Skip to main content

Table 2 Estimated relative risk and 95% CI or credible interval for multicenter randomized controlled trial example

From: Estimating relative risks in multicenter studies with a small number of centers — which methods to use? A simulation study

Modela

RR

95% CI or CrI

GLM-Bin

1.29

0.95–1.75

GEE binomial

1.43

1.01–2.02

GEE binomial, KC-corrected

 

0.96–2.14

GEE Poisson

1.42

1.01–2.01

GEE Poisson, KC-corrected

 

0.95–2.12

GLMM binomialb

–

–

Bayesian binomial GLMM

1.27

1.00–1.65

  1. Abbreviations: CrI Credible interval, GEE Generalized estimating equation, GLM-Bin Log-binomial regression model, GLMM Generalized linear mixed model, KC Kauermann and Carroll, RR Relative risk
  2. Except for the GLM-Bin model, all models are adjusted for center correlation. The KC bias correction in the GEE models adjusts the robust SE for the small number of centers (estimate of RR does not change)
  3. aGLMM Poisson models are excluded because of their poor performance in the simulation study
  4. bModel did not converge
Back to article page

Trials

ISSN: 1745-6215