Shared Savings Integrated Funding Pilots: How to Align Incentives Without Encouraging Cost Shifting or Risk Avoidance

Shared savings is one of the most common “integrated funding pilot” structures because it promises a simple trade: invest upstream, reduce avoidable utilization, and share the financial upside. In practice, shared savings can create new failure modes—cost shifting between payers, risk avoidance of complex people, and “paper savings” that aren’t real cashable reductions. A defensible pilot designs incentive alignment, attribution, and quality safeguards as operational mechanics, not as policy aspirations.

This article sits within Integrated Funding Pilots and should be read alongside evidence and oversight approaches in Using Data for Commissioning & Oversight.

What shared savings pilots are trying to achieve

Shared savings pilots aim to create a financial reason for multiple partners to invest in prevention, stabilization, and navigation—especially where the “spender” and the “saver” are not the same organization. The core challenge is that utilization reductions must be attributable, measurable, and not achieved by simply pushing risk elsewhere (for example, moving costs from inpatient to ED, or from ED to behavioral crisis, or from Medicaid to uncompensated care).

How shared savings fails without tight design

Pilots typically fail in three predictable ways. First, savings are calculated using weak baselines or noisy measures, leading to disputes and loss of trust. Second, incentives inadvertently reward serving easier populations while excluding high-acuity cohorts. Third, partners cannot evidence what operational changes produced the savings, so learning does not scale and oversight bodies see only financial claims without service reality.

Operational Example 1: Quality-gated shared savings tied to “no harm” indicators

What happens in day-to-day delivery

The pilot defines a monthly operations cycle where utilization and safety indicators are reviewed together. A performance analyst produces a standardized pack showing ED use, admissions, and high-cost episodes for the target cohort, alongside “no harm” indicators such as follow-up completion, medication reconciliation timeliness, safety event reporting, and complaint patterns. Operational leads meet with clinical governance to review outliers, confirm whether service changes were implemented, and log remedial actions. Savings are only eligible for distribution if quality gates remain within agreed thresholds and action plans are actively managed.

Why the practice exists (failure mode it addresses)

This exists to prevent the classic shared-savings failure mode where savings are achieved by under-serving, delaying care, or shifting work onto families without support. Without explicit quality gating, the model can reward reductions that are operationally unsafe.

What goes wrong if it is absent

If quality gates are missing or weak, partners may reduce utilization by narrowing eligibility, limiting response, or reducing follow-up intensity. The real-world consequence is increased crisis bounce-back, missed deterioration, medication harm, and safeguarding risk—often surfacing later as sentinel events, litigation, or regulatory scrutiny.

What observable outcome it produces

A quality-gated model produces an auditable link between savings and safe delivery. Oversight can see that reductions in utilization occur alongside stable (or improved) safety indicators, documented decision-making, and a clear trail of actions taken when metrics drift.

Operational Example 2: Attribution rules and cohort “locking” to prevent gaming

What happens in day-to-day delivery

The pilot defines an attribution engine that assigns individuals to the pilot cohort based on objective criteria (for example, enrollment status, geography, risk score thresholds, or confirmed utilization patterns). Once assigned, individuals are “locked” for a defined period so partners cannot remove them when needs escalate. Frontline teams receive a live roster and use it for daily huddles, outreach planning, and escalation decisions. Changes to the roster require documented justification and independent review, with periodic audits to confirm the cohort was managed consistently.

Why the practice exists (failure mode it addresses)

This addresses risk avoidance and cherry-picking—the tendency to focus support on people most likely to show quick savings while excluding those with complex needs who require longer stabilization timelines.

What goes wrong if it is absent

Without clear attribution and cohort locking, the pilot becomes non-comparable over time. Savings claims become disputed because the cohort quietly changes. High-acuity individuals lose continuity, are repeatedly re-triaged, and end up back in crisis pathways, undermining both outcomes and credibility.

What observable outcome it produces

Strong attribution produces stable measurement and defensible savings calculations. It also produces clearer operational learning: teams can compare like with like, understand which interventions work for complex cohorts, and evidence equitable reach rather than selective service delivery.

Operational Example 3: Reinvestment pools with defined allowable uses and tracking

What happens in day-to-day delivery

The pilot sets aside a reinvestment pool funded from a portion of verified savings. A governance group approves reinvestment proposals tied to specific system constraints (for example, after-hours access, rapid follow-up capacity, short-term stabilization supports, or workforce training). Each approved reinvestment has an implementation owner, timeline, and measurement plan. Program managers track delivery monthly, documenting what was implemented, where, and what operational changes occurred (e.g., referral turnaround, response time, follow-up completion). Finance and performance teams reconcile reinvestment spend against expected impacts.

Why the practice exists (failure mode it addresses)

This prevents the failure mode where shared savings becomes a financial distribution exercise rather than a service improvement mechanism. Without reinvestment discipline, pilots generate one-time gains without building sustainable capability.

What goes wrong if it is absent

If savings are distributed without structured reinvestment, partners struggle to sustain improvements. Utilization reductions can reverse, and the pilot becomes vulnerable to political challenge because stakeholders cannot see tangible service benefits or capacity gains.

What observable outcome it produces

A reinvestment pool produces visible system improvements: new capacity, improved timeliness, reduced escalation, and a measurable operational trail linking savings to service redesign. It also strengthens audit defensibility because allowable uses, approvals, and impacts are documented.

What funders explicitly expect to see

Expectation 1: “Cashable” and attributable savings with transparent methodology. Funders expect a baseline method, clear attribution rules, and evidence that reductions represent real system impact, not displaced cost or measurement noise.

Expectation 2: Strong equity and risk safeguards. Oversight bodies expect pilots to demonstrate that high-acuity and marginalized groups are not excluded, with cohort monitoring, access checks, and documented escalation protections.

Making shared savings pilots defensible and scalable

Shared savings can support integrated funding only when incentives are bounded by quality gates, transparent attribution, and reinvestment discipline. The operational question funders ask is simple: can you show what changed on the ground, why it was safe, and how the financial result was produced without shifting risk elsewhere?