Keep three ledgers
Value-based care combines a clinical proposition, an operating business and a payment mechanism. A contract can reward an organization without proving that the intervention improved population outcomes. A clinically useful program can also struggle financially if its cost and payment timing are poorly designed.
I would therefore maintain three ledgers: patient benefit and access; the provider’s economics after delivering the intervention; and the purchaser’s spending after incentives and settlement. A durable arrangement has to explain how those ledgers relate rather than present one favorable number as the whole result.
Separate settlement from causality
A contractual benchmark determines how an arrangement settles. A counterfactual estimates what would have happened without the intervention. They can differ. Actual spending below a benchmark can produce a payment even when the program’s incremental effect is modest or uncertain.
The following example is original illustrative arithmetic, not the formula for a named CMS model. Assume a $100 million contract benchmark, $97 million in actual covered spending, a 50% participant share of the $3 million difference, and $1.8 million of participant program cost. Assume quality conditions are met, and omit corridors, caps and other adjustments to expose the mechanism.
| Ledger | Illustrative calculation | Interpretation |
|---|---|---|
| Contract settlement | $100M − $97M = $3M; 50% share = $1.5M | The benchmark creates a $1.5M participant payment under the stated assumptions. |
| Participant economics | $1.5M payment − $1.8M program cost = −$0.3M | A positive settlement does not cover the participant’s delivery cost. |
| Purchaser spending | $97M actual + $1.5M incentive = $98.5M | The incentive is part of purchaser spending. |
| Causal comparison | Assumed no-program spending of $98M versus $98.5M with incentive | Purchaser spending is $0.5M higher than this assumed counterfactual, despite being below benchmark. |
Change the counterfactual and the estimated program effect changes. Change the benchmark and settlement changes. Neither operation by itself tells us whether patients received better care. The strategic discipline is to show all three questions explicitly.
Underwrite the risk you actually accept
| Exposure | Why it matters | Management response |
|---|---|---|
| Attribution and enrollment | The population can change while staffing and service commitments persist. | Track churn and the continuity of accountability; avoid assuming every enrolled life stays for the full horizon. |
| Risk mix and coding | Payment adjustment and underlying morbidity are related but distinct. | Reconcile clinical need, documented risk and permitted payment rules; do not make coding intensity the care strategy. |
| Benchmark and trend | A favorable settlement may depend on assumptions that do not recur. | Stress-test trend, rebasing and contract terms separately from the intervention effect. |
| Claims and outliers | Incomplete claims and high-cost events can distort early performance. | Use appropriate runout, reserves and sensitivity ranges before distributing apparent surplus. |
| Quality and access | Financial improvement can coexist with delayed or displaced care. | Protect clinical quality, access and patient-choice standards before optimizing return. |
The choice of upside-only, downside risk, prospective care-management funding or another structure should reflect delivery capability and balance-sheet capacity. Transferring more risk to an organization does not automatically give it the tools to manage that risk.
Finance the work before settlement
Care managers, clinical time, analytics, community services and integration require funding before final performance is known. I would model the cash trough as carefully as the expected surplus: launch cost, ongoing delivery, settlement lag, reserves and the point at which additional capital would be needed.
A payer and provider may need to separate payment for essential delivery infrastructure from the outcome-linked component. This can improve participation without abandoning accountability, provided the scope, measures and reconciliation are explicit. A short operational pilot should be paid and evaluated for what it can credibly demonstrate.
Allocate at the intervention level
A population-health budget is a portfolio of interventions with different mechanisms, costs and time horizons. Medication access, post-discharge support, behavioral-health integration and social-care navigation do not have identical return profiles. The relevant question is the incremental benefit of the next unit of each service under real capacity constraints.
I would compare those choices within a common framework of clinical benefit, equity, delivery feasibility and economic exposure. An apparent saving that comes from excluding difficult-to-serve people fails the strategy. A program with a plausible long-term benefit still needs a bounded funding commitment and an evaluation capable of changing the decision.
The board-level test
Before expanding a value-based arrangement, leadership should be able to reconcile the population, contract, clinical mechanism, operating cost and expected cash cycle. It should also know what would happen if the benchmark moved against it, the attributed population changed or the intervention’s effect proved smaller than expected.
The strongest value-based strategy is not the one with the most aggressive savings promise. It is the one whose care model, risk contract and capital structure remain coherent when the favorable assumptions are removed.
Sources & analytical basis
The strategic recommendations are the author’s interpretation. External research and company observations are attributed below; illustrative scenarios are labeled where used.