1月14日晚，EMA和FDA聯(lián)合發(fā)布《藥物研發(fā)中AI良好實踐指導原則》，為藥物全生命周期中利用AI開展證據(jù)生成與監(jiān)測工作提供了通用性指導意見。核心涵蓋10項指導意見。
Guiding principles of good AI practice in drug development
藥物研發(fā)中良好AI實踐指導原則
Artificial Intelligence (AI) has the potential to transform the way drugs (medicines) are developed and evaluated, ultimately improving healthcare. In this context, AI refers to system-level technologies used to generate or analyse evidence across the drug product life cycle, including nonclinical, clinical, post-marketing, and manufacturing phases.
人工智能（AI）有望改變藥物的研發(fā)和評估方式，最終改善醫(yī)療健康服務。在此背景下，AI指的是用于在藥品全生命周期中生成或分析證據(jù)的系統(tǒng)級技術，這些階段包括非臨床、臨床、上市后以及生產(chǎn)階段。
Drugs are authorised based on demonstrated quality, efficacy and safety, and when their benefits outweigh their risks. As new technologies emerge, including AI, it is essential that their use reinforces these requirements for the benefit and safety of patients.
藥物的獲批需基于已證實的質(zhì)量、有效性和安全性，且其獲益需大于風險。隨著以AI為代表的新技術不斷涌現(xiàn)，為了患者的利益和安全，這些技術的應用必須強化上述要求。
The use of AI throughout the drug product life cycle has increased significantly in recent years. The complex and dynamic processes involved in developing, deploying, using, and maintaining AI technologies benefit from careful management throughout the drug product life cycle to ensure outputs are accurate and reliable.
近年來，AI在藥品全生命周期中的應用顯著增加。AI技術的開發(fā)、部署、使用和維護涉及復雜且動態(tài)的流程，對這些流程在藥品全生命周期中進行審慎管理，有助于確保其輸出結果準確、可靠。
Among other innovations, AI technologies are anticipated to support a multi-faceted approach that promotes innovation, reduces time-to-market, strengthens regulatory excellence and pharmacovigilance, and decreases reliance on animal testing by improving the prediction of toxicity and efficacy in humans.
在諸多創(chuàng)新應用中，AI技術有望助力形成一種多維度的方法：推動創(chuàng)新、縮短產(chǎn)品上市時間、提升監(jiān)管效能與藥物警戒水平，同時通過提高對人體毒性和有效性的預測能力，減少對動物實驗的依賴。
This document outlines a common set of principles to inform, enhance, and promote the use of AI for generating evidence across all phases of the drug product life cycle.
本文件闡述了一套通用原則，旨在為AI在藥品全生命周期各階段生成證據(jù)的應用提供指導、予以完善并推動其發(fā)展。
These 10 guiding principles are intended to lay the foundation for developing good practice that addresses the unique nature of these technologies. They will also help cultivate future growth in this rapidly progressing field.
這 10 項指導原則旨在為制定應對此類技術獨特性的良好實踐奠定基礎，同時也將助力推動這一飛速發(fā)展領域的未來成長。
The 10 guiding principles identify areas where the international regulators, international standards organisations, and other collaborative bodies could work to advance good practice in drug development.
這 10 項指導原則明確了國際監(jiān)管機構、國際標準組織以及其他合作機構可著力推進藥物研發(fā)良好實踐的工作領域。
Areas of collaboration include research, creating educational tools and resources, international harmonisation, and consensus standards, which may help inform regulatory policies and regulatory guidelines in different jurisdictions, in line with applicable legal and regulatory frameworks.
合作領域包括開展研究、開發(fā)教育工具與資源、推進國際協(xié)調(diào)以及制定共識標準。這些工作可結合適用的法律法規(guī)框架，為不同司法管轄區(qū)的監(jiān)管政策和監(jiān)管指南提供參考。
As the use of AI in drug development evolves, so too must good practice and consensus standards. Strong partnerships with international public health partners will be crucial to empower stakeholders to advance responsible innovations in this area.
隨著AI在藥物研發(fā)中的應用不斷發(fā)展，相關良好實踐和共識標準也必須與時俱進。與國際公共衛(wèi)生合作伙伴建立緊密的合作關系，對于賦能利益相關方在該領域推進負責任的創(chuàng)新至關重要。
Thus, this initial collaborative work can inform our broader international engagements.
