A Statistical Analysis Plan (SAP) is required for each trial to define guidelines for the analysis and presentation of data at various stages of the trial. Statistical analysis is a clearly defined section within the protocol and should be approved prior to the start of the trial. There may be several SAPs produced during the course of a trial (e.g. those based on the analyses required for the Independent Data Monitoring Committee meetings, interim reporting and final reporting of the trial).
The SAP should include the details of the planned statistical analyses associated with a clinical trial. It should contain the detailed requirements and parameters for the reporting of the results of the trial, the format and content of output reports, and the tests to support the robustness and sensitivity of the analysis conducted.
The SAP should include (as a minimum):
- How the outcome(s) will be measured
- Any transformations on the data likely to be required before analysis
- Appropriate statistical tests to be used to analyse the data
- How mising data will be accounted for in the analysis
- Whether statistical inference will be drawn and if any statistical adjustments for multiple comparisons will be performed
It is recommended that researchers seek statistical advice from a medical statistician at an early stage of trial development to ensure the sample size and analysis plan are appropriate to the project.
Warwick CTU has two SOPs concerned with statistics: SOP 8 - 'Statistical Considerations'; details the statistical input required when designing a clinical trial and SOP 21 - 'Statistical Analysis Plan'; explains the requirements of a SAP. A SAP writing template has been produced to attain consistency when producing the document.