Clinical Trials in the Era of Real-World Evidence

Jie Chen is Chairman and Chief Scientific Officer at Taimei Intelligence Biopharma R&D, Shanghai, and a Visiting Member of the Center for Innovative Study Design at Stanford University. His extensive experience spans biopharmaceutical research, development, real-world evidence studies, and regulatory submissions. Naitee Ting is Vice President at StatsVita and serves as Adjunct Professor of Biostatistics at the University of Connecticut, Columbia University, and Colorado State University. He brings decades of experience in clinical trial methodology and pharmaceutical development. Feinan Lu is Director and Head of Biometrics at Impact Therapeutics, Shanghai. She specializes in the statistical methodology underlying clinical trials and real-world evidence studies supporting regulatory decision-making in drug development.

• Preface xixAcknowledgments xxiiiAcronyms xxv1 Introduction 11.1 Medical Product Development Pathway 11.2 Development of Evidence-Based Medicine 31.3 The 21st Century Cures Act 41.4 Regulatory Guidance 51.4.1 Guidance and related documents by the FDA 51.4.2 Guidance and related documents by the EMA 61.4.3 Guidance and related documents by the NMPA 61.4.4 Guidance and related documents by the ICH 71.4.5 Guidance and related documents by the MHLW of Japan 71.4.6 Guidance and related documents by other regulatory agencies 81.5 Discussion and Summary 91.6 Supplements 112 A Brief History and Critical Components of Clinical Trials 132.1 Lady Tasting Tea 142.2 Alpha 152.3 Permutation Test 172.4 Selection of Control 212.5 Parallel vs crossover trials 252.6 Blinding 252.7 Process of a Clinical Trial 272.7.1 Clinical trial design stage 272.7.2 Trial monitoring or study conduct stage 312.7.3 Analysis and reporting stage 322.8 Supplements 333 Clinical Development Process of a New Drug 353.1 Clinical Development Plan 353.1.1 Clinical development phases 363.1.2 Documents for regulatory submission 373.1.3 Target product profile 383.1.4 Oncology drug development 393.2 Phase I Clinical Trials 403.2.1 Pharmacokinetics (PK) 403.2.2 Tolerability 433.3 Phase II Clinical Trials 463.3.1 Proof of Concept (PoC) trials 483.3.2 Dose-ranging trials 503.3.3 Combined PoC and dose ranging 523.3.4 Example—A Phase II osteoarthritis (OA) trial 543.4 Phase III Trials 573.4.1 Example—Late developed AE (Proteinuria) 613.5 New Drug Application (NDA) 633.6 Phase IV Studies 643.7 Supplements 654 Design Considerations for Phase III Confirmatory Trials 674.1 Drug Label 674.2 Selection of Primary Indication 684.3 Multi-Regional Clinical Trials (MRCT) 704.4 Selection of Endpoint(s) 724.5 Selection of Control 744.6 Selection of Dose(s) 784.6.1 Drug efficacy considerations 794.6.2 Drug safety concerns 804.7 Additional Considerations 814.7.1 Vaccines or drugs for prevention 834.7.2 Drugs treating non-life-threatening diseases 844.7.3 Drugs treating life-threatening diseases 854.8 Supplementary 865 Regulatory Submission and Approval 895.1 International Council of Harmonisation (ICH) 905.2 Prescription Drug User Fee Act (PDUFA) 915.3 Pre-Submission Meetings 935.3.1 End of phase II meeting 935.3.2 Pre-NDA/BLA meeting 945.4 Common Technical Documents and Submission 945.5 Advisory Committee Meetings 975.6 Supplements 1006 Overview on Use of RWD and RWE in Regulatory Setting 1036.1 Categories of RWD and External Data 1036.2 Trial Design and Conduct 1046.2.1 Evidence synthesis 1046.2.2 Use of external data in trial design and conduct 1076.2.3 External data as external controls in trial design 1076.3 Using RWD and RWE to Support Product Approval 1086.3.1 Serving as external controls 1096.3.2 