It is important for each clinical trial to have a clear and comprehensive Statistical Analysis Plan (SAP) to support reproducibility. The SAP is intended to be a comprehensive document that contains a detailed and technical description of the principal features of the analysis outlined in the protocol. It also includes detailed procedures for executing the statistical analysis of the primary and secondary variables and other data. As compared to the protocol, the SAP should contain an in-depth description of the statistical methods to be used and a definition of the statistical output which will be included in the clinical study report.
The SAP is generally developed as a separate document, and written after the protocol has been finalised. Ideally, a biostatistician should develop the SAP with the help of the principal investigator and in alignment with the protocol. The SAP should be finalised prior to data analysis and before treatment un-blinding. Any changes between the methods in the protocol and analysis plan, should be explained in the SAP. The SAP must include a brief summary of the trial, which includes the aims and objectives, trial population, design of the trial, sample size calculations (including the justifications and assumptions), and the randomisation methods. The SAP must describe in detail the statistical methodology i.e. efficacy analysis, safety data analysis, reporting conventions, etc. However, the level of detail depends on the complexity of the trial.
Gamble et al., (2017), in their article recommend a minimum of 55 items that must be included while developing the SAP. These are divided into 6 major sections of:
- Title and Trial Registration (11 items and sub-items)
- Introduction (2 items)
- Study Methods (9 items and sub-items)
- Statistical Principles (8 items and sub-items)
- Trial Population (8 items and sub-items)
- Analysis (17 items and sub-items)
Institutes and agencies involved in the conduct of clinical trials must ensure that there are robust policies and SOPs to support the development of SAP. This template may be used to develop a SAP. However, the template must be customised as per the requirements of the trial.
It should be acknowledged that the SAP is a working document, because the statistical analysis may depend on unpredictable aspects of data, and, new analytical ideas may be developed during the course of the trial. The SAP once developed, should also be reviewed by at least one senior statistician who was not involved in its development.
References and Further reading:
- ICH Harmonised Tripartite Guideline – E9 Statistical Principles for Clinical Trials – current step 4 version, dated February 1998.
- Gamble C, Krishan A, Stocken D, et al., 2017, Guidelines for the Content of Statistical Analysis Plans in Clinical Trials. JAMA; 318(23):2337-2343, available online, (last accessed 26.02.2019), doi:10.1001/jama.2017.18556.
- Clinical Development Services Agency SOP for Writing a Statistical Analysis Plan, version 1.0, 2015.
- MRC Clinical Trials Unit at University College London SOP for Statistical Principles (version 5.0), 2017.