1. Support a transition to net zero by 2050

A footprint assessment describes the health system as it is; a trajectory sets out how it could change. Trajectory development requires more detailed consideration of what is both technically and politically feasible. Modelling and policy should be co-developed, providing a quantitative basis for setting targets, comparing and prioritizing interventions, identifying investment needs, and monitoring progress.

Identification and quantification of interventions should be iterative, involving analysts and clinical, operational, finance and policy experts. This process helps identify opportunities, practical constraints and dependencies that may not be apparent from an emissions assessment alone.

Trajectories should inform policy as one part of a wider assessment of impacts including costs, investment requirements, clinical outcomes, equity, resilience and implementation feasibility.

A credible trajectory must identify the underlying changes required to achieve the modelled pathway and quantify their combined effects. A trajectory based solely on an assumed annual reduction rate, or constrained to meet a predefined target or “fair-share” trajectory of emissions, does nothing to explain how that goal can be reached.

A trajectory that reaches net zero provides a quantified demonstration of how the goal could be achieved under explicit assumptions. Other combinations of interventions, timing and deployment may offer equally credible routes. The health system should retain flexibility to pursue these alternatives as evidence, technologies and delivery conditions develop.

High-emitting and high-ambition health systems should develop quantified trajectories showing how net zero emissions can be reached by 2050 or earlier. Where the interventions currently quantified leave a gap to this goal, that gap must remain explicit and guide further analysis, policy development and investment.

2. Start from a complete health system inventory

Define the time horizon for modelling, key assessment years and calculation intervals.

Projections must start from a documented emissions assessment covering all material sources within a defined boundary, following WHO guidance on measuring GHG emissions in health systems (1) (Fig. 1). Reconcile any differences in coverage and methods between the target base year and the projection’s starting year.

Disaggregate the starting assessment into activity or expenditure volumes and corresponding emissions intensities. Retain the units, emission factors, data sources, coverage and calculation methods for each component, and reconcile these components to the assessment total. Where possible, preserve detail in projections: outputs can always be aggregated; restoring detail once it has been removed may be difficult.

Emissions calculations combine activity volumes, such as electricity consumed, products purchased or kilometres travelled, with corresponding emission factors expressing emissions per unit. Where physical quantities are unavailable, expenditure is usually used to represent purchasing. This distinction supports each stage of trajectory development:

• Growth trajectory: project changes in activity volumes while holding starting-year emission factors constant.

• Reference trajectory: adjust relevant factors to reflect developments such as grid, transport and supply chain decarbonization.

• Health system interventions: account for changes in activity volumes, emissions intensities, or both.

Changes can affect either component, or both. Efficient lighting reduces electricity consumption, while grid decarbonization reduces emissions per kilowatt-hour. Replacing gas heating with a heat pump reduces natural gas consumption, increases electricity demand, and improves heating efficiency. The combined calculation must represent heating demand, equipment efficiency, and fuel and electricity use to reflect the impacts of the intervention.

An unchanged spend-based factor cannot demonstrate the benefit of a lower-emission product purchased at the same price. Equally, a price saving alone does not demonstrate an emissions reduction. Represent these changes using appropriate activity data and comparable product- or process-specific emission factors, following the Toolkit’s guidance on hybrid emissions modelling.

3. Represent future health care needs

Projections must reflect future health care needs and planned changes in service provision, accounting for demographic changes, unmet needs, underserved populations, and improvements in access, coverage, quality and resilience.

Maintaining current levels of provision should not be the default assumption if doing so would perpetuate inadequate access to care. This is particularly important in countries where substantial expansion will be required to deliver universal health coverage.

A growth trajectory (often called a “do nothing” trajectory) usually holds emissions intensities constant at their starting-year values. It shows how emissions would change with health care demand before external decarbonization and additional health system interventions are introduced (Fig. 2).

Relevant drivers of growth include population growth, ageing, disease burden, changes in clinical practice and infrastructure development. Growth assumptions may draw on health and finance ministry expenditure forecasts, national population projections and plans for expanding services or facilities. Select sources relevant to the activity being projected, rather than applying a single growth rate across all emissions categories. Electricity demand, medicine purchases, patient travel and construction will not necessarily grow at the same rate.

For expenditure projections, maintain consistent currencies, price years and valuation bases. Do not add population or economic growth already incorporated in the forecasts, and document assumptions used to extend forecasts beyond their original time horizons.

4. Establish a reference scenario

Growth in clean energy production, electrification of key technologies and processes, and improvements in efficiency will all support reductions in the carbon intensity of health care. A reference scenario must combine projected health care growth with explicit assumptions about these external developments. It provides the comparison against which additional health system interventions are assessed (Fig. 3).

Develop a time series of emission factors reflecting these assumptions. Some factors may remain broadly constant, while others may decline as energy systems and production processes change. Assumptions must reflect the relevant sectors and countries, including those producing imported goods. Apply changes only to the components they affect (for example, national grid decarbonization does not imply an equivalent reduction across an entire supply chain).

Combine projected activity or expenditure volumes with the corresponding emission factors to calculate the reference trajectory. The difference from the growth trajectory shows the contribution of external changes. Represent each change once, whether through activity quantities or emission factors.

