Drexel University Home Pagewww.drexel.edu DREXEL UNIVERSITY LIBRARIES HOMEPAGE >>

iDEA: Drexel E-repository and Archives > Drexel Theses and Dissertations > Health Sciences Theses and Dissertations > Assessing Variabilities for Bioanalytical Methods

Please use this identifier to cite or link to this item: http://hdl.handle.net/1860/3589

Title: Assessing Variabilities for Bioanalytical Methods
Authors: Patel, Binal
Keywords: Public Health;Variability;Bioanalytical Methods
Issue Date: 21-Sep-2011
Abstract: Variability arises from many different sources. In the pharmaceutical industry, it is important that these variabilities are estimated and controlled in order to provide a good quality and safe product. The FDA has certain guidelines dealing with precision that need to be followed. This analysis tested five different methods from SAS® to assess which test should be used. These five tests include methods from PROC VARCOMP – ML, REML, TYPE1 and MIVQUE0 and from PROC MIXED – REML. This analysis simulated data to get an estimate of precision under true values. The simulation only dealt with two parameters where variability could arise and it had four possible values for each. The simulation also took into account the number of labs and number within, which also had four values each, in total created 252 different scenarios. Bias and mean square error (MSE) means were used to evaluate these five methods. Data was also obtained from Merck & Co. and CV (RSD) values were found to check precision of various tests. Generally speaking, the bias and MSE decrease as the sample size increases for all methods. MIVQUE0 was concluded to be the best test for checking variability. If the sample size is high, the testing method does not matter because all will produce the same or similar results, however, the pharmaceutical industry does not have the luxury of having many lab assays available. Therefore, these results are based on providing the best recommendation for the current situation and not an ideal situation.
URI: http://hdl.handle.net/1860/3589
Appears in Collections:Health Sciences Theses and Dissertations

Files in This Item:

File Description SizeFormat
Patel, Binal.pdf1.01 MBAdobe PDFView/Open
View Statistics

Items in iDEA are protected by copyright, with all rights reserved, unless otherwise indicated.


Valid XHTML 1.0! iDEA Software Copyright © 2002-2010  Duraspace - Feedback