How does a clinical encounter turn into a bill?

Documentation first, because everything else is derived from it

A clinician sees a patient and writes a note: the presenting problem, the history, the examination, the assessment, the plan, and what was done. That note is the legal record of the encounter and it is also the only evidence that anything billable happened. The rule that governs the whole chain is that if it is not documented it cannot be coded, and if it cannot be coded it cannot be billed. Not because of a technicality, but because a claim asserts that a specific service was provided for a specific condition, and the note is the proof.

Specificity is the currency. A note recording "diabetes" supports a much less specific code than one recording the type, whether it is controlled, and which complications are present, and the specific version usually supports higher and more defensible reimbursement while also being the clinically better record. This is why organisations invest in clinical documentation improvement: not to inflate anything, but because the note routinely understates what the clinician actually knew and did.

Coding translates the note into two different vocabularies

Coders, who are a trained professional group and not data-entry staff, read the documentation and assign codes from separate systems that answer separate questions.

Diagnosis codes answer why. In most of the world this is the WHO's ICD classification, used in the United States as ICD-10-CM for diagnoses. They establish medical necessity, which is the payer's test of whether the service was warranted for this patient's condition.

Procedure and service codes answer what was done. In US outpatient and professional billing this is CPT, maintained by the American Medical Association, extended by HCPCS Level II for supplies, drugs and services CPT does not cover. Inpatient procedures in the US are coded with ICD-10-PCS. Modifiers attach to a procedure code to qualify it: which side of the body, that two procedures were distinct rather than duplicated, that a service was reduced or discontinued. Modifiers exist because the alternative would be a combinatorial explosion of codes for every circumstance.

The two vocabularies then have to agree. A procedure whose diagnosis code does not support it will be denied for medical necessity even when both codes are individually correct, and reconciling that pairing is much of what a coder is doing.

Inpatient and outpatient reimbursement then diverge in a way worth knowing. Outpatient and professional billing prices individual services, so more services generally means more revenue. Inpatient reimbursement in most systems groups the whole admission into a case-mix category, in the US a diagnosis-related group, derived from the principal diagnosis, procedures, comorbidities and discharge status, and pays a fixed amount for that category. Under grouping, coding accuracy changes which group the stay falls into rather than adding line items, and the incentive shifts from volume to accurate capture of severity.

Charge capture is a separate step and it fails silently

Coding says what happened clinically. Charge capture says what goes on the bill, and the two are not the same activity.

Every billable item an organisation provides has an entry in the charge master: a code, a description, a department, and a gross price. Charge capture is the process by which the things done to a patient become charge lines against that account, and it comes from many places at once. Some is automatic, dropped by the system that performed the work when a laboratory result is verified, a dose is administered from the medication record, or an implant is scanned in theatre. Some depends on a person recording it: a procedure performed at the bedside, supplies used, time-based services where the duration determines the code.

The characteristic failure is not an error, it is an omission. A charge nobody captured produces no rejection, no warning and no queue entry. The account simply closes for less than it should have, and the organisation has no signal that anything happened. This is why charge reconciliation compares what the clinical systems say occurred against what was charged, department by department, and why late charges arriving after a claim has gone out are a routine and expensive part of the cycle. The direction of that error is worth stating plainly: over-capture gets caught by payers and auditors, under-capture gets caught by nobody.

The revenue cycle around it

The cycle starts before the encounter, not after it. At scheduling and registration the organisation verifies who the patient is, who the payer is, whether coverage is active, and whether the planned service needs authorisation. Errors introduced here, a mistyped member identifier or a plan that terminated last month, surface weeks later as denials, and correcting them then costs far more than checking them at the front desk.

After the encounter, coding and charge capture assemble the account. The claim is then priced, scrubbed against edits, and submitted, usually electronically in the standardised transaction formats mandated for the market. The payer adjudicates and returns a remittance advice explaining, line by line, what it paid, what it adjusted and why, using standard reason codes.

That remittance is where the money arithmetic becomes visible, and it contains the distinction candidates most often get wrong. The charged amount from the charge master is largely notional. What the payer will pay is the allowed amount set by the contract or fee schedule between that payer and that provider. The difference between charged and allowed is written off as a contractual adjustment, not chased. Of the allowed amount, the payer pays its share and the rest is patient responsibility: deductible, copayment, coinsurance, plus anything genuinely not covered. A second payer may then be billed under coordination of benefits before the patient is billed for the remainder.

Posting the payment closes the loop, and whatever is left is worked: denials appealed, underpayments identified by comparing the paid amount against the expected contractual amount, patient balances collected or written off. An account is not finished when the claim is submitted, which is the mental model most people arrive with.

Where the whole chain is decided

Every downstream problem in this pipeline traces back to documentation and front-end data, and almost none of it is fixable at the end. A denial for medical necessity is a documentation problem. A denial for eligibility is a registration problem. An underpayment is a contract-modelling problem. A missing charge is a capture problem. Organisations that treat revenue cycle as a collections function staff the back end heavily and rework the same errors forever; organisations that treat it as a documentation and data-quality function push the fix to where it is cheap. The single most useful thing to say about billing is that the bill is a rendering of the clinical record, so the quality ceiling of the bill is set by the note.