From the Editorial Board: Michael E. Chernew, PhD

The concept of premium support has been around since at least the work of Aaron and Reischauer 2 decades ago. The notion was that beneficiaries would receive a fixed “voucher” to purchase coverage and insurers would compete for their business, driving down the cost of coverage and, ideally, care. Crucially, in the Aaron and Reischauer proposal, the voucher amount rose with aggregate medical spending, broadly protecting consumers from undue burden. More recent variants remove this protection, tying the voucher to a different, often slower-rising, index. The appeal of premium support—type programs is that they limit the fiscal liability of payers. The problem, depending on the design, is that they shift the burden to patients. There is very little evidence that the forces of competition in such a model will sufficiently constrain spending.

Yet, it is reasonable to advocate placing the healthcare system on a budget. At some point, we must rein in spending. But we should focus on overall spending growth, not just the payer’s portion. Some entity must bear the risk if spending exceeds the “budget,” and providers are likely the best entity to do so, as they are central to care decisions.

A provider-side premium support model would set a population-based spending target for total medical expenses, and providers would be at risk if they exceed the target. We already have several variants, including a number of accountable care organization models and Maryland’s hospital budgeting system (which is evolving to a more patient-centered model). We are now experimenting with details such as how much risk to impose, how to risk-adjust the budget, how to accommodate small provider groups, and how to attribute patients to providers.

Like the demand-side equivalents, the key is the process for setting the budget target. If we allow the target to rise with spending driven by an unconstrained portion of the system, we will likely not achieve sufficient savings. The forces driving up spending may simply be too strong.

A provider-side premium support model would have the target rise at an affordable rate, perhaps based on gross domestic product. This is not a radical departure from current population-based models. It simply modifies the existing systems of benchmark-setting to meet a budget goal as opposed to reflecting spending outside of the population-based payment system. Although the model cannot be applied universally (some practices are too small), it has the virtue of functioning well even as enrollment in a system grows. Unlike the Sustainable Growth Rate, which had similar goals, the responsibility rests with provider groups, as opposed to collectively.

The key is for the budget to be reasonable (it could be increased if deemed insufficient to accommodate desired innovations). Providers would retain control over resources and share in the efficiencies they create. Most importantly, if providers cannot succeed with a fiscally sustainable rate of spending growth, the system will inevitably fail because beneficiaries and taxpayers will not be able to finance it. Given the early, albeit modest, success of population-based payment models, we should focus on how to transition to a more sustainable model and on the dull, but crucial, details.