以數據分析為基礎的製藥研發流程 : 以治療鐵沉積學名藥為例

NCUIR > School of Management at National Central University > Executive Master of Business Administration > Electronic Thesis & Dissertation > Item 987654321/92222

Please use this identifier to cite or link to this item: http://ir.lib.ncu.edu.tw/handle/987654321/92222

Title:	以數據分析為基礎的製藥研發流程 : 以治療鐵沉積學名藥為例
Authors:	林亭妏;lin, Ting-Wen
Contributors:	企業管理學系在職專班
Keywords:	藥物研發;數據方法分析;統計方法與應用;Bootstrap;Drug development;Data method analysis;Statistical methods and applications
Date:	2023-07-08
Issue Date:	2024-09-19 15:23:42 (UTC+8)
Publisher:	國立中央大學
Abstract:	在高度競爭且邁向全球化的製藥行業中，發展基於數據為基礎的分析方法應用和數據分析等技術已成為各大藥廠著重的目標項目，製藥研發是一個複雜而耗時的過程，傳統上以實驗室實驗為主，然而近年來，基於數據為基礎的方法在製藥研發中得到了廣泛的關注和應用，基於數據為基礎的方法包括數據分析、機器學習和人工智能等技術，能夠將大量的數據轉化為有價值的資訊，並對藥物的發現、優化和評估提供重要的數據支持。本研究以個案公司研發產品-治療鐵沉積學名藥為例，在探討基於數據為基礎的方法中數據分析方法在製藥研發過程中的應用和價值，內文中介紹了製藥研發基本流程及探討了數據方法分析產生的結果差異，並針對目前藥廠常用的統計方法 Bootstrap 的設計與應用做了介紹，再根據個案公司提供的研究數據執行傳統統計學與統計方法Bootstrap 進行數據處理的分析及結果比較。在藥物研究中執行體外試驗-溶離曲線比對試驗 (Dissolution profile comparison) 主要目的皆是為了預測生物等效性 (Bioequivalence, BE) 試驗中體內試驗-藥物的藥動學（Pharmacokinetics）試驗結果，基於研究結果顯示統計方法 Bootstrap 應用，可以補足傳統統計方法因樣本數量較小的情況下執行數據處理所產生較大的變異係數(Coefficient of variation, CV) 問題，藉由這樣的統計方式，在樣本數據不足、非常態分部或存在極端值得情況下皆能補足因變異係數 (Coefficient of variation, CV) 高所帶來的風險，除了使數據可靠度增加，更能有效解決樣本數據和母體數據之間存在的差異性的問題，成為了許多藥物研發公司以及官方單位所接受的統計方法，且合適的數據分析方法能有效的提高藥物研發過程、增加市場的競爭力及降低失敗風險。 ;In the highly competitive and increasingly globalized pharmaceutical industry, the development of data-based analytical methods and the application of data analysis techniques have become prioritized objectives for major pharmaceutical companies. Pharmaceutical R&D is a complex and time-consuming process traditionally focused on laboratory experiments. However, in recent years, data-based approaches have gained significant attention and application in pharmaceutical R&D. These data-based methods include data analysis, machine learning, and artificial intelligence, among other technologies, which can transform large amounts of data into valuable insights and provide essential data support for drug discovery, optimization, and evaluation. This study uses a case company′s research product, a pharmaceutical drug for treating iron deposition, as an example to explore the application and value of data analysis methods based on data-driven approaches in the pharmaceutical research and development process. The study introduces the basic process of pharmaceutical R&D and discusses the differences in results generated by data analysis methods. It also introduces the design and application of Bootstrap, a commonly used statistical software in pharmaceutical companies. Furthermore, the study conducts data processing, analysis, and result comparison using both traditional statistical methods and Bootstrap statistical software based on the research data provided by the case company. In pharmaceutical research, conducting in vitro dissolution profile comparison tests primarily aims to predict the pharmacokinetics outcomes of drugs in vivo bioequivalence trials. Based on research findings, the application of the statistical software Bootstrap has been shown to address the issue of large coefficients of variation resulting from data processing using traditional statistical methods, especially in cases with small sample sizes, non-normal distribution, or the presence of extreme values. This statistical approach not only increases the reliability of the data but also effectively resolves the differences between sample data and population data. As a result, it has become an accepted statistical method by many pharmaceutical research companies and regulatory authorities. Utilizing appropriate data analysis methods can significantly enhance the drug development process, increase market competitiveness, and reduce the risk of failure
Appears in Collections:	[Executive Master of Business Administration] Electronic Thesis & Dissertation

Files in This Item:

File	Description	Size	Format
index.html		0Kb	HTML	28	View/Open

社群 sharing

Loading...