Objective Clinical trials for Alzheimer's disease feature long follow-up periods and high dropout rates, with missing outcome data being commonplace, which can impair the accuracy of treatment effect estimation. This study aimed to compare the applicability and performance of several commonly used and regulatory-recommended missing data handling strategies, including the mixed-effects model for repeated measures (MMRM), standard multiple imputation (MI), reference-based imputation (RBI), and δ-adjusted multiple imputation, in Alzheimer’s disease clinical trials. Methods The data were derived from a multicenter, randomized, double-blind, parallel-group, placebo-controlled clinical trial for Alzheimer’s disease. The endpoint was the Alzheimer's disease assessment scale-cognitive (ADAS-Cog) score, and the change from baseline in ADAS-Cog score at Week 26 was the primary outcome, and the difference between the treatment and placebo groups was estimated. The primary analysis used MMRM under the missing at random (MAR) assumption. Sensitivity analyses were performed using standard MI, reference-based imputation (J2R, CR, CIR), and δ-adjusted multiple imputation. Effect estimates, standard errors, confidence intervals, and P-values were compared across methods. Results Treatment effect estimates were consistent in direction across all methods. Compared with MMRM and MI under the MAR assumption, RBI yielded more conservative estimates under the missing not at random (MNAR) assumption. Under conservative δ settings the conclusions remained robust (all P-values <0.001), indicating that the findings were stable against deviations from MAR to MNAR. Conclusion In this clinical trial dataset, treatment effect inferences show good consistency and robustness across multiple missing data handling methods and missing data mechanisms. The analytical workflow proposed in this paper can serve as a reference for missing data handling and sensitivity analysis in clinical trials for neurodegenerative diseases.
ObjectiveTo review the methodological evolution, applicable scenarios, and HTA use of next-generation population-adjusted indirect comparison methods. MethodsBuilding on traditional indirect comparisons and first-generation population-adjusted methods, this review summarizes four major development paths—weight optimization, unified modelling, marginalization, and subgroup interpolation—and discusses the principles, assumptions, strengths, and limitations of two-stage matching-adjusted indirect comparison (2SMAIC), multilevel network meta-regression (ML-NMR), G-computation, multiple imputation marginalization (MIM), and network meta-interpolation (NMI). Publicly available HTA cases were also examined to identify major appraisal concerns. ResultsCompared with traditional matching-adjusted indirect comparison (MAIC) and simulated treatment comparison (STC), these next-generation methods better address complex evidence networks, transportability to target populations, and incompatibility between conditional and marginal effects. ML-NMR is most useful for anchored evidence networks with multiple studies; 2SMAIC mainly improves precision within the MAIC framework; G-computation and MIM obtain decision-relevant marginal effects through model-based standardization; and NMI enables limited adjustment using subgroup results when individual patient data are unavailable. Published applications remain concentrated in pharmaceuticals. HTA agencies are cautiously open to these methods, but focus on target-population relevance, covariate overlap, key assumptions, and uncertainty analyses. ConclusionsNext-generation population-adjusted indirect comparison methods substantially expand the toolkit for evidence generation when head-to-head trials are unavailable, but they do not remove the dependence on data quality and methodological assumptions. Method choice should therefore be aligned with the evidence structure, data availability, and decision question, with transparent reporting and sensitivity analyses.