Bayesian Model Uncertainty in Mixed Effects Models

2008 ◽  
pp. 37-62 ◽  
Author(s):  
Satkartar K. Kinney ◽  
David B. Dunson
Keyword(s):  
Model Uncertainty ◽  
Bayesian Model ◽  
Mixed Effects ◽  
Mixed Effects Models

scholarly journals Benefits of Bayesian Model Selection and Averaging for Mixed-Effects Modeling

2021 ◽  
Author(s):  
Daniel W. Heck ◽  
Florence Bockting
Keyword(s):  
Model Selection ◽  
Random Effects ◽  
Mixed Models ◽  
Bayesian Model ◽  
Fixed Effects ◽  
Bayes Factor ◽  
Mixed Effects ◽  
Bayes Factors ◽  
Mixed Effects Models ◽  
Within Subjects

Bayes factors allow researchers to test the effects of experimental manipulations in within-subjects designs using mixed-effects models. van Doorn et al. (2021) showed that such hypothesis tests can be performed by comparing different pairs of models which vary in the specification of the fixed- and random-effect structure for the within-subjects factor. To discuss the question of which of these model comparisons is most appropriate, van Doorn et al. used a case study to compare the corresponding Bayes factors. We argue that researchers should not only focus on pairwise comparisons of two nested models but rather use the Bayes factor for performing model selection among a larger set of mixed models that represent different auxiliary assumptions. In a standard one-factorial, repeated-measures design, the comparison should include four mixed-effects models: fixed-effects H0, fixed-effects H1, random-effects H0, and random-effects H1. Thereby, the Bayes factor enables testing both the average effect of condition and the heterogeneity of effect sizes across individuals. Bayesian model averaging provides an inclusion Bayes factor which quantifies the evidence for or against the presence of an effect of condition while taking model-selection uncertainty about the heterogeneity of individual effects into account. We present a simulation study showing that model selection among a larger set of mixed models performs well in recovering the true, data-generating model.

scholarly journals Benefits of Bayesian Model Averaging for Mixed-Effects Modeling

2021 ◽  
Author(s):  
Daniel W. Heck ◽  
Florence Bockting
Keyword(s):  
Mixed Models ◽  
Bayesian Model ◽  
Fixed Effects ◽  
Bayesian Model Averaging ◽  
Model Averaging ◽  
Mixed Effects ◽  
Mixed Effects Models ◽  
Effect Sizes ◽  
Average Effect ◽  
Within Subjects

AbstractBayes factors allow researchers to test the effects of experimental manipulations in within-subjects designs using mixed-effects models. van Doorn et al. (2021) showed that such hypothesis tests can be performed by comparing different pairs of models which vary in the specification of the fixed- and random-effect structure for the within-subjects factor. To discuss the question of which model comparison is most appropriate, van Doorn et al. compared three corresponding Bayes factors using a case study. We argue that researchers should not only focus on pairwise comparisons of two nested models but rather use Bayesian model selection for the direct comparison of a larger set of mixed models reflecting different auxiliary assumptions regarding the heterogeneity of effect sizes across individuals. In a standard one-factorial, repeated measures design, the comparison should include four mixed-effects models: fixed-effects H0, fixed-effects H1, random-effects H0, and random-effects H1. Thereby, one can test both the average effect of condition and the heterogeneity of effect sizes across individuals. Bayesian model averaging provides an inclusion Bayes factor which quantifies the evidence for or against the presence of an average effect of condition while taking model selection uncertainty about the heterogeneity of individual effects into account. We present a simulation study showing that model averaging among a larger set of mixed models performs well in recovering the true, data-generating model.

scholarly journals Effect of tofogliflozin on arterial stiffness in patients with type 2 diabetes: prespecified sub-analysis of the prospective, randomized, open-label, parallel-group comparative UTOPIA trial

2021 ◽  
Vol 20 (1) ◽  
Author(s):  
Naoto Katakami ◽  
◽  
Tomoya Mita ◽  
Hidenori Yoshii ◽  
Toshihiko Shiraiwa ◽  
...  
Keyword(s):  
Risk Factors ◽  
Blood Pressure ◽  
Type 2 Diabetes ◽  
Conventional Treatment ◽  
Mixed Effects ◽  
Mixed Effects Models ◽  
Open Label ◽  
Parallel Group ◽  
History Of

