Date of Award
Spring 5-2024
Document Type
Dissertation
Degree Name
Doctor of Philosophy (PhD)
Department
Pharmaceutical Sciences
First Advisor
Enrique Seoane-Vazquez
Second Advisor
Jason Yamaki
Third Advisor
Beatriz Lopez Bermudez
Fourth Advisor
Rosa Rodriguez-Monguio
Fifth Advisor
Lawrence M. Brown
Sixth Advisor
Richard Beuttler
Abstract
Introduction: In our study, we evaluated the application of ChatGPT4, an artificial intelligence chatbot, in the pharmacy field, specifically the capability of ChatGPT4 to extract and analyze FDA drug information, conduct a cost-effectiveness analysis of two antibiotics, and identify and classify adverse drugs events (ADE). We compare the results of ChatGPT4 with those from manual searches and specialized software programs. Methods: The first study used ChatGPT4 to extract and assess the quality of regulatory information for antibiotics indicated for complicated urinary tract infections (cUTI) approved by the FDA in 2014-2023. The second study involved comparing the cost-effectiveness analysis of cefiderocol and imipenem/cilastatin indicated for cUTI using MS Excel, the software program TreeAge, and ChatGPT4. The third and last study evaluated ChatGPT4’s validity and reliability in identifying and classifying adverse drug events of fluoroquinolone antibiotics. Results: ChatGPT4 was efficient in collecting and evaluating regulatory data, with the Wilcoxon signed-rank test showing a significant difference favoring ChatGPT (p = 1.708 x10-05). ChatGTP4 conducted a CEA that included cost-effectiveness ratios (ICER) and sensitivity analysis. Additionally, ChatGPT4 had an 87.5% overall accuracy in identifying and classifying adverse drug events, identifying common and rare side effects but failing to identify some uncommon ADE. Conclusions: ChatGPT4 is an efficient tool for extracting and evaluating FDA drug information, and it can identify errors in data collection using manual methods. However, ChatGPT4 responses varied under different conditions, such as time of the day, file size, and question complexity. ChatGPT4 performs cost-effectiveness analysis, creating decision trees and calculating ICER, though its accuracy depends on the precision of input data and the way the questions are formulated. ChatGPT4 can identify and categorize most adverse drug events.
Creative Commons License
This work is licensed under a Creative Commons Attribution-Noncommercial-No Derivative Works 4.0 License.
Recommended Citation
Bukhari, K. Artificial Intelligence Applications for Optimizing Population and Clinical Drug Decision-Making Processes. [dissertation]. Irvine, CA: Chapman University; 2024. https://doi.org/10.36837/chapman.000594