How To Prove ROI To Payers And ACOs

Core definitions

• Program launch: first member enrollment date.
• Annual measurement period: each 12-month window after launch.
• Engaged member: eligible member who consented and received program services.
• Program fee PEMPM: per-member-per-month fee paid by payer to vendor.
• Study group: engaged members meeting minimum engagement rules.
• Control group: eligible non-engaged members from the same outreach roster.

Evaluation flow

• Build study/control groups from the same eligibility roster.
• Propensity match control to study.
• Compute relative difference-in-differences savings.
• Subtract fees to get net savings PEMPM.
• Convert to ROI and apply gainshare/repayment terms.

1) Cohort rules that hold up in contract review

• Set minimum continuous engagement (for example 183+ days) to avoid partial-exposure bias.
• Define disenrollment and mortality handling up front.
• Pre-specify power checks and minimum detectable effect targets.

2) Matching approach

• Start with 1:1 greedy matching using a precomputed risk/targeting score.
• Use a strict initial caliper, then expand only by mutual agreement.
• Target standardized differences within ±0.10 for matched covariates.
• Apply explicit outlier handling rules before final matching lock.

3) Savings methodology (relative DID)

Use 12-month pre and 12-month post windows, with identical spend definitions in both groups.

• Control trend = (Control post PEMPM − Control pre PEMPM) / Control pre PEMPM
• Study expected = Study pre PEMPM × (1 + Control trend)
• Gross savings PEMPM = Study expected − Study post PEMPM
• Net savings PEMPM = Gross savings PEMPM − Program fee PEMPM

Include a pre-period parallel-trends check; if violated, rematch or pre-agree on adjustment strategy.

4) ROI conversion

• Net ROI = Net savings PEMPM / Program fee PEMPM
• ROI = 1.0 means savings exactly equal fees.
• ROI > 1.0 means savings exceed fees.

5) Settlement logic (gainshare or repayment)

Use tiered contract rules tied to ROI bands, enrollment volume, and average engagement duration.

• If ROI exceeds threshold, distribute savings by pre-agreed split bands.
• If ROI is below threshold, repay fees up to capped percentages based on cohort depth/tenure.
• Keep all thresholds and caps explicit in the contract appendix.

Governance best practices

• The vendor provides transparent methods, cohort logic, and reproducible code documentation.
• The payer reviews methods and code implementation, then validates final outputs on its data.
• If disputes remain, use a neutral third-party analytic review process.
• Run an early “dress rehearsal” analysis before final measurement close.

Related methods

Previous: Why administrative data is preferred for finding a comparison group

• Do an evaluation on a napkin
• How to evaluate a healthcare program
• Index date in program evaluation
• Small details in evaluation

For more granular data, more recent data, or scientific analysis support, please email us.

Back to Evidence & Performance