ICH Statistical Principles for clinical trials E9

ICH E9 has been developed by an ICH Expert Working Group to harmonise the statistical methodology applied to clinical trials submitted in marketing applications across Europe, Japan and the United States. It builds on earlier regional guidance and related ICH guidelines (such as E1A, E2A–E2C, E3–E8, E10, M1 and M3) and focuses on principles that should guide protocol design, trial conduct, data analysis and reporting under Good Clinical Practice. The guideline is intended for use across all trial phases, with particular emphasis on confirmatory trials that provide the main evidence of efficacy and safety.
The document is a structured PDF with seven main sections and a glossary. 

Section I (Introduction) sets out the background, purpose, scope and direction, explains the role of the trial statistician, and emphasises prospective planning of design and analysis in the protocol and statistical analysis plan. 
Section II (Considerations for overall clinical development) discusses the clinical development plan, confirmatory and exploratory trials, trial populations, primary and secondary variables (including composite, global assessment, multiple and surrogate variables), categorised variables and key design techniques to avoid bias such as blinding and randomisation.
Section III (Trial design considerations) covers design configurations (parallel, crossover and factorial designs), multicentre trials, types of comparison (superiority, equivalence, non inferiority and dose–response), group sequential designs, sample size determination and principles for data capture and processing. 
Section IV (Trial conduct considerations) addresses trial monitoring, interim analysis and early stopping, changes in inclusion and exclusion criteria, accrual rates, sample size adjustment and the role of independent data monitoring committees.
Section V (Data analysis considerations) sets out principles for prespecification of analysis, definition and use of analysis sets (full analysis set and per protocol set), handling of missing values and outliers, data transformations and derived variables, estimation, confidence intervals and hypothesis testing, multiplicity adjustments, subgroup and interaction analyses, covariates and integrity of data and software. 
Section VI (Evaluation of safety and tolerability) gives guidance on selection and collection of safety variables, analysis sets, statistical evaluation of safety data and integrated safety summaries. 
Section VII (Reporting) links to ICH E3 and covers blind review, finalisation of the statistical analysis plan, documentation of deviations from planned analyses, descriptive presentation, interpretation of unplanned analyses and the statistician’s role in clinical study reports and summary documents.

The glossary defines key terms such as intention to treat principle, full analysis set, per protocol set, confirmatory and exploratory trials, equivalence and non inferiority trials, surrogate variables, treatment effect, interim analysis, independent data monitoring committee, statistical analysis plan, frequentist and Bayesian methods. 

For investigator initiated clinical studies, ICH E9 provides a foundational, regulator endorsed framework for planning trial design (including choice of endpoints, type of comparison and sample size), specifying analysis sets and handling missing data and multiplicity when drafting statistical analysis plans and clinical study reports. A practical way to use this guidance document is to refer to Sections II–V during protocol development and to Sections V–VII when finalising the statistical analysis plan, conducting analyses and preparing trial reports and integrated summaries.