Supporting service
Impact and programme evaluation
Quasi-experimental and econometric evaluation of health programmes and financing reforms, household survey analysis and mixed-methods evaluation.
The problem
Governments and funders need to know whether a programme or reform — a new insurance scheme, a payment reform, a service-delivery model — actually changed access, use, quality or financial protection. Randomised trials are often impossible, and routine data are rarely analysed with the rigour needed to separate the effect of a reform from everything else that changed at the same time.
What we deliver
Quasi-experimental evaluation
Difference-in-differences, matching, regression discontinuity, synthetic control and interrupted time series designs.
Evaluation of financing reforms
Effects of insurance, purchasing and payment reforms on use of services, out-of-pocket spending and financial protection.
Household survey analysis
Analysis of national household, expenditure and demographic and health surveys, and design of evaluation survey modules.
Mixed-methods evaluation
Combining quantitative estimates with qualitative work to explain how and why a programme worked.
Methods
- Difference-in-differences and event-study designs
- Propensity score matching and weighting
- Regression discontinuity and instrumental variables
- Interrupted time series
- Survey data analysis
- Mixed methods
How we work
- 01
Define the evaluation question
Agree the outcomes that matter, the comparison that is credible and the data that exist or must be collected.
- 02
Design the evaluation
Choose the strongest feasible design and write a pre-specified analysis plan.
- 03
Collect or assemble data
Use existing survey and routine data where possible; partner with established field teams where new data collection is needed.
- 04
Analyse and interpret
Estimate effects with appropriate robustness checks and explain results in plain language for decision-makers.
Every engagement follows our reproducible pipeline with three expert sign-off gates.
Typical deliverables
- Evaluation design and pre-specified analysis plan
- Analysis code and documented dataset
- Evaluation report and policy brief
- Presentation of findings to decision-makers
Free guides, templates and tools
Guides
Frequently asked questions
Can you evaluate a programme that has already started?
Often, yes. Quasi-experimental designs such as difference-in-differences or interrupted time series can use existing survey and routine data when a randomised design is no longer possible.
Do you collect data?
We design surveys and analysis plans, and partner with established field teams where new data collection is needed.
How long does an impact evaluation take?
Analysis of existing data typically takes 6–10 weeks. Prospective evaluations follow the programme’s timeline.
Discuss your question with us
Tell us what you need and we will reply within one business day with next steps — or book a 20-minute call.