Skip to main content

Table 5 Statistical tests of change difference between control and intervention at the two-year follow-up

From: An educational intervention to prevent overweight in pre-school years: a cluster randomised trial with a focus on disadvantaged families

Intraclass correlation (ICC): Min. n estimate Max. n estimate  
Control 0.0396 0.0443  
Intervention 0.0000 0.0000  
  Estimate SE 95% CI P value
Analysis at the centre level:
 Outcome = final BMI z-score – baseline BMI z-score
  Control 0.55 0.12 0.16 0.94  
  Intervention −0.17 0.10 −0.49 0.16  
  Change difference (centres paired) 0.72 0.21 0.04 1.40 0.043
  Change difference (centres unpaired) 0.72 0.16 0.33 1.10 0.004
Analysis at the child level:
 Outcome = final BMI z-score – baseline BMI z-score, assuming ICC = 0
  Control 0.56 0.10 0.35 0.77  
  Intervention −0.18 0.15 −0.48 0.11  
  Change difference (centres unpaired) 0.74 0.18 0.38 1.10 < 0.001
 Outcome = final BMI z-score – baseline BMI z-score, assuming ICC > 0
  Centre level variance 0.00 0.00    
  Child level variance 0.70 0.11    
  Change difference (centres unpaired) 0.74 0.19 0.37 1.11 < 0.001
 Outcome = final BMI z-score, baseline BMI z-score as covariate, assuming ICC > 0
  Centre level variance 0.00 0.00    
  Child level variance 0.47 0.07    
  Baseline BMI z-score 0.60 0.06    
  Change difference (centres unpaired) 0.49 0.16 0.17 0.80 0.002