Efficient Design & Analysis of Shelf-Life & Stability Studies

Creascience, Montreal
Feb 11, 2005

Aim

Participants will learn:

• Tools for defining shelf-life in a specific context
  
• Specific features of life data
  
• The notion of competing risks
  
• Statistical analysis techniques adapted to life data
  
• Interpretation and scope of results obtained

Target Audience

This session is intended for people who design or analyze shelf-life and stability studies.

The session is geared to the understanding of the key issues in ensuring an efficient set up of shelf-life and stability studies. It also covers powerful statistical tools especially designed to analyze life data

Prerequisites

This course introduces the important ideas in statistics and data analysis applied to the field of shelf-life and stability studies. It assumes that participants have no previous knowledge of statistics or that they have not used it for a long time.

Covered Topics

• General Overview of Shelf-Life
  
  
  
  • What is shelf-life?
    
  • Why study shelf-life?
    
  • Objectives of shelf-life studies
    
  • What are stability studies?
    
  • Difference between shelf-life and stability studies
  
  
• Introduction to Life Data
  
  
  
  • What are life data?
    
  • Defining the event of interest that causes the failure of the product
    
    
    
    • Distinguishing features of life data
    
    
    
    
    • Censoring
    
    
    
    
    • Truncation
  
  
• Designing Efficient Studies
  
  
  
  • Study Definition
    
    
    
    • Study objectives
      
    • Project feasibility
      
    • Study and control populations
    
    
  • Identification and selection of variables
    
    
    
    • Factors and outcome
      
    • Defining end of life of a sample
    
    
  • Other measurement issues
    
    
    
    • Censoring
      
    • Competing risks
      
    • Non destructive vs. destructive testing
    
    
  • Experimental design, sample size and statistical power
    
    
    
    • Overall sample size and allocation to sampling times
      
    • Issues with destructive testing
      
    • Specific problems in sensory evaluation
      
    • Dynamic designs
  
  
• Statistical Analysis of Life Data
  
  
  
  • Reasons why usual analysis methods (ANOVA, standard regression) fail
    
  • Goals of survival analysis
    
  • Methods for analyzing life data
    
    
    
    • Ways to represents results with survival curves
    
    
  • Non-parametric estimation: Kaplan - Meier
    
    
    
    • Goal
      
    • Assumptions
      
    • Results and interpretation
      
    • Comparing several curves
    
    
  • Parametric estimation
    
    
    
    • Goal
      
    • Assumptions
      
    • Different distributions: Weibull, gamma, exponential, etc.
      
    • Model fitting and assessment of goodness-of-fit
      
    • Results and interpretation
    
    
  • Semi-parametric estimation
    
    
    
    • Goal
      
    • Modeling factor effects
      
    • Results and interpretation
  
  
• A note on real-time vs. accelerated testing
  
  
  
  • Why use accelerated shelf-life testing (ASLT)?
    
  • The principle underlying ASLT
    
  • Key to successful ASLT
    
  • Problems in ASLT

Course Format

Different teaching tools are used: lectures, group discussions and interpretation of statistical software output

Software

This course is not software-specific

Course Duration

1 day

For inquiries about the course content, please contact Natalie Rodrigue at Creascience at 514 840 9220 x 27

Specific Conditions

N/A