TY - JOUR
T1 - Individual participant data meta-analysis to examine interactions between treatment effect and participant-level covariates
T2 - Statistical recommendations for conduct and planning
AU - Riley, Richard D.
AU - Debray, Thomas P.A.
AU - Fisher, David
AU - Hattle, Miriam
AU - Marlin, Nadine
AU - Hoogland, J.
AU - Gueyffier, Francois
AU - Staessen, Jan A.
AU - Wang, Jiguang
AU - Moons, Karel G.M.
AU - Reitsma, Johannes B.
AU - Ensor, Joie
N1 - Funding Information: R.D.R. and J.E. were supported by funding to Keele University for an awarded National Institute for Health Research (NIHR) Clinical Trials Unit Support Funding Opportunity entitled “Supporting efficient/innovative delivery of NIHR research”. T.P.A.D., J.H., K.G.M.M., and J.B.R. were supported by a TOP grant of the Netherlands Organisation for Health Research and Development (ZonMw) entitled “Promoting tailored healthcare: improving methods to investigate subgroup effects in treatment response when having multiple individual participant datasets” (Grant number: 91215058). N.M. is funded by an NIHR Doctoral Fellowship (DRF‐2018‐11‐ST2‐077). The i‐WIP project was funded by NIHR Health Technology Assessment (programme 12/01). This publication presents independent research funded by the NIHR. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR, or the Department of Health and Social Care. Funding Information: NIHR Health Technology Assessment Programme, programme 12/01; National Institute for Health Research (NIHR) Clinical Trials Unit Support Funding, TOP grant of the Netherlands Organisation for Health Research and Development (ZonMw), 91215058; NIHR Doctoral Fellowship, DRF‐2018‐11‐ST2‐077; Netherlands Organisation for Health Research and Development, Keele University Funding information Funding Information: information NIHR Health Technology Assessment Programme, programme 12/01; National Institute for Health Research (NIHR) Clinical Trials Unit Support Funding, TOP grant of the Netherlands Organisation for Health Research and Development (ZonMw), 91215058; NIHR Doctoral Fellowship, DRF-2018-11-ST2-077; Netherlands Organisation for Health Research and Development, Keele UniversityR.D.R. and J.E. were supported by funding to Keele University for an awarded National Institute for Health Research (NIHR) Clinical Trials Unit Support Funding Opportunity entitled “Supporting efficient/innovative delivery of NIHR research”. T.P.A.D., J.H., K.G.M.M., and J.B.R. were supported by a TOP grant of the Netherlands Organisation for Health Research and Development (ZonMw) entitled “Promoting tailored healthcare: improving methods to investigate subgroup effects in treatment response when having multiple individual participant datasets” (Grant number: 91215058). N.M. is funded by an NIHR Doctoral Fellowship (DRF-2018-11-ST2-077). The i-WIP project was funded by NIHR Health Technology Assessment (programme 12/01). This publication presents independent research funded by the NIHR. The views expressed are those of the author(s) and not necessarily those of the NHS, the NIHR, or the Department of Health and Social Care. Publisher Copyright: © 2020 The Authors. Statistics in Medicine published by John Wiley & Sons, Ltd.
PY - 2020/7/10
Y1 - 2020/7/10
N2 - Precision medicine research often searches for treatment-covariate interactions, which refers to when a treatment effect (eg, measured as a mean difference, odds ratio, hazard ratio) changes across values of a participant-level covariate (eg, age, gender, biomarker). Single trials do not usually have sufficient power to detect genuine treatment-covariate interactions, which motivate the sharing of individual participant data (IPD) from multiple trials for meta-analysis. Here, we provide statistical recommendations for conducting and planning an IPD meta-analysis of randomized trials to examine treatment-covariate interactions. For conduct, two-stage and one-stage statistical models are described, and we recommend: (i) interactions should be estimated directly, and not by calculating differences in meta-analysis results for subgroups; (ii) interaction estimates should be based solely on within-study information; (iii) continuous covariates and outcomes should be analyzed on their continuous scale; (iv) nonlinear relationships should be examined for continuous covariates, using a multivariate meta-analysis of the trend (eg, using restricted cubic spline functions); and (v) translation of interactions into clinical practice is nontrivial, requiring individualized treatment effect prediction. For planning, we describe first why the decision to initiate an IPD meta-analysis project should not be based on between-study heterogeneity in the overall treatment effect; and second, how to calculate the power of a potential IPD meta-analysis project in advance of IPD collection, conditional on characteristics (eg, number of participants, standard deviation of covariates) of the trials (potentially) promising their IPD. Real IPD meta-analysis projects are used for illustration throughout.
AB - Precision medicine research often searches for treatment-covariate interactions, which refers to when a treatment effect (eg, measured as a mean difference, odds ratio, hazard ratio) changes across values of a participant-level covariate (eg, age, gender, biomarker). Single trials do not usually have sufficient power to detect genuine treatment-covariate interactions, which motivate the sharing of individual participant data (IPD) from multiple trials for meta-analysis. Here, we provide statistical recommendations for conducting and planning an IPD meta-analysis of randomized trials to examine treatment-covariate interactions. For conduct, two-stage and one-stage statistical models are described, and we recommend: (i) interactions should be estimated directly, and not by calculating differences in meta-analysis results for subgroups; (ii) interaction estimates should be based solely on within-study information; (iii) continuous covariates and outcomes should be analyzed on their continuous scale; (iv) nonlinear relationships should be examined for continuous covariates, using a multivariate meta-analysis of the trend (eg, using restricted cubic spline functions); and (v) translation of interactions into clinical practice is nontrivial, requiring individualized treatment effect prediction. For planning, we describe first why the decision to initiate an IPD meta-analysis project should not be based on between-study heterogeneity in the overall treatment effect; and second, how to calculate the power of a potential IPD meta-analysis project in advance of IPD collection, conditional on characteristics (eg, number of participants, standard deviation of covariates) of the trials (potentially) promising their IPD. Real IPD meta-analysis projects are used for illustration throughout.
KW - effect modifier
KW - individual participant data (IPD)
KW - meta-analysis
KW - subgroup effect
KW - treatment-covariate interaction
UR - http://www.scopus.com/inward/record.url?scp=85083976083&partnerID=8YFLogxK
U2 - https://doi.org/10.1002/sim.8516
DO - https://doi.org/10.1002/sim.8516
M3 - Article
C2 - 32350891
SN - 0277-6715
VL - 39
SP - 2115
EP - 2137
JO - Statistics in medicine
JF - Statistics in medicine
IS - 15
ER -