Missed Charge Capture: The 3-8% of Revenue You Delivered But Never Billed

The most expensive money in a billing operation is the money you already earned and never asked for. It is not the denial you did not appeal. It is not the underpayment you did not notice. It is the service you delivered, documented in the chart, and never put on a claim at all. Industry audits put that pool at three to eight percent of gross patient revenue, and unlike every other leakage category, it never appears on an AR aging report because there is no claim on file for it to age against.

This is what a missed charge capture workflow surfaces. Not fraud, not upcoding, not aggressive coding. Just billing for the services already documented in the medical record, when someone or something between the visit and the claim let them fall off.

1. The pool nobody sees on the AR report

The math is honest but ugly. A 300 bed community hospital at $600 million in gross patient revenue running a middle-of-the-band five percent missed-charge rate is walking away from $30 million a year. A 20 provider medical group at $12 million in revenue at the same rate is walking away from $600,000. A 60 provider multi-specialty group at $45 million is walking away from $2.25 million. This is not a rounding error. It is a category of leakage bigger than the denial pool for most organizations, and it is systematically hidden because it lives before the claim exists.

Every dashboard your CFO looks at starts its clock at claim creation. AR aging days: measured from claim date. Denial rate: percent of claims denied. Days in AR: aging of open claims. None of these instruments ever look upstream at whether the claim should have been larger. They cannot. There is no data field for "the charge that never got dropped."

2. Why the EHR does not catch it

EHRs enforce documentation, not billing completeness. A chart can close with a signed note, complete assessment, finalized orders, and satisfied care plan and still have zero charges attached, because the billing module is a downstream consumer of orders, procedures, and encounter charges, not a validator against the clinical narrative.

Concrete example. A provider documents in the visit note: "Administered 40mg ketorolac IM in office for acute back pain." The note is complete. The chart closes. The provider moves on. What was supposed to happen in parallel is an order in the medication administration record that generates a billable line item for the injectable plus the administration code. What actually happens if the provider forgets that step is: the note documents the service, the MAR does not, and the claim goes out short by roughly forty dollars for the drug and its administration.

Multiply that pattern across every documented-but-not-ordered supply, every add-on code left off, every procedure the surgeon added to the case mid-op and forgot to charge for, every ancillary service captured in the modality system but never bridged to billing. The pool fills up one small omission at a time.

3. The six places charges actually get lost

Timing gaps, supply and injectable omissions, and stripped add-on codes typically account for 60 to 75 percent of first-cycle missed charge recovery. Focus a workflow on those three first; layer in professional-technical, ancillary bridges, and intra-op capture once the top three are clean.

4. The recovery ceiling for your book

The missed charge rate ranges 3 to 8 percent depending on specialty mix, EHR configuration, and how mature the existing charge capture discipline is. Cut those numbers in half or double them accordingly. Any provider organization that has never run a documentation-to-claim cross-check is at the higher end. Any organization with active CDI, dedicated ancillary bridges, and a chargemaster review cadence is at the lower end.

If you want a locality-adjusted reference for what your average claim value should be, use the RCM Pulse Report. It returns the Medicare-benchmarked allowed amount for your specialty's dominant CPTs in your specific locality. When your realized average per encounter is significantly below that benchmark on the same case mix, missed charges are usually a large part of the gap.

5. A five-step reconciliation workflow

Step 1: Extract structured data from every clinical source of truth

The chart, the medication administration record, the anesthesia record, the ops log, the ancillary systems (imaging, pathology, PT, respiratory), and any modality that captures a delivered service. Extract to a structured store with encounter id as the join key.

This is the biggest one-time investment. Every clinical system has an export path; not every system has a clean one. Getting these extracts stable is what makes everything downstream possible.

Step 2: Extract the claim file for the same encounters

Pull the claim as sent for every encounter in scope. Line-item level. Same encounter id join key. Now you have two datasets keyed on encounter: what the chart says was delivered, and what the claim says was billed.

Step 3: Run the cross-check

For each encounter, identify every billable service referenced in the clinical narrative. For each one, check whether a corresponding line item exists on the claim. If the note says "administered 40mg ketorolac IM" and no HCPCS J-code for the drug plus administration CPT appears on the claim, flag as a candidate missed charge.

This is where AI helps most. The clinical narrative is unstructured prose. Reliably mapping "administered 40mg ketorolac IM" to J1885 plus the 96372 administration code, and checking both against the claim, is a language-model task that used to require a coder to do encounter by encounter. Now it runs on every encounter within 24 to 72 hours of chart close.

Step 4: Human coder or CDI reviewer confirms and adds the charge

Every candidate is reviewed by a specific coder or CDI reviewer, not the AI alone. The reviewer confirms the documentation supports the charge (this is what keeps the workflow compliant), then adds the charge to the encounter and reopens the claim if it has already been sent.

The queue is owned and deadlined. Not "someone will get to it." Named reviewer, service-level agreement of 72 hours from flag to disposition, escalation path for the ones that are ambiguous.

Step 5: Root-cause feedback to the source

Every recovered charge tells you something about a specific provider, a specific service, a specific EHR workflow, or a specific ancillary bridge that keeps dropping charges. Aggregate by cause. If Dr. Chen keeps missing injectable administration codes, that is a training feedback item. If the imaging ancillary system keeps failing to bridge technical component charges, that is an engineering ticket. If the OR case entry does not have a field for intra-op unplanned procedures, that is a workflow redesign.

The feedback loop is what turns the audit into an ongoing monitor and eventually into a prevention system. Skip this step and next quarter's cross-check will surface the same categories again.

