, it’s not the prices, stupid. There was virtually no excess medical inflation (medical inflation above general inflation) for 2023 or 2024. In fact, prices for retail drugs (net of rebates) rose at a rate below inflation. There will certainly be cases of rising prices driving spending, but on average, price growth is not the problem. This does not mean high-priced products and services are not an important component of spending growth, but instead it implies that their contribution to spending growth on average stems from their greater use, not rising prices.
Similarly, non-medical spending by private health insurers, which includes profits, grew at 4.4 percent rate, which is below overall spending growth. As the study notes, the increased medical spending was unanticipated by many insurers, which led to reductions in nonmedical insurance expenditures, the subcategory that includes underwriting gains or losses and thus where profits (or surpluses in the case of non-profit insurers) are recorded.
What Is Driving Spending Growth
The main driver of spending growth is greater volume and intensity of care. Volume refers to the number of encounters (admissions, visits, etc.) and intensity refers to the mix of services (high-cost versus low-cost admissions, shifts from an inpatient to an outpatient setting or from an office to a hospital outpatient department, or the use of expensive vs less-expensive drugs). Most decompositions of health spending growth follow the national health accounts framework, focusing on the sector getting paid (hospital, physicians, retail drugs). This may mask some underlying dynamics related to mix that are important.
Coding Intensity
Payment for health care services is based on service codes and the coding system is dependent on coding patterns. Spending may rise if the care delivered is coded differently, even if the underlying delivery of care is unchanged. There is some evidence from recent years of an uptick in coding for sepsis, greater use of higher acuity evaluation and management codes and use of new Evaluation & Management codes.
The drivers of greater coding intensity are unclear. Coding concerns are not new, but new technologies enabled by artificial intelligence (AI) and ambient scribe technology may be accelerating the trend. Some of the coding may be accurate. But if payment rates are based on earlier coding patterns, the payment rates may not be appropriate. As a result, greater coding intensity increases spending and may add very little clinical value.
AI-Enabled Medical Services
Apart from the role of AI technology in supporting administrative activities such as coding, AI offers great potential to improve the value and efficiency of care. New AI-enabled services, particularly diagnostic services, can better direct care, eliminating unnecessary, potentially harmful, and costly services.
It stands to reason that such tools will lower spending, but realization of that promise depends on how AI services are priced and how providers respond. If the new services are paid for by fee for service, price will likely exceed marginal cost and use may grow beyond what would be optimal. (Because of the potential for quality improvement, the optimal level of use would be above the money-saving level.) Moreover, such tools may require use of other, potentially expensive diagnostic services. For example, AI tools that help diagnose heart disease may require CT-scans that would otherwise not occur. Finally, the productivity gains from AI may free up resources to deliver services that would otherwise not be used.
We are very early in the adoption of AI-based services into the health care system, and it is unlikely that such services contributed significantly to the 2024 spending trend. But going forward, monitoring and evaluating the impact of these services will be a first order concern.
Changes In Health Care Infrastructure, Provider Consolidation And Shifts In Patient Flows
The infrastructure of health care is constantly changing. New outpatient facilities (independent and system-affiliated) are opening and providers are consolidating. Private equity firms have a growing presence in the market. These developments may have important consequences for spending. The shift to lower price settings may lower spending, but integration of physician practices with health systems may raise it because, in general, systems are paid more. Expanding infrastructure may also lead to greater utilization of care. Shifts in patients towards higher-priced providers within sectors (e.g., from low-priced to high-priced hospitals) may also increase spending.
Much of the related policy attention has been focused on antitrust issues and private equity, both of which are important, but the impact of the evolving infrastructure and changing patient flows extends well beyond these issues and remains poorly understood. The key issue is the balance between, on the one hand, efficiency-generating shifts toward lower-priced or better-quality care and, on the other hand, inefficient shifts towards high-priced settings, higher-priced providers within settings, or potentially inappropriate use of services.
Use Of Expensive Products
A non-trivial, though likely not the dominant, driver of spending growth is the increased use of expensive products. Prescription drugs, both in the retail setting and those covered by the medical benefit, garner the most attention. GLP-1s, used to treat diabetes and obesity, are the sentinel example. Despite declines in prices, increased utilization drove up spending on these medications. Yet other products matter as well, including skin substitutes, whose use has skyrocketed. As with all products and services, though more saliently for expensive ones, the core policy questions involve limiting use to situations where the clinical benefit is sufficient to justify the cost (net of any offsets elsewhere) and restraining prices without unduly hampering innovation. Policies such as greater bundling of similar medications, reforming the Average Sales Price+ 6 percent payment policy for drugs under Medicare Part B. and ensuring value is a cap on price should be explored. CMS has been very active in this area, launching several new financing models, including the GLOBE model, the GUARD model and the Generous model, on top of very active implementation of Inflation Reduction Act policies related to drug pricing. Monitoring the impact of these demonstrations on prices, spending, access and innovation will be important.
Looking Forward
Health care spending growth continued at an unsustainable pace in 2024. Early reports suggest spending growth in 2025 will remain elevated. Such growth challenges policy makers and private payers alike.
Reactions often involve efforts to shift the financial burden to other stakeholders. For example, reductions in the federal share of Medicaid spending (the federal Medicare assistance percentage, or FMPAP) shift funding from the federal to state governments; decreases in marketplace subsidies shift some of the burden to individuals, as do employer increases in employee premium contributions. In some cases, shifting who pays may induce reductions in aggregate spending, but such decreases—for example in the case of reductions from higher out-of-pocket cost sharing—may result in lower use of high value services. Our ultimate goal should be to reduce spending in the least deleterious manner possible.
In that spirit, several options include:
- Focusing on strategies to reduce low-value care and inappropriate coding in fee-for-service settings. The WISeR model and private utilization management programs seek to accomplish this goal. The devil is always in the details.
- Improving designs of alternative payment models (APM) that create incentives for providers to practice efficiently. Benchmark-setting rules and risk adjustment are likely the greatest leverage points, but it is also important to consider APM programs holistically; Maintaining too many constantly evolving APM experiments will likely be counter-productive.
- Regulating areas where markets fail. This may include price regulation (including Medicare fee schedule improvement), standardization to support choice, and simplification of administratively burdensome regulations (including broad revision of programs to improve quality). System simplification should be a guiding principle
- Improving market mechanisms to induce more efficient care-seeking behavior and pricing, which may involve antitrust enforcement, providing better consumer information, improving choice support tools, and creating benefit packages based on the principles of value-based insurance design. But market mechanisms have limits and past efforts have not proven very successful. Thus, pursuit of more efficient markets should not forestall necessary regulation.
The specifics of these strategies will be central to establishing a fiscally sustainable health care system. But the spending growth we have experienced, and will experience in the future, reflect system design choices. Our ability to support access to high-quality care at a cost that is affordable in aggregate will require redoubled efforts to reform both health care financing and delivery.
HENRY KOTULA 