Nuffield Trust 5 August 2015
- When introducing new ways of delivering health care services it is important to understand what works, why it works, and what impact the new services are having on costs and outcomes. One of the recurrent problems when evaluating the impact of new care models on outcomes is how to know ‘what would have happened anyway’.
- Retrospective matched control analysis measures the impact of an intervention in terms of differences in the outcome relative to a matched control group.
- This guide outlines ten steps towards retrospective matching to evaluate new health and care service models of interest to those involved in evaluation at a local, regional or national level.
- Clarify the aims of the service and the evaluation
- Decide on the number of people needed to demonstrate an effect
- Ensure permission is granted to access person-level datasets
- Ensure there are data on who received the new service, and some
information about the service received
- Identify the potential control population
- Create longitudinal patient-level histories of service use
- Identify matched controls
- Monitor outcome variables for those receiving the new service and
matched controls
- Undertake summative analysis
- Continuously monitor