Supporting population extrapolation 1096.3.3 Supporting label expansion to different indications 1126.3.4 Serving as sole evidence 1126.3.5 Supporting other label modifications 1136.4 PMRs and PMCs 1146.5 Discussion and Summary 1156.6 Supplements 1157 Single-Arm Trials 1197.1 Necessary Conditions 1207.1.1 Life-threatening or serious conditions with no efficacious treatments 1207.1.2 Rare diseases 1207.2 Desirable Conditions 1217.2.1 Well-understood natural history of the rare diseases 1217.2.2 Well-understood mechanism of action of the drug 1217.2.3 Adequate dose optimization 1227.2.4 Substantial treatment effects 1247.2.5 Translation of SEs into clinical benefits 1257.2.6 Favorable benefit-risk profile 1267.2.7 Totality of evidence 1277.2.8 Planning/initiation of confirmatory trials 1287.3 Other Considerations 1287.3.1 Well-defined estimands 1287.3.2 Adaptive designs 1297.3.3 External controls 1297.3.4 Communication with regulatory agencies 1307.4 Examples 1317.4.1 Blinatumomab for relapsed/refractory B-cell acute lym- phoblastic leukemia 1317.4.2 Retifanlimab for locally advanced or metastatic squa- mous carcinoma of the anal canal 1337.4.3 PI3K inhibitor for hematologic malignancies 1347.5 Conclusion and Summary 1367.6 Supplements 1368 Externally Controlled Trials 1398.1 Types of External Controls 1398.1.1 Historical control 1408.1.2 Contemporaneous control 1418.1.3 Non-concurrent control 1428.1.4 Historical-contemporaneous control 1428.1.5 Synthetic control 1438.1.6 Hybrid control 1438.1.7 Virtual control 1448.2 External Data as a Sole Control Group 1458.2.1 Selection of covariates 1458.2.2 Choice of distance metrics 1468.2.3 Matching 1468.3 External data to augment concurrent controls in RCTs 1488.4 Assessment of fit-for-use external data 1498.5 General Considerations—External Controls 1518.6 Targeted-Learning Roadmap 1538.6.1 Step 0: Clearly defined research question 1548.6.2 Step 1: Observed data and its generating mechanism 1548.6.3 Step 2: Statistical model and targeted statistical esti- mand (parameter) 1558.6.4 Step 3: Causal model and causal estimand 1558.6.5 Step 4: Statistical estimand versus causal identifiability 1568.6.6 Step 5: Estimation of statistical estimand 1578.6.7 Step 6: Result interpretation and sensitivity analysis 1588.7 Discussion and Summary 1598.8 Supplements 1619 Master Protocols 1639.1 Types and Features of Master Protocols 1639.2 Estimands in Master Protocols 1659.3 Multiplicity 1679.4 Master Protocols Using External Controls 1689.4.1 Estimands in master protocols using EDE 1699.4.2 Special considerations for different types of master pro- tocols using EDE 1729.5 Case Studies for ECTs 1739.5.1 MASTER KEY project 1739.5.2 MORPHEUS 1749.6 Discussion and Summary 1769.7 Supplements 17710 Decentralized Clinical Trials 17910.1 Elements of DCTs 18010.1.1 DHTs 18010.1.2 Participant screening, recruitment, and retention 18210.1.3 Dispensing medication 18310.1.4 Remote data acquisition 18310.1.5 Outcome/endpoint assessment and data acquisition 18410.2 Regulatory Guidance and Framework on DCTs 18510.3 Statistical Challenges and Considerations 18610.3.1 Estimands 18610.3.2 Trial design 18810.3.3 Data management plan 19110.3.4 Statistical analysis plan 19210.3.5 Missing data 19310.3.6 Study conduct 19410.3.7 Reporting and monitoring of safety events 19610.3.8 Other considerations 19710.4 Examples 19810.4.1 Decentralization by necessity 19810.4.2 Decentralization for operational benefits 19810.4.3 Decentralization to address unique scientific questions 19910.4.4 Decentralization for validation of endpoints 20010.4.5 Decentralization for validation of DCT platform 20010.5 