Multispecialty Group billing and revenue cycle, 50-state coverage.

Because the three levers that decide it are set separately for each line, usually by different people at different times. Credentialing status is set once per provider per payer and then stops being watched. Evaluation and management leveling discipline follows whichever specialty has had a coding audit recently. Payer contract variance depends on whether that line's fee schedule was ever loaded into the practice management system. A line where all three are attended performs well and a line where two of them are unattended does not, and the gap between them has nothing to do with clinical productivity. That is why we measure per line and per provider rather than at the group level, where the two failure modes cancel each other out and everything looks normal.

Provider by provider, against the peer group inside the same specialty, and in both directions. The group average is the wrong number to watch because it hides two opposite problems that offset. One physician under-codes out of caution and gives away revenue that was documented and earned, so the loss never appears anywhere as a denial. Another over-codes and builds an audit exposure that stays silent until a payer looks at it. Averaged together they read as normal. The correction is documentation coaching at the point the note is written, not a billing adjustment made afterwards, because a leveling pattern is a habit rather than a data entry error.

A provider who is not yet effective with a payer generates encounters that cannot be billed to that payer. The important part is what happens next: those claims are held rather than denied, so they never enter the aging report and never appear in any number the group reviews at the monthly meeting. The loss is real and invisible at the same time. In the 28-provider group we published as our physician group underpayment recovery case study, $273K sat parked across four providers and three payers in exactly that pattern. The control is to track enrollment effective dates per provider and per payer against the start date, and to age the hold queue and report it alongside AR rather than behind it.

Contract variance is the gap between the rate your contract says and the rate the remittance advice actually paid. A group does not see it when no expected rate exists to compare against, which happens whenever the payer fee schedules were never loaded into the practice management system. Whatever the 835 says then gets posted as correct by definition. Commercial contracts in most markets are written as a percentage of the Medicare Physician Fee Schedule, and that reference point moves every year, so an unloaded contract drifts quietly rather than breaking visibly. In the group we published, zero of eleven contracts were loaded and 18.4 percent of lines were paid below the contracted rate once we replayed 24 months of remittance data against the loaded schedules.

Whoever performed the substantive portion, under the split or shared visit rules at 42 CFR 415.140. The point that matters operationally is that this is a documentation decision made in the room, not a billing decision made weeks later from a note that was not written to answer the question. Groups that settle it at the billing desk get it wrong in both directions: they give away revenue when the physician's work is real but not captured in a way that carries it, and they create exposure when the visit is reported under the physician without the record to support it. The fix is a standing rule per service line, applied when the note is written.

It should sit inside it rather than beside it. The performance threshold for the CY 2026 performance period is 75 points, set in the CMS CY 2026 Physician Fee Schedule final rule at 90 FR 49266, and it is met or missed at the group level even though the work to meet it happens line by line with different measure sets per specialty. Groups that run quality reporting as a separate annual compliance exercise abstract the same clinical data twice and find out where they stand too late in the performance year to change it. Building measure capture into the coding workflow removes the duplicate abstraction and turns the threshold into something the group can steer toward during the year rather than report on after it.