Lesson 13 — Billing does not pass the audit

Controlled substances need batch-precise traceability. Billing data alone doesn’t carry it.

A controlled-substance audit (HMG / BetmG in Switzerland) requires traceability at the batch and serial level — for each dose dispensed, the hospital must be able to demonstrate provenance from manufacturer to patient. The audit treats this as binary: either the records can produce the chain, or they cannot.

Billing data records what was charged. It does not record what batch was charged. The two are different shapes of truth, and one of them passes the audit.

What billing gives you

Charge-event records typically carry:

• Patient encounter ID
• Material ID (drug or device)
• Quantity dispensed
• Date/time of charge
• CHF amount

What they do not typically carry:

• Batch number (manufacturer lot)
• Serial number (per-unit identifier for tracked items)
• Source-location reconciliation (which inventory bin the unit came from)

A hospital can answer “how many doses of fentanyl did we dispense last quarter” from billing data. It cannot answer “which batch of fentanyl produced these specific doses” — and the regulator wants the second answer.

What the audit actually needs

Controlled-substance regulation requires:

• A batch-precise inventory record — opening balance, receipts from supplier (with batch numbers), dispensations (with patient encounter linkage), wastage, closing balance. Every unit accounted for.
• A double-witness signature on dispensation events for Schedule I substances.
• A discrepancy investigation trail for any inventory variance, no matter how small.
• A multi-year retention of these records, audit-readable on demand.

Most hospitals maintain these records — but in a parallel system separate from the main billing flow. The billing data and the controlled-substance ledger are different stores.

What changed in the analysis

Two structural moves:

1. signal_controlled_substance_traceability stopped trying to reconcile billing alone against an audit prior. The signal now flags charge events for controlled substances and pairs them with a gap analysis — what record fields are present, what fields are missing, what would be needed to pass an audit on this evidence.
2. thesis_controlled_substance_gap became explicit about the question it’s answering: “are the controlled-substance charge events accompanied by audit-grade evidence in the available data?” — not “are the controlled substances accounted for”, which is a question billing data cannot answer regardless.

The Wisdom anchored the truth that the analytical layer can only surface what’s documented in the available data. If batch information lives in a parallel system the pack doesn’t ingest, the analysis cannot speak to it. The honest framing is “this is what we can see; this is what we cannot”.

What stays interesting

• A controlled-substance charge with batch number recorded in the billing data is a positive signal — the hospital has integrated its substance ledger with its billing flow. Worth surfacing as best practice.
• A controlled-substance charge with patient encounter linkage missing is a real audit risk on its own — separate from the batch question, and surfaceable from billing alone.
• A growing volume of controlled-substance charges without supporting fields can indicate a system change that broke the linkage. Trend matters.

Why this is in a Lessons journal

Because the gap between “what the data can show” and “what the audit requires” is one of the most important boundaries in clinical analytics. A signal that claims to demonstrate audit-readiness when the underlying data cannot do so is worse than no signal. The Wisdom anchored the honesty: name the gap, surface what is surface-able, defer the rest to the system that actually carries the truth. The Lesson is that some questions have data shapes that don’t fit the available data — and the right move is to say so.

Sources

• Wisdom: smebit_controlled_substance_hmg.yaml
• Provider: a hospital pharmacist with controlled-substance compliance responsibility
• Anchor signal: signal_controlled_substance_traceability
• Anchor thesis: thesis_controlled_substance_gap
• Date Wisdom captured: 2026-05-08