Ask a risk adjustment team how they performed last year and most will tell you their capture rate. Ask what CMS actually paid on, and the room gets quieter. The gap between those two answers is where revenue integrity in risk adjustment lives, and for most plans nobody owns it, because coding owns capture, encounters owns submission, and finance owns the projection, and the loss happens in the space between them.
Revenue integrity is not a bigger capture number. It is the discipline of making sure every condition a member actually has ends up recognized by CMS, that nothing recognized is indefensible, and that you can see the money at every step in between. This post covers where the leaks are, which ones analytics can close, and how to measure the thing that pays instead of the thing that is easy to count.
The focus here is Medicare Advantage, where payment runs through the CMS-HCC model and the encounter data system.
Where Revenue Integrity in Risk Adjustment Breaks Down
RAF revenue leaks at four separate points, and most plans have real visibility into only one of them.
A condition is never documented. The member has it, the provider treats it, and nothing in the note supports coding it. This is the leak everyone knows about, and it is the one every vendor sells against.
A condition is documented but never coded. The evidence sits in a chart nobody read, usually because the chart was never selected for review. NLP finds these, but only if the clinical documents are actually in the system.
A condition is coded but never lands on a risk-eligible encounter. The coder did the work. The diagnosis went onto a claim that was not risk-eligible, or onto an encounter that was never generated, or onto a file that failed validation and sat waiting for a correction that came after the deadline.
A condition is submitted but never accepted. The encounter was rejected, or the diagnosis was flagged as not risk-eligible, or a later void removed it. Internal reporting counts it. CMS does not.
Only the first two are coding problems. The last two are operational, and they are the ones most analytics stacks cannot see, because the data lives in response files rather than in charts.
The Number Most Plans Measure Is the Wrong One
Capture rate measures what you sent. Revenue integrity measures what CMS took. Those are different numbers, and only one of them is on the payment.
Three CMS files close that distance. The MAO-002 report tells you whether an encounter was accepted or rejected. The MAO-004 report tells you which diagnoses on accepted encounters were treated as eligible for risk adjustment, and which were rejected or deleted. Read together, they turn a submission log into an accounted-for number.
The Model Output Report, or MOR, goes one step further. It shows the HCCs CMS actually recognized for each member. Comparing your internal capture against the MOR is the single most useful reconciliation in risk adjustment, because the delta between the two is your real leak, expressed in conditions rather than in guesses.
The Monthly Membership Report, or MMR, tracks RAF and payment at the member level as scores move across the Initial, Mid-Year, and Final reconciliation cycles. Plans that only look at the Final run find out about problems when it is too late to correct them. Plans that watch RAF progression across all three cycles catch drift while the correction window is still open.
Revenue Integrity Runs in Both Directions
Most revenue integrity conversations are about money left on the table. Half of the exposure is on the other side.
A condition captured without documentation that supports it is revenue you will give back, usually with company. RADV audits sample members and extrapolate the findings across the contract, so a handful of unsupported captures in a sample becomes an error rate applied to a population. The repayment is not the value of the codes the auditor found. It is the value of what those codes imply about everything the auditor did not look at.
This is why capture and defensibility cannot be run as separate programs with separate scorecards.
A plan that hits its RAF target with weak evidence has not protected revenue, it has moved the loss into a future year and added interest.
Real revenue integrity means every captured condition traces back to a documented line in a chart and a coder who signed it, and that you can produce that trail without a scramble.
Want to see where your RAF is leaking?
What Revenue Integrity Analytics Should Give You
Five capabilities separate analytics that protect revenue from dashboards that describe it after the fact.
A dollar value on every gap and every rejection. Conditions and error counts do not help a team decide what to work. Expected financial value does, and it is the only ranking that survives contact with a queue that cannot be finished.
Closed-loop reconciliation against CMS responses. If the analytics never ingest 999, TA1, 277CA, MAO-002, and MAO-004, they are reporting on what left the building rather than on what counted.
Validation before submission, not error reports after. Catching a non-risk-eligible claim carrying a risk diagnosis, or a missing diagnosis on an encounter about to go out, is worth more than a report explaining why the file failed.
RAF progression across payment cycles. Scores move between Initial, Mid-Year, and Final. Analytics that show only a point-in-time number hide the drift that finance most needs to see.
Traceability from chart to CMS outcome. Every captured diagnosis should link back to the document and the coder it came from, and forward to what CMS did with it. That single thread is what makes revenue both provable and defensible.
How Invent Health Supports Revenue Integrity
Invent Health runs revenue integrity as one connected loop rather than three disconnected reports, and it starts with analytics, not coding.
Risk Analytics decides what is worth working. Suspect logic runs across historical, lab, pharmacy, comorbidity, and chart-derived signals, drawing on EMR data through CCD and FHIR. It scores the medical HCC RAF and the pharmacy side, validates internally captured diagnoses against CMS-recognized HCCs in the MOR, and tracks RAF progression across the Initial, Mid-Year, and Final cycles using the MMR. Corrections and gaps are ranked by expected financial value, so teams work the money rather than the list.
The Coder Workbench confirms the condition is real. AI-assisted coding reads the clinical documentation, shows the supporting evidence, and recommends ICD-10-CM codes for a certified coder to confirm and sign. It is a two-pass, coder-in-the-loop model. The AI surfaces what matters, a person makes the call, and every code stays linked to the chart for audit-defensible lineage.
