Medical Science Liaisons (MSL) | Medical Writers | Regulatory Affairs | Medical Affairs | Pharmaceutical Physicians | Clinical Investigators | Clinical Research Scientists | Statisticians transferring from other areas

This course is primarily concerned with statistical methodology for the design and analysis of clinical trials, planned and conducted within the pharmaceutical industry. Much of the methodology presented is in fact applicable on a broader basis and can be used in observational studies and in clinical trials outside of the pharmaceutical sector; nonetheless the primary context is clinical trials and pharmaceuticals.

The course is aimed at non-statisticians working in the Pharmaceutical Industry and will be suitable for physicians, investigators, clinical research scientists,medical science liaison (MSL), medical writers, regulatory personnel, statistical programmers and senior data managers. Statisticians moving from other areas of application outside of pharmaceuticals may also find the course useful in that it places the methods that they are familiar with in context for their new environment.

The course topics subdivided by section are detailed below.

- Placebo’s and Blinding
- Randomisation
- Signal and Noise
- Endpoint Types
- Superiority, Equivalence and Non-Inferiority

- Sample and Population
- Sample Statistics and Population Parameters
- The Normal Distribution
- Sampling and Standard Errors

- Confidence Intervals
- Hypothesis Testing
- The p-Value
- One-Tailed and Two-Tailed Tests

- The Principle of ITT
- Examples in the Application of ITT
- The Practice of ITT
- The Per-Protocol Set
- Withdrawals and Missing Data

- Sensitivity and Specificity
- Other Quantities
- Prevalence
- ROC Curves

- Statistical Testing
- Unpaired t-Test
- Non-Parametric Tests
- The Chi-Squared Test for Binary Data
- Odds Ratio and Relative Risk

- Type I and Type II Errors
- Power
- Calculating Sample Size
- Link Between p-Values and Confidence Intervals

- Inflation of the Type I Error
- How Multiplicity Can Arise?
- Bonferroni Correction
- Hochberg Correction
- Regulatory View on Multiplicity
- Importance of Pre-Planning
- Avoiding Adjustment including Hierarchical Testing
- Subgroup Testing
- Interim Analysis

- Time to Event and Censoring
- Kaplan-Meier Curves
- Logrank Test
- Hazard Rate and Hazard Ratio
- Independent Censoring

- Demonstrating Similarity
- Establishing Equivalence
- Establishing Non-Inferiority
- Choice of Delta – Non-Inferiority
- Switching between Non-Inferiority and Superiority

This advanced course follows on from the introductory Level 1 course and covers more advanced topics. The Level 1 course is aimed at non-statisticians working in the pharmaceutical industry and is structured to enable the understanding of statistical concepts and methods used in the design and analysis of clinical studies. The Level 2 course is aimed at the same target audience and has the same primary objective but covers a broader and more in-depth range of topics.

The advanced course is divided into 14 sections and development follows a logical step-by-step structure. It is assumed that participants will already have completed Level 1. Each of the 14 sections concludes with a workshop-exercise to assess understanding of the material covered and a 70% mark (based on at most 2 attempts) is required on each of those exercises to receive a certificate of completion for the course.

The course topics subdivided by section are detailed below.

- Stratified Analysis for Continuous Endpoints
- Cochran-Mantel-Haenszel Test for Binary Endpoints
- Stratified Logrank Test for Time-to-Event Endpoints
- Limitations and Extensions

- Simple Linear Regression
- Multiple Linear Regression
- Logistic Regression
- Proportional Hazards Regression
- Negative Binomial Regression
- Univariate versus Multivariate
- Correlation

- Analysis of Covariance (ANCOVA)
- Regression Towards the Mean
- Least Squares Mean
- Treatment x Covariate Interactions
- Prognostic versus Predictive
- Modelling
- Logistic Regression
- Negative Binomial Regression
- Cox Proportional Hazards Model

- Avoiding Missing Data
- Classifying Missing Data
- Missing Completely at Random
- Missing at Random
- Missing Not at Random
- Multiple Imputation
- Pattern Mixture Models
- MMRM for Repeated Measures

- Estimands:
- What are they?
- Why are they interesting?
- Estimand Strategies
- Treatment Policy
- Hypothetical
- Composite
- Principal Stratum
- While On-Treatment
- Case Study in Diabetes

- Recap Level 1, Section 8
- Golden Rule
- Reporting p-Values
- Combining Hierarchical and α-Splitting Methodologies
- Closed Testing with Case Studies
- Revisiting Workshop-Exercise Level 1, Section 8
- Exploratory Analyses
- Revisiting Interim Analyses

- Frequentist versus Bayesian Methodologies
- Bayes Rule
- Prior and Posterior Distributions
- Case Study – Ovarian Cancer
- Practical Applications

- Objectives of Meta-Analysis
- General Aspects of Conduct
- Fixed and Random Effects
- Meta-Analysis versus Pooling
- Assessing Heterogeneity
- Publication Bias and Funnel Plots

- Network Meta-Analysis (NMA) for Indirect Comparisons
- Case Study – Psoriasis
- Network Geometry
- Bayesian Rank Analysis
- Surface Under the Cumulative Ranking (SUCRA) Curve
- Indirect Comparisons and Effect Modifiers
- NMA versus Meta-Analysis

- Recap, Level 1, Section 9
- Proportional Hazards Assumption
- Restricted Mean Survival Time
- Accounting for Cross-Over
- Cumulative Incidence Functions

- Observational Studies and Forms of Bias
- Regulatory Acceptance
- Real-World Evidence
- Selection Bias in Observational Studies
- Propensity Score Matching
- Case Study
- Stratification and Inverse Propensity Score Weighting

- Primary Endpoint Safety
- Safety Data at the Trial level
- Incidence Rates
- Tables and Graphs for Reporting
- Data and Safety Monitoring Boards (DSMBs)
- Signal Detection
- Proportional Reporting Ratios
- Bayesian Neural Networks

- Recap, Level 1, Section 10
- p-Values for Non-Inferiority and Equivalence
- Assay Sensitivity
- Sample Size – Non-Inferiority
- Analysis Sets
- Bioequivalence
- Biosimilarity
- Case Study– Neovascular AMD

- Definition of Adaptive Designs
- Restricted and Flexible Adaptations
- Type I Error Control using a Combination Test
- Operational Bias
- Increasing Sample Size
- Seamless Phase I/Phase II
- Changing the Primary Endpoint
- Enrichment

*'Access to the 2nd edition of Richard kay's book 'Statistical Thinking for Non-Statisticians in Drug Regulation' in eBook format will be included as part of the Advanced level package at no extra cost. The book published by Wiley is widely recognised as a leading text in the field.'*