Assessing the impact of case-mix heterogeneity in individual participant data meta-analysis: Novel use of I² statistic and prediction interval

Abstract

Case mix differences between trials form an important factor that contributes to the statistical heterogeneity observed in a meta-analysis. In this paper, we propose two methods to assess whether important heterogeneity would remain if the different trials in the meta-analysis were conducted in one common population defined by a given case-mix. To achieve this goal, we first standardize results of different trials over the case-mix of a target population. We then quantify the amount of heterogeneity arising from case-mix and beyond case-mix reasons by using corresponding I2 statistics and prediction intervals. These new approaches enable a better understanding of the overall heterogeneity between trial results, and can be used to support standard heterogeneity assessments in individual participant data meta-analysis practice.

Efficacy and Safety of Systemic Treatments Among Colorectal Cancer Patients: A Network Meta-Analysis of Randomized Controlled Trials

“…Third, the limitation of the meta-analysis in considering differences across studies due to patient characteristics must be addressed. Recent methods have been proposed to account the heterogeneity between trial results in meta-analyses ( 126 , 127 ); however, the application of NMA requires further investigation. Fourth, the current study did not yield an ideal index for both efficacy and safety from six evaluated endpoints to select the ideal regimen for the treatment of advanced or metastatic CRC.…”

Parametric G‐computation for compatible indirect treatment comparisons with limited individual patient data

“…These results support previous findings comparing outcome regression and weighting in different contexts. 24,26,91,92 We can now capitalize on the advantages offered by outcome regression with respect to weighting in our scenario, for example extrapolation capabilities and increased precision. Outcome regression methods are also appealing because they make different modeling assumptions than weighting.…”

“…23,24 Outcome regression approaches such as STC are appealing as these tend to be more efficient than weighting, providing more stable estimators and allowing for model extrapolation. 25,26 We view extrapolation as an advantage because poor overlap, with small effective sample sizes and large percentage reductions in effective sample size, is a pervasive issue in HTA. 27 While extrapolation can also be viewed as a disadvantage if it is not valid, in our case it expands the range of scenarios in which population adjustment can be used.…”

Challenges in Systematic Reviews and Meta-Analyses of Mediation Analyses

“…The fixed-effect component in this model describes the summary indirect effect of on via the mediator in the target population , and the variance component 2 describes the heterogeneity between the individual indirect effect effects across studies even after being standardized over the same covariate distribution of the trial population . Alternative random-effect models for summarizing ˆ ( , ) can also be extended from previous works [31].…”