In this course, a variety of case studies will be used to present the most important aspects to consider for a sound determination of product shelf-life or stability. Starting from the assessment of the differences between shelf-life and stability studies, participants will learn for each type of study how to design efficient experiments to accurately determine the failure time of products. The issues discussed include the timepoint selection, how to handle destructive testing, the experiment size and the choice of samples.
Shelf-life data possess specific features so that their design and analysis require adapted statistical tools. Key elements in the study design are presented and survival techniques used to analyse life data are covered and illustrated. The principle of accelerated shelf-life testing (ASLT) along with the conditions for a successful use will be discussed.
- Difference between Shelf-Life & Stability Studies
- Introduction to Life Data
- Defining the Event of Interest that Causes the Failure of the Product
- Specific Features of Life Data and Impact on Statistical Analysis
- Typical Layout of Results
- Designing Efficient Studies
- Impact of End of Study Time, Censoring, Competing Risks, Destructive & Non-destructive Testing, Sample Variability
- Experimental Design: Sample Selection, Sample Size & Statistical Power, Dynamic Design
- Analysis of Stability Studies
- Analysing Time to Event Data to Estimate Shelf-Life
- Reasons Why Usual Analysis Methods Fail
- Kaplan-Meier Non-Parametric Survival Curve Estimation
- Parametric Survival Curve Estimation
- Overview of Semi-Parametric Survival Curve Estimation
- Why Accelerated Shelf-Life Testing (ASLT) Models?
- The tools available to define shelf-life in a specific context
- The specific features of life data
- The notion of competing risks and its impact on the results
- What statistical analysis techniques adapted to life data are available
- The modeling techniques especially designed for life data
- How to interpret results and what the scope of the results are
Recommended Duration: 1 day(s)