How to run your ehrQL over multiple time periods

Sometimes you may need to run the same ehrQL separately over multiple months or years. This may be needed for instance to capture discrete periods of patient activity through time, or where there are multiple study index dates.

Other than making a separate ehrQL file for each period required, there are two supported options:

1. Pass parameters via the project.yaml

• Pass in a custom parameter to each ehrQL action (in the project.yaml) that allows us to use the same code with a different date each time. Each period is a separate action.
• This creates a full patient-level dataset for each period, allowing concatenation and further analysis.
• If you are running for many periods (e.g. every month for several years) you may need to write a script to generate all the actions required. It may be computationally expensive and require more storage space.
• Find more details in the how-to guide on re-using ehrQL with custom parameters.

2. The measures framework

• The measures framework is used to calculate quotients (i.e. a numerator divided by a denominator) for each period, which can be broken down by other variables (e.g. age bands).
• The periods ("intervals"), numerators, denominators and groups are all defined using ehrQL in a modified dataset definition.
• This approach does not generate a full patient-level dataset for each period - only a single measures output, showing the numerators and denominators for each measure for each period.
• This is ideal if you are only planning simple calculations on your results (e.g. trends/rates over time) or as an initial step to check that your overall results look sensible.
• This approach creates smaller files and should generally run faster than the Parameters approach.
• Find more details in the guide to the measures framework.