To optimize designs for longitudinal studies analyzed by mixed-effect models with binary outcomes, the Fisher information matrix can be used. Optimal design approaches, however, require a priori knowledge of the model. We aim to propose, for the first time, a robust design approach accounting for model uncertainty in longitudinal trials with two treatment groups, assuming mixed-effect logistic models. To optimize designs given one model, we compute several optimality criteria based on Fisher information matrix evaluated by the new approach based on Monte-Carlo/Hamiltonian Monte-Carlo. We propose to use the DDS-optimality criterion, as it ensures a compromise between the precision of estimation of the parameters, and hence the Wald test power, and the overall precision of parameter estimation. To account for model uncertainty, we assume candidate models with their respective weights. We compute robust design across these models using compound DDS-optimality. Using the Fisher information matrix, we propose to predict the average power over these models. Evaluating this approach by clinical trial simulations, we show that the robust design is efficient across all models, allowing one to achieve good power of test. The proposed design strategy is a new and relevant approach to design longitudinal studies with binary outcomes, accounting for model uncertainty.
LalondeRLKowalskiKGHutmacherMM, et al.Model-based drug development. Clin Pharmacol Ther2007; 82: 21–32.
2.
MouldDRUptonRN. Basic concepts in population modeling, simulation, and model-based drug development. CPT Pharmacometrics Syst Pharmacol2012; 1: e6–e6.
3.
MouldDRUptonRN. Basic concepts in population modeling, simulation, and model-based drug development – part 2: introduction to pharmacokinetic modeling methods. CPT Pharmacomet Syst Pharmacol2013; 2: e38–e38.
4.
UptonRNMouldDR. Basic concepts in population modeling, simulation, and model-based drug development: part 3 – introduction to pharmacodynamic modeling methods. CPT Pharmacomet Syst Pharmacol2014; 3: e88–e88.
5.
SyrjänenRKKilpiTMKaijalainenTH, et al.Nasopharyngeal carriage of Streptococcus pneumoniae in Finnish children younger than 2 years old. J Infect Dis2001; 184: 451–459.
6.
MestreTABeaulieu-BoireIAquinoCC, et al.What is a clinically important change in the Unified Dyskinesia Rating Scale in Parkinson’s disease?Parkinsonism Relat Disord2015; 21: 1349–1354.
7.
AtkinsonADonevATobiasR. Optimum experimental designs, with SAS, Oxford, New York: Oxford University Press, 2007.
8.
PronzatoLPázmanA. Design of experiments in nonlinear models: asymptotic normality, optimality criteria and small-sample properties, New York: Springer Science & Business Media, 2013.
9.
FedorovVLeonovS. Optimal design for nonlinear response models, Boca Raton: CRC Press, 2013, pp. 20–29.
10.
MentréFMalletABaccarD. Optimal design in random-effects regression models. Biometrika1997; 84: 429–442.
11.
NybergJBazzoliCOgungbenroK, et al.Methods and software tools for design evaluation in population pharmacokinetics-pharmacodynamics studies. Br J Clin Pharmacol2015; 79: 6–17.
12.
NguyenTTMentréF. Evaluation of the Fisher information matrix in nonlinear mixed effect models using adaptive Gaussian quadrature. Comput Stat Data Anal2014; 80: 57–69.
13.
RiviereM-KUeckertSMentréF. An MCMC method for the evaluation of the Fisher information matrix for non-linear mixed effect models. Biostatistics2016; 17: 737–750.
14.
Mielke T. Approximation of the Fisher information and design in nonlinear mixed effects models. PhD Thesis. Otto-von-Guericke-Universitat Magdeburg, Germany, 2012.
15.
WaiteTWWoodsDC. Designs for generalized linear models with random block effects via information matrix approximations. Biometrika2015; 102: 677–693.
16.
Waite TW. Design of experiments with mixed effects and discrete responses plus related topics. PhD Thesis. University of Southampton, United Kingdom, 2012.
17.
UeckertSMentréF. A new method for evaluation of the Fisher information matrix for discrete mixed effect models using Monte Carlo sampling and adaptive Gaussian quadrature. Comput Stat Data Anal2017; 111: 203–219.
18.
Riviere M-K and Mentré F. R package MIXFIM, version 1.0 http://mc-stan.org/ (accessed 17 August 2015).
19.
Stan Development Team. Rstan: the R interface to Stan, version 2.12.0. http://mc-stan.org/ (accessed 19 January 2018).
20.
AtkinsonACBogackaB. Compound D- and Ds-optimum designs for determining the order of a chemical reaction. Technometrics1997; 39: 347–356.
21.
RetoutSCometsESamsonA, et al.Design in nonlinear mixed effects models: optimization using the Fedorov-Wynn algorithm and power of the Wald test for binary covariates. Stat Med2007; 26: 5162–5179.
