The case for a system of record

Ask a plan leader to map how a non-medical benefit moves and you’ll get a confident answer at the two ends. At the front, there’s the benefits catalog: what’s offered, to which population, under which plan. At the back, there’s finance and claims: what got paid, booked, and reported. Both ends are systematized, owned, and audited.

Now ask what connects them. The answers get vaguer: a vendor portal here, a care management note there, a shared spreadsheet, an email confirming a delivery. The middle, where the benefit actually reaches a member, is the least instrumented part of the whole chain. And it’s the part regulators, auditors, and finance are increasingly asking about first.

Why the middle got left out

This isn’t negligence; it’s history. Claims systems were built for medical services that generate claims. Benefit catalogs were built to define and communicate offerings. Neither was designed for a world where a meaningful share of value is a ride, a meal, or an hour of in-home help delivered by a third party and funded from a blend of sources. The supplemental benefit rules1 say what a plan may offer. They do not supply the mechanism that proves it arrived.

So the middle got handled the way new problems always get handled first: manually, with whatever tools were nearby. That worked when non-medical benefits were a rounding error. It works far less well now that they’re a competitive and compliance priority.

Categories emerge when a job that used to live in the margins of existing systems becomes too important, and too scrutinized, to keep doing by hand.

What a system of record is

A system of record sits between the catalog and the ledger and does one job well: it records what actually happened. Not the intent (the catalog has that) and not the accounting entry (the ledger has that), but the delivery event itself, with member, benefit, date, vendor and funding source, captured once, at the moment it occurs, and kept as an immutable record.

From that single spine, four things stop being projects and start being queries:

• Execution: a running ledger of what was delivered, not just authorized.
• Sequencing: the logic that keeps benefits and funding sources from colliding or duplicating. This is the operational form of the coordination responsibility and covered-benefit boundaries a D-SNP already has to document in its State Medicaid agency contract.2
• Visibility: a live view of benefit spend across vendors and programs.
• Attribution: reporting that ties every aggregate number back to the individual events beneath it, records that must survive 10 years and be produceable to HHS or the Comptroller General on request.3

Why now

Three forces are converging. Non-medical benefits are growing as a share of what plans offer. Oversight of how those dollars move is already specific: a plan offering SSBCI must keep written eligibility policies, document every eligibility determination, publish its objective criteria on a public-facing website, and hold those standards unmodified for the full coverage year.4 And the manual approach that carried teams this far collapses exactly when volume and scrutiny both rise, which is now. When a job gets important enough and watched enough at the same time, it stops being a spreadsheet and becomes infrastructure.

The bet

The bet behind a system of record is simple: the record of what was delivered is about to be as load-bearing as the record of what was billed. Plans that treat it that way, capturing the delivery event as first-class data rather than reconstructing it under audit pressure, will spend less time reconciling, defend their spend more easily, and be able to prove the value they’re already creating.

The catalog says what you promised. The ledger says what you paid. The system of record says what you actually did. For non-medical benefits, that middle statement is quickly becoming the one that matters most.