One Price for the Whole Episode Instead of Each Step
A bundled payment covers everything involved in treating a condition, from the procedure through recovery, at a single negotiated amount. It transfers the cost of complications to the providers who can prevent them.
The Unit of Payment Determines the Behaviour
Under fee for service, every element of an episode of care is billed separately: the surgeon, the anaesthetist, the facility, the imaging, the implant, the rehabilitation, and any readmission that follows.
Each participant is paid for what they do and none is responsible for the total. If a complication requires a further procedure, that procedure generates additional revenue for everyone involved in it.
A bundled payment changes the unit. A single amount covers the entire episode, defined in advance, and the participants divide it. If the episode costs less, they keep the difference. If it costs more, they absorb it.
Defining the Episode Is the Whole Design
An episode requires three boundaries and each is contested.
The trigger is the event that starts it, typically an admission or a procedure. The duration extends the episode forward, commonly thirty, sixty, or ninety days after discharge. The scope defines which services are included and which are excluded as unrelated.
| Design Choice | Consequence |
|---|---|
| Longer post discharge window | Captures readmissions and recovery |
| Shorter window | Complications fall outside the bundle |
| Broad service inclusion | Stronger coordination incentive |
| Narrow inclusion | Costs shift to excluded services |
Anything outside the bundle is still paid the old way, so every boundary is a place where cost can be moved rather than reduced. Designing an episode is mostly an exercise in anticipating where the leakage will go.
Where It Has Worked
Bundles perform best where the episode is well defined, the procedure is elective and planned, the patient population is relatively predictable, and there is genuine variation in cost that is not driven by patient severity.
Joint replacement is the standard example and has been the most studied. Evaluations of mandatory and voluntary programmes have generally found modest reductions in episode spending, driven substantially by reduced use of institutional post acute care, meaning fewer discharges to skilled nursing facilities and more to home with support.
That is a real efficiency where the shift is clinically appropriate, and it is also the finding that generates the most concern, since the same shift could reflect under provision of needed rehabilitation. Studies examining outcomes have generally not found deterioration, which is reassuring without being conclusive.
Where It Works Poorly
Chronic conditions fit badly, because there is no natural trigger and no endpoint. Managing diabetes is not an episode, it is a continuous state, and forcing it into an episode structure produces arbitrary boundaries.
Complex or unpredictable care also fits poorly, since a bundle transfers variance to the provider and providers cannot bear variance they cannot manage. That is why bundles include outlier provisions removing extremely expensive cases from the arrangement, which protects providers and reduces the incentive at exactly the cases where cost is highest.
The Selection Problem
The predictable hazard is that a provider paid a fixed amount per episode prefers episodes likely to be straightforward.
Risk adjustment varies the payment according to patient characteristics, which addresses it partially and imperfectly, since the observable characteristics do not capture everything that makes a patient complex.
Evaluations have looked for evidence that participating providers shifted toward healthier patients, with mixed findings. The concern is well founded even where the evidence is ambiguous, because the incentive is unambiguous.
Who Holds the Risk
A structural question with real consequences is which entity is accountable for the bundle. Where a hospital holds it, the hospital controls facility costs and post acute referrals but not physician decisions. Where a physician group holds it, the reverse applies.
Programmes have experimented with both, and the arrangements that perform best generally involve genuine gainsharing between the facility and the clinicians, since neither can reduce total episode cost alone.
That requires legal structuring, because arrangements paying physicians based on reducing services raise questions under laws restricting payments that could influence clinical decisions, and specific waivers have been required for these programmes to operate.
The Policy Trajectory
Bundled payment programmes have been introduced, expanded, made voluntary, made mandatory in selected markets, cancelled, and reintroduced across successive administrations, which makes the evidence base harder to interpret than it should be.
The consistent finding across evaluations is modest savings in well defined surgical episodes, achieved principally through post acute care, without detectable harm to outcomes. That is a real but limited result, and it is considerably smaller than the enthusiasm the model has periodically attracted.
The Bottom Line
Bundled payment reduces the unit of payment from a service to an episode, which makes somebody responsible for the total cost of treating a condition rather than for their own contribution to it. It works where episodes are definable and predictable, principally elective surgery, and produces modest savings mostly by reducing institutional rehabilitation. Its structural weakness is that everything outside the episode boundary remains paid the old way, which means the design of the boundary determines whether cost is reduced or merely relocated.