An ACO director I spoke with earlier this year had just finished their first REACH performance period. They had a population health platform. They had a care management team. They had attributed beneficiaries and a benchmark they were confident they could beat.

Then the reconciliation reports came back, and the numbers didn't match what they expected. Attribution had shifted more than they anticipated. Claims data they thought they had captured wasn't reconciling cleanly with what CMS saw. The shared savings they had planned for were significantly smaller than projected.

They had a great plan, but the data infrastructure underneath all of it wasn't built for what REACH actually requires and that gap cost them.

What Makes REACH Different From MSSP at the Infrastructure Level

ACO REACH and MSSP share a lot of surface-level similarities. Both are Medicare shared savings programs. Both use benchmarks, attribution, and quality performance to determine savings. Both require your organization to engage with CMS claims and clinical data.

But REACH has specific infrastructure requirements that MSSP doesn't force you to confront in the same way.

In REACH, you are expected to ingest and work with bulk Medicare claims data directly. CMS pushes CCLF files, Fee Reduction files, and risk score reports to participating organizations. What you do with that data including how quickly you can process it, reconcile it against your clinical records, and turn it into actionable care management, determines how well your care teams can actually perform against the benchmark.

If your infrastructure can't receive, process, and reconcile those files in a timely way, your care managers are working from information that is already weeks behind. And in a program where performance is measured in dollars saved per attributed beneficiary, weeks of lag is not a rounding error.

The Claims Reconciliation Problem Most REACH ACOs Don't See Coming

Here is the specific problem that trips up REACH participants who are otherwise well-organized.

Your EMR captures what happens clinically. Your care managers document outreach, gap closures, and care plan updates. Your quality team tracks measure performance. All of that activity is real, and it represents genuine work on behalf of your attributed population.

But CMS measures your performance based on claims, not on what's documented in your EMR. There is always a lag between when care is delivered and when the claim appears in the data CMS is using to evaluate you. There is also often a gap between what your EMR says happened and what the claim reflects — because of coding variation, documentation gaps, and the natural latency of a complex billing system.

Organizations that don't have infrastructure to reconcile clinical activity and claims end up managing their performance against a picture that is incomplete. They are making decisions about which patients to prioritize, which gaps to chase, and where to allocate care management resources based on data that doesn't match what CMS will ultimately use to score them.

The ACOs that perform consistently in REACH have built a data environment that brings both streams together. They know the difference between a gap that is open in claims, a gap that is clinically addressed but not yet confirmed in claims, and a gap that is confirmed on both sides. That distinction drives entirely different care management workflows, and it can't be made without the right infrastructure.

What REACH Requires Before Your AI Strategy Makes Sense

There is a significant amount of conversation in the ACO space right now about AI-powered care gap closure, predictive risk stratification, and automated outreach. Some of those tools are genuinely useful. But nearly every one of them has the same dependency: clean, unified, current data.

AI doesn't fix a data problem. It inherits one. If your claims and clinical records are reconciled manually by one analyst on a two-week lag, a predictive model built on top of that environment is going to produce predictions with a two-week lag baked in. Faster processing of unreliable inputs is not better performance.

The organizations I've seen get meaningful results from AI-adjacent tooling in the REACH context are the ones that had already solved the data foundation problem first. They had one environment where claims and clinical data lived together, governed by consistent logic, accessible to every team that needed it. The more sophisticated tooling came after that foundation was in place.

If your organization is evaluating AI tools for REACH performance, the first question worth asking is not which tool to buy. It's whether your data environment can support what any of those tools actually require.

What REACH-Ready Infrastructure Actually Looks Like

The organizations that are set up well for REACH have a few things in common that have nothing to do with which platform they bought.

They own their data pipeline. The CCLF files and clinical data live in an environment the organization controls, not a vendor's warehouse that requires a support ticket to access. When CMS releases updated files, the organization can process them on its own timeline.

They have built attribution logic that matches CMS's methodology. Attribution in REACH is more complex than in MSSP. Organizations that have reconciled their internal view of their attributed population against the CMS methodology have far fewer surprises at reconciliation time.

They can stratify their open gaps by status. Not every gap is equally actionable. The organizations earning shared savings have learned to separate genuinely open gaps from gaps that are clinically closed but not yet confirmed in claims, and they have different workflows for each. That stratification requires infrastructure.

And they have made deliberate decisions about what to focus on. REACH rewards depth over breadth. The ACOs earning the most consistent shared savings are the ones that picked a focused set of quality and utilization targets and built their care management workflows around those specific priorities.

If You're Preparing for a New Performance Period

The window between performance periods is the right time to ask honest questions about whether your data infrastructure is actually built for what REACH requires.

Not whether you have a platform. Not whether your team is working hard. Whether the data your care managers are acting on is current enough, accurate enough, and specific enough to your contract to actually drive the shared savings you're projecting.

DAXHS offers a VBC Readiness Assessment designed specifically to help ACOs understand where their data infrastructure stands relative to what their contracts require. It's a structured two-week diagnostic that includes a working session with your leadership and data team, a review of your current EMR and analytics environment, and a written report with findings and a prioritized roadmap.

A lot of REACH participants come in expecting a technology conversation. They leave with a much clearer picture of where their data foundation is strong, where it has gaps, and what to address first.

Want to receive a DAXHS VBC Readiness Assessment? Sign up with the link!

Alex Choquette is the CEO and Co-Founder of DAX Healthcare Solutions. She works with ACOs, FQHCs, and independent physician groups on the data infrastructure and operational realities of value-based care.