The method.
A ledger is only as good as its rules. This page is the boring one: where numbers come from, what units we use, and when entries get re-reviewed.
Three kinds only
Peer-reviewed studies (Europe PMC / Crossref verified), national health-system publications (NHS England, WHO), and department-level audits submitted through the editor.
t CO₂e per year
Always tonnes of CO₂-equivalent, always per year, always per average English trust unless the entry says otherwise. No 'carbon saved' without a unit.
Quarterly, dated
Entries re-review quarterly; the REV date on every page is the last full pass. Grade changes are announced in the ledger-update email.
One trust's desflurane exit, in five numbers.
How a ledger entry is computed, end to end — using the measured programme numbers.
| Step | What you pull | Example value | Where it goes |
|---|---|---|---|
| 01 | 12 months of volatile agent purchases, by agent, from pharmacy ledgers | desflurane 412 bottles × 240 ml | Volume of each agent used |
| 02 | Density and bottle contents (agent-specific, from the datasheet) | 240 ml × 1.52 g/ml ≈ 365 g/bottle | Mass of agent vaporised |
| 03 | GWP₁₀₀ factor per agent (desflurane ≈ 2,640; sevoflurane ≈ 195; IPCC AR6 / EAUK) | agent mass × 2,640 | CO₂e baseline |
| 04 | Same arithmetic for the replacement agent at equal MAC-hours | sevoflurane equivalent ≈ 11.1 t CO₂e | CO₂e after the switch |
| 05 | Difference, stated per year, with the source line | ≈ 188 t CO₂e/yr saved | The ledger entry — grade A, published programme |
What A, B and C mean.
| Grade | Standard | Example |
|---|---|---|
| A | Peer-reviewed department-level measurement or RCT | Trust-wide desflurane exit, measured (BJA 2025) |
| B | Cohort study or published modelling | Volatile→TIVA cost-effectiveness model (Anaesthesia 2026) |
| C | Credible estimate; local audit needed before publication | MH community-visit clustering (this ledger, flagged) |
Add your department's number.
Baseline
Pull the dataset your action names — gas logs, prescribing reports, meter data. 12 months, no gaps.
Intervention
Run the action. Record what actually changed, including what didn't work — failed switches are data too.
Report
Same units as the ledger (t CO₂e/yr). Email the editor; entries gain a source line and a grade.