Performance Characterization of EDA and Its Potential to Improve Decision Making in Product Batch Release - Good Cascade Impactor Practices, AIM and EDA for Orally Inhaled Products - page 221

Fig. 8.53 OIP product w9k001 : plot showing paired outcomes for grouped stages (G) and
LPM / SPM ratio (R) for correct decisions represented by “0,” type I errors represented by “1,” and
type II errors represented by “2,” with percent of total for each category
Fig. 8.54 Plot illustrating the ability of stage group 2 to predict MMAD ; the width of the range of
acceptable MMAD values is only slightly larger than the uncertainty interval of the prediction
( MMAD range/uncertainty interval for prediction = 1.4 levels of discrimination per micron), indi-
cating a relatively low capability to discriminate different values of MMAD compared to the range
of acceptance limits
The assessment process described for IPAC-RS file code w9k001 was also
applied to the other products in the IPAC-RS database (Table 8.10 ).
The APSD characteristics of some of the products were more challenging than
others to successfully model and expand the range for assessment, especially for the
two products designated by file codes w9j901 and w9j601 for which there were only