6. Where the pool is biggest by specialty

Infusion and procedural specialties are where the pool concentrates. Primary care yields less per encounter but frequently misses lower-dollar codes like vaccine administration and chronic care management that add up at volume. In every specialty, running the reconciliation on the top 20 CPTs by dollar volume is where the first pass earns its keep.

7. Charge capture vs upcoding, the compliance line

Missed charge capture recovery and upcoding are not the same thing, and any workflow worth running has to enforce the distinction hard.

Upcoding is billing a higher-level code than the documentation supports. It is a compliance violation. CMS and OIG treat it as fraud. It is the reason many organizations avoid systematic charge recovery work in the first place, out of legitimate fear that a poorly disciplined workflow will bill charges the chart cannot back up.

Missed charge capture is billing exactly the codes the documentation already supports, when those codes were left off the claim. Every recovered charge in a properly run workflow has to trace back to a specific chart entry, procedure note, medication administration record, or ancillary system record that already exists in the medical record before the recovery starts.

8. Five metrics to instrument

1. Missed charge detection rate. Percent of gross revenue flagged as candidate missed charges each week. Reveals the pool size and its trend by specialty.
2. Confirmation rate on flagged candidates. Percent of AI-flagged candidates that a human coder confirms as valid. Target: 80 percent or higher. Under 60 percent means the flagging model is too aggressive or the training data has drift.
3. Time-to-recovery. Days from chart close to added charge posted. Under 5 days keeps you inside every payer's initial filing window and avoids corrected-claim workflows.
4. Recovery per encounter. Average dollar recovery per encounter with at least one added charge. Trend up over time as coders get faster and the flagging gets sharper.
5. Repeat pattern index. Percent of flagged charges that trace to a small number of providers or workflows. Rising index means the root-cause feedback loop is not working; falling index means it is.

9. Charge capture software: what it does and where it falls short

Every EHR ships with something that looks like charge capture software baked in. Every hospital IT team has bolted at least one third-party charge capture tool on top of it. And still, the 3 to 8 percent pool persists. Understanding what the software is actually doing, and what it is not, is how you decide what to build or buy next.

What charge capture software does

Most charge capture solutions in market today do three things. They provide a mobile or tablet interface for providers to enter charges at the point of care (rounding apps for hospitalists, procedure logs for surgeons). They apply a chargemaster lookup so the entered service maps to the right CPT and HCPCS with the right modifier. And they push the captured charge into the billing system on a scheduled bridge.

Category leaders here are pMD, Ingenious Med, Epic's own charge capture module, and MEDITECH's Expanse. Prices range from $75 per provider per month at the light end to enterprise per-encounter contracts at hospital scale.

Where charge capture software falls short

Everything in the chart the provider did not personally enter into the charge capture tool. That is the hole. If a nurse administered the injectable, the drug is in the medication administration record, not in the provider's charge capture app. If the surgeon added an unplanned procedure and dictated it in the op note but did not open the charge capture screen after, it is documented and not billed. If the ancillary system captured the imaging read fee, the bridge to billing failed silently, and the technical component posted while the professional component did not. Charge capture software solves the entry-error problem for the provider. It does not solve the reconciliation problem between the full medical record and the claim.

Clinical documentation improvement (CDI) software

CDI software is the adjacent category that reads the clinical narrative and flags documentation gaps that would affect coding accuracy. Category leaders are 3M 360 Encompass, Iodine Software, Solventum, and Nuance CDE One. What CDI software does well: DRG assignment accuracy on the inpatient side, mortality risk adjustment, and severity of illness capture. What it does not do: cross-check the ambulatory chart against the outbound claim for missed line items. CDI is optimized for the DRG payment model, not for the fee-for-service line-item reconciliation that a missed charge audit actually needs.

A custom AI biller that does the reconciliation

The gap in the market is the encounter-level cross-check we described in section 5. Our AI Biller does this by reading the clinical narrative (visit note, procedure note, MAR, anesthesia record, op note, ancillary system outputs), identifying every billable service referenced, and comparing against what actually made it onto the claim. When something is on the chart and not on the claim, it surfaces the candidate charge to a licensed coder for confirmation before the added line item posts. Delivered in 30 days. Sits inside your PM environment. Costs less than one coder per quarter.

The right combination depends on your operation. Charge capture software solves point-of-care entry. CDI software solves DRG and severity capture. A custom AI biller solves the encounter-to-claim reconciliation gap that neither of the first two categories was built to solve. Most mature operations end up needing all three in the stack.

10. Where to start Monday morning

1. Pick your two biggest specialties by revenue. These are where the pool is largest and where the workflow investment earns back fastest. Infusion, orthopedics, dermatology, GI, and cardiology are usually candidates. Skip primary care for the first pass unless it dominates your book.
2. Pull the last 30 days of encounters for those specialties. Chart notes, MAR, ops log, and the corresponding claim file, all keyed on encounter id.
3. Score 100 encounters by hand. A coder reads the chart, lists every billable service referenced, and compares to the claim. Categorize the misses (supply, add-on, timing, split, ancillary, intra-op). This is a two-day effort and gives you the specific pattern map for your organization.
4. Extrapolate the pool. Missed charges per encounter times encounter volume times average claim value equals the annualized pool. Now you have a real number to justify the workflow build.
5. Pick the single biggest recurring pattern and fix its source. Not just recover the charges, fix why they keep getting dropped. Provider training, EHR order set, ancillary bridge, chargemaster entry. Recovery plus root-cause. Every cycle.

11. Frequently asked questions