Discussion and Summary 20010.6 Supplements 20111 Drug Development for Rare Diseases 20311.1 Regulatory Guidance 20411.1.1 Guidance and related documents by the FDA 20411.1.2 Guidance and related documents by other agencies 20611.2 Challenges in Rare Disease Drug Development 20911.2.1 Study design 20911.2.2 Study conduct 21111.2.3 Statistical analyses 21211.2.4 Ultra rare diseases 21311.2.5 Others challenges 21311.3 Strategies to Address the Challenges 21411.3.1 An overall development strategy 21411.3.2 Trial design 21511.3.3 Trial conduct 21811.3.4 Statistical analyses 21911.3.5 Ultra rare diseases 22011.3.6 Other considerations 22011.4 Use of RWD and RWE 22111.4.1 Natural history studies 22111.4.2 Trial design 22311.4.3 Trial conduct 22511.4.4 Analytical methods for data analysis 22611.4.5 Model-informed drug discovery and development and in silico trials 22711.4.6 Collaborative efforts 22711.5 Case Studies 22811.5.1 NHS to support product effectiveness 22811.5.2 Use of RWD/RWE to support regulatory decisions in various settings 22811.6 Discussion and Summary 23011.7 Supplements 23212 Time-to-Event Analysis with Treatment Switches 23512.1 Scenarios of Treatment Switching 23612.1.1 Control group to receive new treatment 23612.1.2 Treatment group to receive control drug 23612.1.3 Control and treatment groups to receive other drugs 23812.2 Study Designs 23812.3 Strategies to Handle Treatment Switching 24012.3.1 Treatment policy strategy 24012.3.2 Hypothetical strategy 24012.3.3 Principal stratum strategy 24112.3.4 Composite variable strategy 24112.3.5 While-on-treatment strategy 24112.4 Analytical Methods 24212.4.1 Counterfactual methods 24212.4.2 Non-counterfactual methods 24312.4.3 Choice of appropriate methods for adjusting treatment switching 24512.5 Considerations 24712.5.1 Study Conduct 24712.5.2 Statistical analysis plan 24812.6 Communication with regulatory agencies 24912.7 Case Studies 25012.7.1 Olaparib for patients with prostate cancer 25012.7.2 Pazopanib for patients with renal cell carcinoma 25112.7.3 The LATITUDE Study 25212.8 Discussion and Summary 25312.9 Supplements 25313 Precision Medicine 25513.1 Regulatory Activities and Approvals 25613.1.1 Precision medicine at the FDA 25613.1.2 Precision medicine at the EMA 25713.1.3 Precision medicine at the NMPA 25813.1.4 Regulatory approvals of precision medicine 25913.2 Biomarkers 26013.2.1 Types of biomarkers 26013.2.2 Discovery, development, and validation of biomarkers 26413.3 Study Design 26713.3.1 Enrichment designs 26713.3.2 SMART Designs 28113.4 Analytic Methods and Applications 28313.4.1 Model-informed drug development (MIDD) 28413.4.2 In silico models 28513.4.3 Health digital twins 28613.4.4 Single-cell omics 28713.5 Optimal Treatment Regimes 29713.5.1 Single decision point 29713.5.2 Multiple decision points 30613.6 Discussion and Summary 31113.7 Supplements 31314 Vaccine Effectiveness Studies 31714.1 General Considerations 31914.2 Early Phase Development 32014.2.1 Immune response 32014.2.2 Immunogenicity 32114.2.3 Early phases of clinical development 32214.2.4 Novel design of confirmatory trials 32314.2.5 Vaccine challenge studies 32414.3 Design Considerations 32514.3.1 Endpoints 32514.3.2 Study population 32614.3.3 Randomization 32714.3.4 Adaptive design 32814.3.5 Sample size 32914.4 Vaccine Effectiveness 33114.4.1 Rationale for using RWD & RWE in vaccine trials 33114.4.2 Design options using RWD & RWE 33214.4.3 Challenges when using RWD & RWE in vaccine trials 33214.5 Vaccine Safety 33314.5.1 A general strategy for safety evaluation 33314.5.2 A major safety endpoint in a primary hypothesis 33414.5.3 Pharmacovigilance and risk management 