Trend Analysis on Stability
Able to detect If
A batch is out of trend with respect to historical batches
One or more observation is out of trend within a batch
Possible Models
Simple Linear Regression Model
Random Coefficients Regression Model
Data Selection
Minimum 3 Batches
Minimum 4 Timepoints
Identifying OOT
3 or 4 observations
2 timepoints may be sufficient to highlight an OOE
Stability data from historic product lots to calculate trends
Single real time storage condition
More data may be needed due to high method variability, lot to lot variability
Trend assessment to be based on all available timepoints for the selected lots
Data values with one significant figure more than the specification
Definitions
OOS
All reportable results not meeting established specifications. A confirmed reportable value that is outside the acceptable specification criteria as stated in the product or analyte specification (e.g. CoAs, USP).
OOT
A stability result that does not follow the expected trend, either in
comparison with other stability batches or with respect to previous results collected during a stability study
OOE
OOE results for the purposes of this document are anomalous, unexpected or unusual findings, which have not been classified as either OOS or OOT, for both qualitative and quantitative testing.
Outlier
A single result, which is markedly different from others in a series as confirmed by statistical evaluation.
Used to understand the relationship between stability data and time - can be used to predict an expected stability result over time
More flexible model that allow for multivariate inputs and covariates
Adv: Parameters in the model easily interpreted
Adv: Few restrictions on the design space
Disadv: The error terms assume constant variance
Disadv: The degradation rate is assumed to be linear
Process flow
Compare a Stability test result with the trend limits
If the stability test result is out of trend limits, investigate
Define level of investigation
Attention to process/manufacturing changes
Should not immediately require a new stability timepoint
Evaluate if a value is significantly different from T0
If there are no trend limits, evaluate against historic stability data
Periodic reassessment of the defined limits
Assess prediction intervals according to a defined interval (annual or every 3 years)
Assessment of trend limits canbe used to evaluate site or post-change differences