以人工智慧推動藥品真實世界數據之商業模式：生技醫藥產業案例分析

Abstract

隨著醫療科技進步與資料數位化普及，真實世界數據（Real-World Data；RWD）在藥品開發、上市後監測與臨床決策中的應用價值日益受到重視。相較於傳統隨機對照試驗，RWD更貼近實際臨床場域與多樣化病患族群。然而，RWD本身具有異質性高及結構不一的特性，使其在轉化為真實世界證據（Real-World Evidence；RWE）的過程中面臨多重挑戰。人工智慧（Artificial Intelligence；AI）技術被視為加速此轉化的關鍵工具，尤其在資料清理、萃取、整合與預測分析等環節具潛力。然而，AI如何賦能RWD於生技醫藥產業，進而塑造可持續的商業模式與價值共創機制，仍缺乏系統性探討。 本研究採取質性研究方法，透過多重個案分析與專家訪談，剖析AI技術在RWE商業化過程中的角色與運作邏輯。案例研究選取五家具代表性的醫療科技企業—IQVIA、Flatiron Health、Aetion、TriNetX與Oncoshot，涵蓋不同規模、區域與技術特性。資料來源包含官方網站、年報、商業報導與學術文獻。分析面向包括商業模式共通性與異質性、資料處理策略、AI應用比較及生態系角色定位。另納入來自產業的專家觀點，補充實務場域對AI應用、資料治理與跨部門合作的觀察。 研究結果顯示，AI在RWD商業化過程中不僅是技術輔助工具，更深刻影響企業的資料治理架構與跨域合作。成功企業普遍具備以下特徵：（1）清晰定位資料處理面向，建立差異化競爭優勢；（2）問題導向與價值共創，聚焦產業痛點並與生態夥伴共創價值；（3）以合規為優先的AI演進，並採分階段導入策略；（4）多元目標客群開拓，延伸服務對象涵蓋藥廠、醫院、政府與學研機構；（5）多元創新與營收模式，結合資料授權、平台訂閱與解決方案等模組化設計；（6）具備整合生態系與平台主導潛力，逐步由利基型服務商轉為協作推動者。 綜合而言，AI已成為推動RWD價值鏈升級的系統性力量。對台灣而言，未來可從資料互通性標準起步，結合策略導向設計，建構具在地適應性的AI-RWD商業生態系。

With the advancement of medical technology and data digitization, Real-World Data（RWD）has become increasingly valuable in drug development, post-marketing surveillance, and clinical decision-making. Compared to randomized controlled trials, RWD better reflects real-world settings and diverse patient groups. However, its heterogeneity and lack of standardization pose challenges in transforming RWD into Real-World Evidence（RWE）. AI is seen as a key enabler, especially in data cleaning and analysis. Yet, how AI supports sustainable business models in the RWE biopharmaceutical industry remains underexplored. This study uses a qualitative approach, combining multiple case studies and expert interviews to examine how AI drives RWE commercialization. Five companies—IQVIA, Flatiron Health, Aetion, TriNetX, and Oncoshot—were selected to reflect diverse sizes, regions, and technologies. Data sources include websites, reports, industry news, and academic literature. The analysis focuses on business models, data strategies, AI applications, and ecosystem roles. Expert insights complement observations on data governance and cross-sector collaboration. Findings show that AI not only serves as a technical tool but also shapes data infrastructure and collaboration mechanisms. Success factors include: (1) clear positioning of data capabilities; (2) problem-driven value creation with ecosystem partners; (3) compliance-first, phased AI deployment; (4) diversified customer targets; (5) flexible revenue models including data licensing and modular services; and (6) potential to evolve into ecosystem orchestrators. In conclusion, AI is a systemic force advancing the RWD value chain. For Taiwan, building interoperability standards and strategic collaboration will be key to developing a locally adaptive AI-RWD ecosystem.