Article type
Year
Abstract
Background:
A key question for meta-analysis is to reliably assess whether treatment effects vary across different participant subgroups. Traditionally, these interactions have been estimated using approaches known to induce aggregation bias, so we previously recommended a within-trial approach to provide unbiased estimates for binary or ordered-categorical patient-level treatment-covariate interactions. However, patients, clinicians, and policymakers need reliable estimates of treatment effects within specific covariate subgroups, on relative and absolute scales, in order to target treatments appropriately, which estimation of an interaction effect itself does not provide.
Objective and Methods:
Our objective is to describe further developments to the “within-trial” framework by providing practical methods to (1) estimate within-trial interactions across two or more subgroups; (2) estimate subgroup-specific (“floating”) treatment effects that are compatible with the within-trial interactions and make maximum use of available data; and (3) clearly present these data using novel implementation of forest plots. We describe the steps involved and show how the methods can be applied using detailed aggregate (“summary”) data either extracted from trial publications or obtained directly from trial authors. We demonstrate the within-trial framework by applying it to two examples taken from previously published meta-analyses in which detailed aggregate data were available.
Results:
In our first example, a meta-analysis of the effects of interleukin-6 antagonists on outcomes for patients hospitalised with COVID-19, we show how the method can be used for a binary covariate. Our second example, a meta-analysis of the effects of postoperative radiotherapy on survival of patients with non-small cell lung cancer, allows us to demonstrate the framework for a categorical covariate with three levels. We demonstrate how to implement these methods using a newly developed command (metafloat) in Stata.
Conclusions:
Our within-trial framework allows straightforward estimation of a range of within-trial treatment-covariate interactions, compatible subgroup-specific treatment effects, and corresponding forest plots to clearly and effectively demonstrate how treatment effects differ across patient subgroups.
Patient, public, and/or healthcare consumer involvement: One patient and one member of the public are involved in the dissemination of the results of this study, including advising on how best to present the methodology to various audiences.
A key question for meta-analysis is to reliably assess whether treatment effects vary across different participant subgroups. Traditionally, these interactions have been estimated using approaches known to induce aggregation bias, so we previously recommended a within-trial approach to provide unbiased estimates for binary or ordered-categorical patient-level treatment-covariate interactions. However, patients, clinicians, and policymakers need reliable estimates of treatment effects within specific covariate subgroups, on relative and absolute scales, in order to target treatments appropriately, which estimation of an interaction effect itself does not provide.
Objective and Methods:
Our objective is to describe further developments to the “within-trial” framework by providing practical methods to (1) estimate within-trial interactions across two or more subgroups; (2) estimate subgroup-specific (“floating”) treatment effects that are compatible with the within-trial interactions and make maximum use of available data; and (3) clearly present these data using novel implementation of forest plots. We describe the steps involved and show how the methods can be applied using detailed aggregate (“summary”) data either extracted from trial publications or obtained directly from trial authors. We demonstrate the within-trial framework by applying it to two examples taken from previously published meta-analyses in which detailed aggregate data were available.
Results:
In our first example, a meta-analysis of the effects of interleukin-6 antagonists on outcomes for patients hospitalised with COVID-19, we show how the method can be used for a binary covariate. Our second example, a meta-analysis of the effects of postoperative radiotherapy on survival of patients with non-small cell lung cancer, allows us to demonstrate the framework for a categorical covariate with three levels. We demonstrate how to implement these methods using a newly developed command (metafloat) in Stata.
Conclusions:
Our within-trial framework allows straightforward estimation of a range of within-trial treatment-covariate interactions, compatible subgroup-specific treatment effects, and corresponding forest plots to clearly and effectively demonstrate how treatment effects differ across patient subgroups.
Patient, public, and/or healthcare consumer involvement: One patient and one member of the public are involved in the dissemination of the results of this study, including advising on how best to present the methodology to various audiences.