The moment it solves
The claim left. The underpayment did not announce itself.
A payment that looks fine against a national benchmark can still be short against the rate this practice negotiated. Detection recall is a sensing metric. It is not a recovery rate, and this study is not a payer leaderboard.
How it works
Sensing, then a queue. Staff still decide.
- 01
Claim Radar keeps reading remittances
Always-on. Contract-aware. Not a nightly export someone has to remember to run.
- 02
ReQover makes a queue
Appealable work versus documentation/education. Not every flag is cash.
- 03
Staff decide
ReQover does not promise recoveries and does not guarantee savings.
Evidence of value
Classed evidence. Not every number is cash.
MEASURED PERFORMANCE
97.8%
Tag-path recall on the downcode-detection review log
91 / 93 review-log rows detected; effective 91 / 92 valid rows. Do not imply every flagged underpayment is recoverable.
How this was measured
OpAgent Benchmark: downcode-detection tag-path recall on the review log.
Sample: 91 / 93 review-log rows · 91 / 92 valid rows
Detection is not recovery. Do not create payer leaderboards from this study.
Source: OpAgent Benchmark
MEASURED PERFORMANCE
46 / 47
Queue split: appealable vs documentation/education
46 appealable · 47 documentation/education. Not every flagged underpayment is recoverable.
How this was measured
OpAgent Benchmark: staff queue split from the downcode-detection study.
Sample: 93 review-log rows
A queue split is not a recovery rate.
Source: OpAgent Benchmark
Methodology / caveat
Tag-path recall is a detection metric on the downcode-detection review log. Do not imply every flagged underpayment is recoverable. Do not publish payer leaderboards from this study.
OpAgent Benchmark · Claim Radar / ReQover · detection ≠ recovery
Where it connects
The rest of OpAgent is already in the room.
- CQO
The understanding layer that keeps working after the encounter.
- DataQorp
Declared policy, kept separate from what the remittance actually did.
- CodeIQ Visit
The visit coding that should have been supportable before the claim left.