Skip to main content
. 2009 Jun 23;338:b2242. doi: 10.1136/bmj.b2242

Table 2.

 Two level logistic regression model* of odds of being prescribed antibiotic in each network (3296 patients from 384 clinicians). