Skip to main content

Table 3 The multilevel logistic regression model with clinician and hospital random effects

From: Effect of a multi-faceted quality improvement intervention on inappropriate antibiotic use in children with non-bloody diarrhoea admitted to district hospitals in Kenya

Variable N Levels Odds ratio 95% CI p-value
Group 566 Control    0.843
  594 Intervention 0.85 0.17-4.30  
Survey 275 Baseline    0.159
  222 1st Follow-up 0.37 0.14-0.96  
  220 2nd Follow-up 0.42 0.16-1.16  
  443 3rd Follow-up 0.41 0.17-0.98  
Group × Survey 103 Intervention × Baseline    0.048
  119 Intervention × 1st Follow-up 1.03 0.28-3.72  
  109 Intervention × 2ndFollow-up 0.30 0.08-1.15  
  263 Intervention × 3rd Follow-up 0.30 0.09-1.02  
Crackles 742 Absent    < 0.001
  56 Present 6.98 2.61-18.67  
  362 Not recorded 2.11 1.29-3.44  
Random-effects N Component Estimate 95% CI  
Clinician 308 Standard deviation 1.20 0.92-1.57  
Hospital 8 Standard deviation 0.90 0.51-1.57