Model‐Based Meta‐Analysis: Optimizing Research, Development, and Utilization of Therapeutics Using the Totality of Evidence

Abstract

Model‐based meta‐analysis (MBMA) is a valuable component of the quantitative pharmacology toolkit for model‐informed drug discovery and development. It enables principled decision making with a totality of evidence mindset through integration of internal and external data across multiple dimensions (e.g., targets/mechanisms, molecules/drugs, doses/regimens, diseases/indications, populations, endpoints, and clinical trial designs). MBMA distinguishes itself from traditional meta‐analysis by infusing pharmacologic plausibility into the statistical rigor that typifies meta‐analytic data integration. This is possible through mechanism‐informed formulation of pharmacologically inspired cause–effect and dose‐response relationships, time course of treatment effects, and interrelationships between proximal and distal outcomes of modulation of disease biology and pathophysiology. In this review, we offer a question‐based approach to enhance appreciation of the value of MBMA across the continuum from drug discovery and translational research through clinical development, comparative effectiveness research, and postapproval optimization of therapeutics using illustrative examples across therapeutic areas.

Population Pharmacokinetics, Efficacy Exposure-response Analysis, and Model-based Meta-analysis of Fenebrutinib in Subjects with Rheumatoid Arthritis

“…Modeling of historical meta-analysis data sought to leverage publicly available information to enable direct and indirect comparisons against competitors and between study populations. Such an analysis is particularly valuable at the end of Phase 2 milestone for informing go/no-go decisions to proceed to Phase 3 (36). First, MBMA aimed to quantitatively compare fenebrutinib against currently marketed therapies, through an indirect comparison of the performance of placebo and adalimumab historically versus that observed in the fenebrutinib Phase 2 trial.…”

Beyond Randomized Clinical Trials: Use of External Controls

“…These include meta-regression methods, 37 which use baseline covariate information to explain part of the between-trial heterogeneity, network meta-analytic approaches to synthesize evidence from multiple treatments, 38 meta-analyses for specific data types, 39 multivariate meta-analysis to handle several end points, 40 or pharmacometric model-based meta-analysis. 41 The main strengths of the meta-analytic approaches are their ability to account for unexplained between-trial heterogeneity, their flexibility to adapt to different types of source data, and the possibility to integrate both individual patient data and aggregate data from publications. Central to meta-analytic approaches is the Meta-analysis Methods to synthesize the evidence from several sources of control data.…”

“…Depending on the type of relevant source data, more complex meta‐analytic approaches may be needed for the evidence synthesis. These include meta‐regression methods, which use baseline covariate information to explain part of the between‐trial heterogeneity, network meta‐analytic approaches to synthesize evidence from multiple treatments, meta‐analyses for specific data types, multivariate meta‐analysis to handle several end points, or pharmacometric model‐based meta‐analysis …”

Quantitative prediction of bone mineral density by using bone turnover markers in response to antiresorptive agents in postmenopausal osteoporosis: A model‐based meta‐analysis

“…Model‐based meta‐analysis is a superior approach commonly used to compare efficacy indirectly and quantitatively by established pharmacodynamic (PD) models based on aggregate study‐level data 23 . Additionally, it was developed afterwards to summarize relationship between different endpoints 24,25 . This study aims to investigate the probable common relationship between BTMs (serum CTX [sCTX] and P1NP) and BMD at lumbar spine (LS) and total hip (TH).…”