22.
SölknerJ. Choice of optimality criteria for the design of crossbreeding experiments. J Anim Sci1993; 71: 2867–2873.
23.
BucklandSTBurnhamKPAugustinNH. Model selection: an integral part of inference. Biometrics1997; 53: 603–618.
24.
AtkinsonAC. DT-optimum designs for model discrimination and parameter estimation. J Stat Plan Infer2008; 138: 56–64.
25.
NguyenTTBénechHDelaforgeM, et al.Design optimisation for pharmacokinetic modeling of a cocktail of phenotyping drugs. Pharm Stat2016; 15: 165–177.
26.
SchorningKBornkampBBretzF, et al.Model selection versus model averaging in dose finding studies. Stat Med2016; 35: 4021–4040.
27.
AokiYRöshammarDHamrénB, et al.Model selection and averaging of nonlinear mixed-effect models for robust phase III dose selection. J Pharmacokinet Pharmacodyn2017; 44: 581–597.
28.
BuatoisSUeckertSFreyN, et al.Comparison of model averaging and model selection in dose finding trials analyzed by nonlinear mixed effect models. AAPS J2018; 20: 56–56.
KieferJWolfowitzJ. The equivalence of two extremum problems. Can J Math1960; 12: 363–366.
31.
OgungbenroKAaronsL. Population Fisher information matrix and optimal design of discrete data responses in population pharmacodynamic experiments. J Pharmacokinet Pharmacodyn2011; 38: 449–469.
32.
KuhnELavielleM. Maximum likelihood estimation in nonlinear mixed effects models. Comput Stat Data Anal2005; 49: 1020–1038.
33.
ClopperCJPearsonES. The use of confidence or fiducial limits illustrated in the case of the binomial. Biometrika1934; 26: 404–413.
34.
BlazerDBurchettBServiceC, et al.The association of age and depression among the elderly: an epidemiologic exploration. J Gerontol1991; 46: M210–M215.
35.
LandermanLRMustilloSALandKC. Modeling repeated measures of dichotomous data: testing whether the within-person trajectory of change varies across levels of between-person factors. Soc Sci Res2011; 40: 1456–1464.
36.
NguyenTTBazzoliCMentréF. Design evaluation and optimisation in crossover pharmacokinetic studies analysed by nonlinear mixed effects models. Stat Med2012; 31: 1043–1058.
37.
MehtäläJAuranenKKulathinalS. Optimal designs for epidemiologic longitudinal studies with binary outcomes. Stat Methods Med Res2015; 24: 803–818.
38.
CookRDWongWK. On the equivalence of constrained and compound optimal designs. J Am Stat Assoc1994; 89: 687–692.
39.
YeattsSDGenningsCCroftonKM. Optimal design for the precise estimation of an interaction threshold: the impact of exposure to a mixture of 18 polyhalogenated aromatic hydrocarbons. Risk Anal2012; 32: 1784–1797.
40.
WinkensBSchoutenHJAvan BreukelenGJP, et al.Optimal designs for clinical trials with second-order polynomial treatment effects. Stat Methods Med Res2007; 16: 523–537.
41.
Alhorn K, Schorning K and Dette H. Optimal designs for frequentist model averaging. arXiv:180705234 [stat], http://arxiv.org/abs/1807.05234 (2018, accessed 1 August 2018).
42.
DuboisALavielleMGsteigerS, et al.Model-based analyses of bioequivalence crossover trials using the stochastic approximation expectation maximisation algorithm. Stat Med2011; 30: 2582–2600.
43.
SavicRMMentréFLavielleM. Implementation and evaluation of the SAEM algorithm for longitudinal ordered categorical data with an illustration in pharmacokinetics–pharmacodynamics. AAPS J2010; 13: 44–53.
44.
PlanELMaloneyAMentréF, et al.Performance comparison of various maximum likelihood nonlinear mixed-effects estimation methods for dose-response models. AAPS J2012; 14: 420–432.
BazzoliCRetoutSMentréF. Design evaluation and optimisation in multiple response nonlinear mixed effect models: PFIM 3.0. Comput Methods Programs Biomed2010; 98: 55–65.
52.
LestiniGDumontCMentréF. Influence of the size of cohorts in adaptive design for nonlinear mixed effects models: an evaluation by simulation for a pharmacokinetic and pharmacodynamic model for a biomarker in oncology. Pharm Res2015; 32: 3159–3169.
53.
StrömbergEAHookerAC. The effect of using a robust optimality criterion in model based adaptive optimization. J Pharmacokinet Pharmacodyn2017; 44: 317–324.
54.
FangXLiJWongWK, et al.Detecting the violation of variance homogeneity in mixed models. Stat Methods Med Res2016; 25: 2506–2520.
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