因此，這項初步的合作成果可為我們更廣泛的國際合作提供參考。
Principles
指導原則
1.Human-centric by design
設計以人類為中心
The development and use of AI technologies align with ethical and human-centric values.
AI技術的開發(fā)與使用需與倫理及以人為本的價值觀保持一致。
2.Risk-based approach
基于風險的方法
The development and use of AI technologies follow a risk-based approach with proportionate validation, risk mitigation, and oversight based on the context of use and determined model risk.
AI技術的開發(fā)與使用需遵循基于風險的方法，根據(jù)使用場景和已確定的模型風險，采取相應的驗證、風險緩解及監(jiān)督措施。
3.Adherence to standards
遵守標準
AI technologies adhere to relevant legal, ethical, technical, scientific, cybersecurity, and regulatory standards, including Good Practices (GxP).
AI技術需遵守相關的法律、倫理、技術、科學、網(wǎng)絡安全及監(jiān)管標準，包括良好實踐規(guī)范（GxP）。
4.Clear context of use
明確的使用場景
AI technologies have a well-defined context of use (role and scope for why it is being used).
AI技術需具備明確的使用場景（即其使用的角色與范圍）。
5.Multidisciplinary expertise
多學科專業(yè)知識
Multidisciplinary expertise covering both the AI technology and its context of use are integrated throughout the technology’s life cycle.
涵蓋AI技術及其使用場景的多學科專業(yè)知識，需貫穿該技術的整個生命周期。
6.Data governance and documentation
數(shù)據(jù)治理與文件記錄
Data source provenance, processing steps, and analytical decisions are documented in a detailed, traceable, and verifiable manner, in line with GxP requirements. Appropriate governance, including privacy and protection for sensitive data, is maintained throughout the technology’s life cycle.
數(shù)據(jù)來源的溯源、處理步驟及分析決策需以詳細、可追溯、可驗證的方式記錄，符合 GxP 要求。在技術的整個生命周期中，需持續(xù)維護恰當?shù)闹卫泶胧舾袛?shù)據(jù)的隱私與保護。
7.Model design and development practices
模型設計與開發(fā)實踐
The development of AI technologies follows best practices in model and system design and software engineering and leverages data that is fit-for-use, considering interpretability, explainability, and predictive performance. Good model and system development promotes transparency, reliability, generalisability, and robustness for AI technologies contributing to patient safety.
AI技術的開發(fā)需遵循模型、系統(tǒng)設計及軟件工程領域的最佳實踐，并使用適用的數(shù)據(jù)，同時考慮可解釋性、可說明性與預測性能。良好的模型與系統(tǒng)開發(fā)需提升人工智能技術的透明度、可靠性、通用性與穩(wěn)健性，以助力患者安全。
8.Risk-based performance assessment
基于風險的性能評估
Risk-based performance assessments evaluate the complete system including human-AI interactions, using fit-for-use data and metrics appropriate for the intended context of use, supported by validation of predictive performance through appropriately designed testing and evaluation methods.
基于風險的性能評估需對包含人機交互的完整系統(tǒng)進行評估，使用適用于預期使用場景的適用數(shù)據(jù)與指標，并通過恰當設計的測試與評估方法驗證預測性能。
9.Life cycle management
生命周期管理
Risk-based quality management systems are implemented throughout the AI technologies’ life cycles, including to support capturing, assessing, and addressing issues. The AI technologies undergo scheduled monitoring and periodic re-evaluation to ensure adequate performance (e.g., to address data drift).
基于風險的質(zhì)量管理體系需貫穿AI技術的整個生命周期，包括支持問題的捕獲、評估與解決。AI技術需接受定期監(jiān)測與周期性重新評估，以確保其性能達標（例如應對數(shù)據(jù)漂移）。
10.Clear, essential information
清晰的關鍵信息
Plain language is used to present clear, accessible, and contextually relevant information to the intended audience, including users and patients, regarding the AI technology’s context of use, performance, limitations, underlying data, updates, and interpretability or explainability.
需使用通俗易懂的語言，向目標受眾（包括用戶與患者）呈現(xiàn)清晰、易獲取且符合場景的信息，內(nèi)容涵蓋AI技術的使用場景、性能、局限性、底層數(shù)據(jù)、更新及可解釋性。
京公網(wǎng)安備11010802046783號
