The Denial Was Born Three Weeks Ago
A denial lands on a Tuesday. It feels like a Tuesday problem, so someone works it on Tuesday. But the mistake that caused it wasn’t made on Tuesday. It was made three weeks earlier, at a registration desk, or in an eligibility check that ran at the wrong time, or in an authorization nobody was watching. The remittance is just the delivery notice for a decision made weeks upstream, and by the time it arrives, the trail has gone cold and the fix is expensive.
This is why working denials faster has a ceiling. You’re catching the output of factories that are still running. Point a practice at a lower denial rate and the ones who actually get there did the unglamorous thing: they went upstream and shut the factories off. There are five of them, and they run in the order the money moves through the practice.
Factory One: Registration
A transposed policy number. A stale address. A subscriber name that doesn’t match the payer’s file. Tiny keystrokes at intake, made by a busy person during a check-in, that turn into denials three weeks later with rework stapled to them. Nobody was careless; the desk was moving fast because the lobby was full.
The fix is structural, never a memo about being more careful. Digital intake that pulls data instead of asking a human to retype it. Required fields that can’t be skipped past. A same-day verification pass on every new registration, before the error has time to travel. Every field a person doesn’t rekey is an entire category of denial that stops being manufactured.
Factory Two: Eligibility Timing
Coverage checked once at scheduling and never again misses everything that changed between then and the visit: the plan that lapsed, the deductible that reset, the patient who switched jobs. Coverage never checked at all is a coin flip you submitted as a claim. The timing is the whole problem, because eligibility is true only as of the moment you check it.
So check it on a schedule. At booking, and again in the days before the visit, with any mismatch routed to a person while the patient is still reachable, instead of after the payer answers for you. A coverage problem caught two days early is a phone call. The same problem caught after submission is a denial, an appeal, and sometimes a patient who now owes money they didn’t expect.
Factory Three: The Authorization Runway
Authorizations rarely fail on merit. They fail on timing: the visit that outran the number of sessions approved, the renewal nobody was tracking, the service that needed sign-off nobody requested. Every one of those is a calendar problem dressed as a clinical one.
So treat every authorization like a runway with a length. Visits remaining. Days remaining. Surface the ones about to run out while there’s still time to do something, instead of discovering the overrun in a denial. An auth that expires on a dashboard is a task somebody handles. An auth that expires inside a claim is money gone, plus a patient who got care nobody could bill for. The authorization guide covers the full build.
Factory Four: Coding Without Guardrails
The same visit coded three ways by three clinicians isn’t a disagreement about medicine. It’s a denial generator, and it’s also a pattern payer analytics can read across your whole claim history before you’ve ever looked at it yourself. The spread costs you twice: the conservative coders leave money on the table, and the outliers draw attention you didn’t ask for.
The fix respects clinical judgment while putting rails on the parts that keep tripping. Written thresholds for the choices that recur, so the common cases stop being personal style. A review trigger for the genuinely unusual. And a feedback loop that shows each clinician their own pattern, framed as a pattern, never as an accusation about a single claim. Most coding denials trace to a handful of code-and-payer pairs, and those pairs are sitting in your own remittance data the first time anyone groups it. The coding piece is the full version.
Factory Five: Submission Without Edits
The last factory is the gap between charge entry and the payer. Claims that leave without being scrubbed against rules you already know. Claims batched weekly, so a defect made on Monday isn’t discovered until it bounces the following week, cold and aged.
Daily submission with edits applied is the cheapest denial prevention that exists, and it’s a schedule decision, not a purchase. It catches the defect the day it’s made, while the visit is fresh and the fix is a two-minute correction instead of a three-week investigation.
The Loop That Closes the System
The five factories stay shut only if the downstream side reports back. Every denial that still gets through gets traced to the factory that made it, and that factory gets adjusted. That’s the pairing with the working-denials system: one drains the backlog, the other shuts off the supply, and together they move a practice out of arguing with payers and into running a process where there’s simply less to argue about. What that looks like once it’s counted honestly is the denial rate guide.
Where to Start
Pull one month of denials, tag each one with the factory that made it, and fund the factory with the biggest total first; the data picks the order for you. Then grab 30 minutes with us. Prep nothing. We’ll show you the factory map built from real remittance data, and you’ll see which of the five is manufacturing most of yours.