Accountable care organizations, or ACOs, were created to change how provider groups are paid. Under Medicare payment systems, each visit or service generates a separate payment, so more activity can improve the financial result. An ACO instead makes providers jointly responsible for the cost and quality of care delivered to a defined population. When the group produces savings, it may share in them.
This was the premise behind CMS’s announcement of a large new cohort of ACOs at the end of 2013, part of the broader landscape of CMS innovation models. During the same period, CMS sought input on changes to the more demanding Pioneer program, which had lost participants. Together, those developments showed that the basic concept attracted interest while its more demanding form was difficult to implement.
The problem with paying for volume
The ACO premise is simple: paying for volume can produce more volume. When payment is tied to individual services, providers have a financial reason to deliver activity but less responsibility for the total cost and quality of care. An ACO instead makes a provider group responsible for the care of a population.
That arrangement is intended to encourage care coordination and savings while maintaining quality. The organization is judged on more than what happens during one visit. It must account for the performance of the broader population assigned to it.
The concept is straightforward, but putting it into operation is difficult. A group cannot be held meaningfully responsible unless it knows which patients belong to it and how performance will be measured. The benchmark also must give the group a realistic chance to produce savings. These questions can determine whether the model works outside theory.
The issue is not simply whether providers can reduce waste. It concerns how responsibility is assigned and measured. The surrounding rules determine whether that responsibility is clear enough to affect decisions.
Attribution makes the population hard to define
Attribution was one of the main implementation problems. Patients do not always remain within one organized network or return to the same provider for every service. They may choose different providers over time, making it difficult to determine which patients belong to which ACO when performance is measured.
Attribution is more than an administrative task. It determines whose care is included in an organization’s results. If the assigned population is unclear, an ACO may be held responsible for spending or quality outcomes it does not fully control. It may also receive credit for patients whose care it manages only in part.
This issue reveals a basic tension in the model. An ACO needs a defined population to accept responsibility for outcomes, but patients may experience health care as a series of separate choices among providers. The policy seeks responsibility for a population even when individual care relationships remain divided.
The same problem affects reported savings. Before the system can determine whether an organization improved, it must decide which patients belong in the comparison. Without that decision, the link between the organization’s actions and its measured results remains uncertain.
Benchmarks can penalize prior efficiency
Benchmarking creates another difficulty. To determine whether an organization generated savings, its performance must be compared with a standard based on its past performance or another measure. A group that already operated efficiently could therefore be judged against a strong starting point.
Improvement is harder to show for a group that began with low spending or effective care. An organization with higher spending may have more visible opportunities to reduce costs, while an efficient group may have fewer ways to produce additional savings.
This concern helps explain why attrition from the Pioneer program mattered. Some organizations that appeared positioned to succeed found the risk unattractive under the terms offered. Their decisions showed that participation depends on the structure of the risk as well as on the appeal of coordinated, outcome-focused care.
An ACO model can therefore be attractive in principle and difficult to accept. If the benchmark does not account for the starting point, a group may have little room to show improvement. If the risk is out of proportion to the opportunity for savings, participation may be difficult to maintain.
The lesson for transactional care
Direct-to-consumer telehealth sits at the opposite end of the payment spectrum from the ACO model. It is transactional by design and is paid per interaction, with no accountability for outcomes beyond retention. This describes the incentive structure, not the quality of every provider.
The distinction matters when a person chooses a GLP-1 provider. Since the payment model is transactional, the quality of follow-up must be assessed rather than assumed. For people considering modern medication costs, our GLP-1 clinic rankings consider follow-up quality alongside the initial interaction.
The comparison concerns more than payment. It also concerns the responsibility assigned to the provider. An ACO is designed to make providers jointly responsible for the cost and quality of care for a population. Direct-to-consumer telehealth is organized around individual transactions. A telehealth provider may offer follow-up, but its continuity must be evaluated separately from the payment model.
The ACO experience, also illustrated by CMS Selects 27 ACOs, offers a clear lesson about outcome-based payment. Paying for outcomes instead of visits is a sound idea, but implementation depends on the rules around it. Attribution determines which patients count. Benchmarking affects whether savings can be demonstrated. Risk distribution affects whether organizations participate and remain in the model. The stated goal matters, but the incentives and obligations determine how much responsibility providers can exercise in practice.