The objective of this research is to compare potency estimation between the traditional regression approach and bioequivalence testing. Aquatic toxicity testing determines which level a chemical becomes harmful to the aquatic species. My comparison is conducted through simulation of a data set to generate p-values to indicate whether or not each test finds evidence of significant difference between a treatment and a control concentration. The critical concentration is a 0 if there was no evidence of a difference detected or is the concentration at which it has effects on the number of young. The question asked is, which statistical analysis method is best suited to determine the critical concentration level of a chemical associated with a specified level of inhibition of control response (birth of young)? This question is relevant to the field because if an analysis method deems a dangerous concentration of a chemical to be safe, it could be wrong and have detrimental effects on aquatic species. What I find is that TST tests resulted in the concerning levels of toxicity being far higher than the concentrations chosen by the RI25 and IC25. Based on previous reading, I expected the TST results to be more statistically conservative than the RIp or ICp which means TST tests result in higher concentrations of a drug being deemed safe. Moving forward, I could explore possible application of this approach on other types of data. I could also look into other response types like growth or survival in addition to the number of young. While this experience is not directly related to my hope of becoming an actuary, it has taught me that many doors will open to a variety of careers with the knowledge and skills I’ve gained in my degree and the statistical theory I’ve utilized in this project. [tags STEM,Data Science and Statistics,Potency Estimation]
Author(s): Chloe Wakim and Jing Zhang
Advisor: Jing Zhang, Statistics
You must be logged in to post a comment.