Evaluating public health measures
Difference-in-differences and synthetic control designs compare places that adopted a measure with comparable places that did not, and trends before adoption are examined before any effect is reported.
The Global Centre for Risk and Innovation (GCRI)
Causal inference asks whether an intervention or exposure actually changed health outcomes, compared with what would otherwise have happened. In this field it evaluates public health measures, service investments and environmental exposures where randomisation is often impossible, and states the assumptions each estimate rests on.
GCRI · Established · service level R0
Public health impact evaluation is the use of causal inference to judge whether an intervention or exposure changed health outcomes compared with what would otherwise have happened. It covers public health measures, service investments and environmental exposures where randomisation is often impossible, and states each estimate's assumptions.
The working screen behind this method: its structure, terms and states. Names are placeholders and values illustrative.
Health · Causal inference for public health · Run R-9605
| Main specification | 0.06 | Yes |
| Drop early adopters | 0.10 | Yes |
| Synthetic control | 0.11 | Yes |
Causal inference and impact evaluation develop methods to determine whether and how interventions or exposures cause outcomes. They separate genuine effects from correlation. In health and biosecurity, causal inference shows which risk reduction measures actually work. It supports allocating resources toward interventions with demonstrated effect across connected risks. The canonical description is on the STEM Commons.
Difference-in-differences and synthetic control designs compare places that adopted a measure with comparable places that did not, and trends before adoption are examined before any effect is reported.
Time-series and case-crossover designs relate daily exposure to outcomes in the same population, controlling for season and trend, to estimate lagged exposure-response functions.
New facilities, fee removals or outreach programmes are evaluated with stepped rollout, regression discontinuity or matching, so the effect on use and outcomes is separated from background change.
Each term as this page uses it. Where a standard-setting body defines the term, the wording follows that body.
| Term | Definition |
|---|---|
| Counterfactual | What would have happened without the intervention or exposure, which every causal estimate approximates. |
| Interrupted time series | A design that compares an outcome's trend before and after an intervention. |
| Confounding | Distortion of an estimated effect by a factor linked to both the exposure and the outcome. |
How far each division sits from a large city, with the roads and airports that carry patients and supplies, shown for the Amazon Basin as an example geography for this subject.
Data from the source register feeds the methods; calibration and an independent test gate the estimate and its limits; the decision stays with the competent authority.
Fig. 01Nodes carry their type, metrics and state; connectors show what feeds what; the gates are the validation steps; the timeline is one application. Design of the method's use, not a record of a run.
Datasets and records used by causal inference and impact evaluation on this platform, with the role each plays and the label its values carry. Sources are listed in the source register.
| Data | Role | Label |
|---|---|---|
| Routine health information systems | Outcomes before and after interventions | REPORTED |
| Policy and measure trackers | Timing and stringency of interventions | DECLARED |
| Daily mortality and admissions | Outcome series for exposure studies | OFFICIAL |
| Reanalysis and air quality fields | Exposure series | MODELLED |
Identification assumptions are probed through placebo tests, falsification checks and sensitivity analysis. Replication across settings tests external validity.
| Check | How it is done |
|---|---|
| Pre-trend and placebo tests | Effects are tested in periods and places where none should exist |
| Sensitivity to unmeasured confounding | Bounds or E-values state how strong hidden confounding would need to be to explain the result |
| Replication across settings | Effects are compared across countries or periods before they are generalised |
Unobserved confounding and spillovers can bias estimates. Effects found in one context may not transfer to another.
Economic Development stewards causal inference and impact evaluation for the Consortium and reviews how it is applied across platforms.
The current service level is R0: information and enquiries. An enquiry receives a reference and names the next step; applications, payments, compute allocations and registrations open only when their team, terms and review capacity exist.