Potential sources for forecasts include national energy and transport projections, sectoral transition plans, and external forecasts and scenarios of future emissions and economic output (3–8). Supplier transition plans must be scrutinized for emissions coverage, implementation measures, timelines and evidence of delivery before reductions are assigned to purchases; a stated target alone does not establish the timing or scale of future reductions.

Scenarios can be used to test assumptions about possible external conditions and explore different rates of decarbonization. Current-policy scenarios reflect measures already implemented, while scenarios based on Nationally Determined Contributions assume delivery of pledged policies that may not yet be supported by effective implementation. Comparing these cases can show how strongly the health system’s trajectory depends on wider policy delivery.

Where detailed projections are unavailable, document and test any proxy assumptions or decisions to hold factors constant. When combining emissions and economic projections, use a consistent scenario and model, with matching geographical coverage and years. Check emissions coverage and the treatment of removals.

5. Quantify interventions and their combined effects

Identify the list of interventions that can be modelled against the reference scenario (Fig. 4). Distinguish them from changes already included in that scenario so that each effect is represented once. Identification and quantification should be iterative. Initial estimates can inform discussions about priorities and delivery, while those discussions refine the modelled coverage, timing and uptake.

For each intervention, specify the activity affected, the expected change in volumes or emissions intensity, implementation requirements and rollout timing, and the supporting evidence. Record existing uptake, eligible activity, achievable coverage and rollout rates.

Distinguish technical potential from assumed deployment, and make dependencies on financing, infrastructure or policy explicit. When defining deployment rates, consider the use of “no-scrappage” approaches where assets are replaced at the end of their useful life, or early replacement options. The calculation method must be able to represent the improvement being assessed.

Include any material emissions increases associated with interventions, even where they reduce emissions overall, such as increased electricity demand from electric vehicles, cleaning and transportation requirements for reusable products, or construction or equipment purchases. Moving or outsourcing activities must not create apparent reductions by removing their emissions from the calculation.

For each projected year, apply interventions to the reference scenario one at a time, retaining earlier changes and recalculating emissions after each addition. Calculating each intervention independently against the same reference scenario can overstate the combined reduction where their effects overlap.

Each intervention must therefore act on the activity that remains after preceding changes. For example, if some prescribing shifts from metered-dose inhalers (MDIs) to dry powder alternatives, savings from introducing lower-carbon MDIs must be counted only for those that are dispensed.

Calculate each intervention’s contribution as emissions before its addition minus emissions afterwards. These contributions sum to the combined reduction from the reference trajectory. Document the calculation order, because interactions can affect the savings allocated to individual interventions. Calculation order should not be interpreted as a ranking of policy priority or cost–effectiveness. Further guidance is provided in the GHG Protocol Scope 2 Guidance (9) on accounting for exported electricity without crediting avoided grid emissions against the health system’s footprint, and in Appendix B of the Policy and Action Standard (10) on assessing interactions between interventions.

Aggregate the results into a whole-system trajectory, checking that coverage matches the agreed boundary and that calculations reconcile to the starting assessment. Present absolute emissions over time, contributions from modelled changes and any remaining gap to interim and final targets.

When reporting, distinguish avoided emissions relative to the growth or reference scenario from reductions against the base year: a growing health system may avoid substantial emissions while its absolute emissions continue to rise. Relevant emissions intensity measures can be reported alongside absolute emissions, such as emissions per unit of health care activity or per unit of expenditure at constant prices. Define the denominator and explain changes in service mix, coverage and methods that affect interpretation. Falling intensity can demonstrate improvement while absolute emissions rise, but does not by itself demonstrate progress towards an absolute emissions target.

6. Support monitoring and allow for alternative pathways

Use alternative scenarios to explore how the pathway to a target could be revised in response to observed progress, new evidence or changing delivery conditions. Examine influential assumptions about health care growth, external decarbonization, intervention effects and deployment. Include adverse conditions, such as slower grid decarbonization or delayed implementation (Fig. 5).

Use observed activity and implementation data to test the assumptions underlying the projections. Compare observed changes with the modelled pathways, explain material differences, and update assumptions about health care growth, external decarbonization and intervention deployment. Changes to total emissions alone cannot establish that interventions are being delivered as planned.

Indicators should reflect the change being modelled. Possible examples include the proportion of lighting replaced by LEDs, electricity use per square metre of gross floor area in comparable facilities, the share of electricity consumption supplied by on-site renewable generation, the proportion of vehicles that are battery electric, the share of inhaler prescribing using clinically appropriate lower-emission options, or the proportion of relevant purchases meeting a defined supplier criterion. Interpret these alongside health care access, quality and other relevant service outcomes. NHS England’s Green Plan Guidance (11) provides 19 examples of metrics for tracking delivery, covering emissions, implementation and climate adaptation.

Some interventions may reduce emissions without producing a detectable change in the inventory, particularly where estimates use broad spend-based factors. Follow the Toolkit’s guidance on hybrid emissions modelling to understand how to integrate modelled reductions or report them separately. Compare original projections with updated estimates based on observed implementation and activity, using the differences to revise deployment assumptions and identify additional interventions.

Update projections routinely as new activity and implementation data become available. Undertake a comprehensive review of the model’s structure, methods and assumptions at least every five years, or sooner when new evidence, policies or implementation progress materially change the outlook. Preserve previous versions and explain revisions so that results can be compared with the original target and earlier projections.