Table 2 Summary of the adjustment of mixed effects models, including random effects for country, to observed hip fracture incidence rates.

From: Quantitative modelling of hip fracture trends in 14 European countries: testing variations of a shared reversal over time

Random effects

Model 1

Model 2

Model 3

Women

Men

Women

Men

Women

Men

SDintercept

0.348

0.454

0.348

0.457

0.346

0.457

SDyear(linear)

0.060

0.077

0.060

0.077

SDyear(quadratic)

0.014

0.000

SDresidual

0.176

0.189

0.173

0.185

0.173

0.185

p-value (LR test, each model vs. the previous)

<0.001

<0.001

0.860

>0.999

Proportion (%) of variability explained compared to the fixed effects only model

47.6

56.6

48.6

57.4

48.5

57.4

Bayesian Information Criterion (BIC)

−506.3618

−352.0886

−517.2905

−364.5618

−497.2848

−343.7998

Correlation between predicted and observed values

0.9935

0.9906

0.9938

0.9910

0.9938

0.9910

Relative Squared Error (RSE)

1.281

1.864

1.230

1.778

1.232

1.778

Relative Absolute Error (RAE)

9.997

12.193

9.784

11.935

9.781

11.935

  1. SD - standard deviation; LR - likelihood ratio.
  2. Model 1: random intercept.
  3. Model 2: Model 1 + random linear term for calendar year.
  4. Model 3: Model 2 + random quadratic term for calendar year.