33414.6 Discussion and Summary 33514.7 Supplements 33615 Sensitivity Analyses in Clinical Trials 33915.1 Different Types of Analyses 33915.1.1 Primary analysis 33915.1.2 Sensitivity analysis 34015.1.3 Supplementary analysis 34115.1.4 Exploratory analysis 34215.2 Rationales for Sensitivity Analysis 34215.3 Considerations for Sensitivity Analysis 34315.3.1 Three-layer sensitivity analyses 34415.3.2 Order of sensitivity analyses 34515.4 Methods for Sensitivity Analyses 34615.4.1 Observed data and distributional assumptions 34615.4.2 Missing mechanism and missing data imputation 34715.4.3 Sensitivity analysis for causal inference 34915.5 Summary and Conclusion 35015.6 Supplements 35116 Safety Evaluation 35716.1 Safety databases 35716.1.1 Safety data from clinical trials 35716.1.2 Spontaneous adverse event reporting systems 35816.2 Statistical Methods 36016.2.1 Proportional reporting ratio and relative reporting risk 36116.2.2 Empirical Bayes method 36116.2.3 Bayesian signal detection 36316.2.4 Likelihood ratio test 36416.2.5 Tree-based scan test statistics 36516.2.6 Machine learning and deep-learning methods 36716.2.7 Sequential tests 36816.3 A Case Study 37416.3.1 Targeted maximum likelihood estimation 37616.4 Discussion and Summary 37816.5 Supplements 37917 Estimands in RWE Studies 38117.1 Frameworks for Defining Estimands 38117.1.1 ICH E9(R1) Statistical Principles for Clinical Trials: Ad- dendum: Estimands and Sensitivity Analysis in Clinical Trials 38217.1.2 Causal inference framework 38317.1.3 Target trial 38517.1.4 Targeted learning framework 38617.2 Estimands in RWE Studies 38817.2.1 Population—heterogeneity, eligibility, and representation 38817.2.2 Treatment—adherence, preferences, and dynamic treat- ment regimes 38917.2.3 Endpoints—Surrogate endpoints, clinical outcomes, and survival 38917.2.4 Intercurrent events—drug-induced, behavioral and non- behavioral adherence 39017.2.5 Population-level summary 39217.2.6 Sensitivity analysis 39317.2.7 Similarities and differences in defining an estimand be- tween TCTs and RWE studies 39317.3 Examples of Estimands in TCT and RWE Studies 39317.3.1 Traditional clinical trials 39317.3.2 Pragmatic trials 39417.3.3 Single-arm trials 39517.3.4 Observational studies 39717.3.5 Longitudinal study with a static treatment regime 39817.3.6 Longitudinal study with dynamic treatment regimes 40017.4 Discussion and Summary 40317.5 Supplements 40318 A Roadmap for Formulating RWE Studies 40518.1 Stakeholders and Research Questions 40518.2 Study Design 40618.3 Fit-for-Purpose RWD 40618.4 Treatment Regimes 40718.5 Possible ICSs 40718.6 A Roadmap 40818.7 Discussion and Summary 40818.8 Supplements 41019 Artificial Intelligence and Machine Learning in Clinical Stud-ies  41119.1 Study Design and Planning 41219.1.1 Determination of design parameters 41219.1.2 Natural history studies 41219.1.3 Genotypic and phenotypic profiling 41519.2 Study Conduct 41619.2.1 Site selection and optimization 41719.2.2 Patient identification, recruitment, and retention  41719.2.3 Decentralization 41819.2.4 Safety monitoring 41819.2.5 Risk-based monitoring  41819.2.6 Predictive analytics 41919.3 Data Analytics 42319.3.1 Adaptive clinical trial designs 42319.3.2 Data cleaning 43019.3.3 Missing data imputation 43119.3.4 Integration of multiple data sources 43219.3.5 Generation of synthetic controls 43319.3.6 Reporting and visualization 43519.4 Prediction of Clinical Trial Outcomes 43619.4.1 Factors influencing clinical trial outcomes 43719.4.2 Prediction of probability of trial success 43819.4.3 Prediction of trial enrollment and duration 44319.5 Discussion and Summary 44519.6 Supplements 446Index 573