Encounter Submissions makes sure it counts. One engine generates and validates clean submissions, 837P, 837I, and DME for Medicare EDPS and Edge Server XML for ACA, and flags format errors, missing diagnoses, risk-impacting diagnosis errors, and non-risk-eligible claims carrying risk diagnoses before a file leaves. It then reconciles what comes back, matching MAO-002 and MAO-004 for Medicare and the Edge Server reports for ACA, so a diagnosis that did not count returns as a known condition needing a corrected encounter, ranked by its RAF value, instead of reappearing as a fresh suspect.
Because analytics, coding, and encounter submission share one platform, the money is visible at every step, from the chart through coder validation to the 837, the CMS response, and the payment cycle it lands in.
Why Revenue Integrity Gets Harder in 2027
Most plans still treat retrospective chart review as the correction of last resort. Miss something during the year, catch it in a sweep, submit it as a chart review, recover the revenue.
That path narrows next year. The CY 2027 Rate Announcement ends unlinked chart reviews beginning with the 2027 payment year, which means a diagnosis found in a chart review has to tie back to an encounter CMS accepted. The July HPMS memo on CY 2027 risk adjustment implementation lays out what plans need in place. A condition with no accepted encounter behind it no longer has a route to payment, which moves encounter acceptance from an operations metric to a revenue metric.
Two other pressures compound it. The V28 model removed a large share of diagnosis codes from risk-adjustment mapping and re-based the values, so fewer conditions carry payment and each one matters more. And RADV audits are moving toward every contract every year, with sample findings extrapolated across the contract, which raises the cost of a weakly supported capture well above the value of the code itself.
The plans that hold their revenue through the change will be the ones that already measure capture at acceptance and can trace every dollar back to a chart.
See Where Your Revenue Is Going
The fastest way to size this is to reconcile a payment year you have already closed. Compare what your team captured against what the MOR recognized, then look at what MAO-002 and MAO-004 say about the difference. The number is usually larger than expected, and most of it is recoverable in the years ahead once you can see it.
Frequently asked questions
What is revenue integrity in risk adjustment?
Revenue integrity in risk adjustment is the discipline of making sure every condition a member has is captured, documented, submitted, and accepted by CMS, and that nothing accepted is indefensible at audit. It covers both under-capture and over-capture, because both cost money, just in different years.
How can health plans improve revenue integrity through risk adjustment analytics?
By measuring capture at CMS acceptance rather than at submission. That means reconciling submitted diagnoses against MAO-002 and MAO-004, comparing internal capture against the MOR, tracking RAF across payment cycles in the MMR, and ranking every gap and correction by expected financial value rather than count.
Where does risk adjustment revenue leak?
At four points. Conditions never documented, conditions documented but never coded, conditions coded but never placed on a risk-eligible encounter, and conditions submitted but never accepted by CMS. The first two are coding problems. The last two are operational and are the ones most analytics stacks cannot see.
What is the difference between capture rate and accepted capture?
Capture rate measures what a plan submitted. Accepted capture measures what CMS recognized for risk adjustment after processing. Only accepted capture appears in payment. Plans that report on capture rate alone can show a strong year internally while losing revenue that never reached the risk score.
What is MAO-004 and how does it affect revenue?
MAO-004 is the Medicare Advantage report showing which submitted diagnoses CMS accepted, rejected, or deleted for risk adjustment. It is where a plan learns that a documented, coded, submitted condition did not count. Without it, teams re-chase conditions they already captured and never see the revenue impact.
What is the MOR and why does it matter for revenue integrity?
The Model Output Report shows the HCCs CMS recognized for each member. Comparing internal capture against the MOR reveals the real gap between what a plan believes it captured and what CMS actually used to calculate payment, expressed in specific conditions rather than in estimates.
How does the MMR help track risk adjustment revenue?
The Monthly Membership Report tracks member enrollment, RAF, and payment over time. Because scores move across the Initial, Mid-Year, and Final reconciliation cycles, watching RAF progression in the MMR surfaces drift early enough to correct, instead of after the final run closes the window.
Can over-capture hurt revenue integrity?
Yes, and often more than under-capture. RADV audits sample members and extrapolate findings across the contract, so unsupported captures in a small sample can produce a repayment applied to the full population. Revenue captured without defensible documentation is generally revenue borrowed from a later year.
How do rejected encounters cause revenue loss?
A diagnosis on a rejected encounter never reaches the risk score, even when the chart fully supports it. The coding work was done and the revenue still does not arrive. Because internal reports usually mark the condition as captured, the loss stays invisible until someone reconciles the CMS response files.
How should plans prioritize encounter corrections?
By expected financial value, aging, and the likelihood the correction succeeds. A prioritized worklist ranked on RAF impact puts the highest-value corrections in front of a team that cannot work every rejection, which matters most late in a cycle when the correction window is closing.
What changes for risk adjustment revenue integrity in 2027?
Unlinked chart reviews end with the 2027 payment year, so chart review diagnoses must tie back to an encounter CMS accepted. Combined with V28 lowering the number of payable conditions and annual RADV auditing, encounter acceptance becomes a revenue metric rather than an operations metric.
How does Invent Health support revenue integrity?
Risk Analytics ranks gaps and corrections by financial value and reconciles capture against the MOR and MMR. The Coder Workbench confirms each condition against documented evidence with a certified coder. Encounter Submissions validates files before they leave and reconciles MAO-002 and MAO-004 responses after.