Abstract Background Tofogliflozin, an SGLT2 inhibitor, is associated with favorable metabolic effects, including improved glycemic control and serum lipid profile and decreased body weight, visceral adipose tissue, and blood pressure (BP). This study evaluated the effects of tofogliflozin on the brachial-ankle pulse wave velocity (baPWV) in patients with type 2 diabetes (T2DM) without a history of apparent cardiovascular disease. Methods The using tofogliflozin for possible better intervention against atherosclerosis for type 2 diabetes patients (UTOPIA) trial is a prospective, randomized, open-label, multicenter, parallel-group, comparative study. As one of the prespecified secondary outcomes, changes in baPWV over 104 weeks were evaluated in 154 individuals (80 in the tofogliflozin group and 74 in the conventional treatment group) who completed baPWV measurement at baseline. Results In a mixed-effects model, the progression in the right, left, and mean baPWV over 104 weeks was significantly attenuated with tofogliflozin compared to that with conventional treatment (– 109.3 [– 184.3, – 34.3] (mean change [95% CI] cm/s, p = 0.005; – 98.3 [– 172.6, – 24.1] cm/s, p = 0.010; – 104.7 [– 177.0, – 32.4] cm/s, p = 0.005, respectively). Similar findings were obtained even after adjusting the mixed-effects models for traditional cardiovascular risk factors, including body mass index (BMI), glycated hemoglobin (HbA1c), total cholesterol, high-density lipoprotein (HDL)-cholesterol, triglyceride, systolic blood pressure (SBP), hypertension, smoking, and/or administration of drugs, including hypoglycemic agents, antihypertensive agents, statins, and anti-platelets, at baseline. The findings of the analysis of covariance (ANCOVA) models, which included the treatment group, baseline baPWV, and traditional cardiovascular risk factors, resembled those generated by the mixed-effects models. Conclusions Tofogliflozin significantly inhibited the increased baPWV in patients with T2DM without a history of apparent cardiovascular disease, suggesting that tofogliflozin suppressed the progression of arterial stiffness. Trial Registration UMIN000017607. Registered 18 May 2015. (https://www.umin.ac.jp/icdr/index.html)

The Use of Theory of Linear Mixed-Effects Models to Detect Fraudulent Erasures at an Aggregate Level

2021 ◽  
pp. 001316442199489
Author(s):  
Luyao Peng ◽  
Sandip Sinharay
Keyword(s):  
Type I Error ◽  
Real Data ◽  
Mixed Effects ◽  
Error Rates ◽  
Mixed Effects Models ◽  
Type I ◽  
Aggregate Level ◽  
Linear Mixed Effects Models ◽  
Linear Mixed Effects ◽  
Best Linear Unbiased

Wollack et al. (2015) suggested the erasure detection index (EDI) for detecting fraudulent erasures for individual examinees. Wollack and Eckerly (2017) and Sinharay (2018) extended the index of Wollack et al. (2015) to suggest three EDIs for detecting fraudulent erasures at the aggregate or group level. This article follows up on the research of Wollack and Eckerly (2017) and Sinharay (2018) and suggests a new aggregate-level EDI by incorporating the empirical best linear unbiased predictor from the literature of linear mixed-effects models (e.g., McCulloch et al., 2008). A simulation study shows that the new EDI has larger power than the indices of Wollack and Eckerly (2017) and Sinharay (2018). In addition, the new index has satisfactory Type I error rates. A real data example is also included.

scholarly journals ALLY in fighting COVID-19: magnitude of albumin decline and lymphopenia (ALLY) predict progression to critical disease

2021 ◽  
pp. jim-2020-001525
Author(s):  
Johanna S van Zyl ◽  
Amit Alam ◽  
Joost Felius ◽  
Ronnie M Youssef ◽  
Dipesh Bhakta ◽  
...  
Keyword(s):  
Medical Records ◽  
Resource Planning ◽  
Mixed Effects ◽  
Mixed Effects Models ◽  
Disease Course ◽  
Risk Models ◽  
Simple Tool ◽  
Laboratory Results ◽  
Risk Of Disease ◽  
Critical Patients

The global severe acute respiratory syndrome coronavirus 2 (SARS-CoV-2) pandemic leading to coronavirus disease 2019 (COVID-19) is straining hospitals. Judicious resource allocation is paramount but difficult due to the unpredictable disease course. Once hospitalized, discerning which patients may progress to critical disease would be valuable for resource planning. Medical records were reviewed for consecutive hospitalized patients with COVID-19 in a large healthcare system in Texas. The main outcome was progression to critical disease within 10 days from admission. Albumin trends from admission to 7 days were analyzed using mixed-effects models, and progression to critical disease was modeled by multivariable logistic regression of laboratory results. Risk models were evaluated in an independent group. Of 153 non-critical patients, 28 (18%) progressed to critical disease. The rate of decrease in mean baseline-corrected (Δ) albumin was −0.08 g/dL/day (95% CI −0.11 to −0.04; p<0.001) or four times faster, in those who progressed compared with those who did not progress. A model of Δ albumin combined with lymphocyte percentage predicting progression to critical disease was validated in 60 separate patients (sensitivity, 0.70; specificity, 0.74). ALLY (delta albumin and lymphocyte percentage) is a simple tool to identify patients with COVID-19 at higher risk of disease progression when: (1) a 0.9 g/dL or greater albumin drop from baseline within 5 days of admission or (2) baseline lymphocyte of ≤10% is observed. The ALLY tool identified >70% of hospitalized cases that progressed to critical COVID-19 disease. We recommend prospectively tracking albumin. This is a globally applicable tool for all healthcare systems.

Sign in / Sign up

Export Citation Format

Share Document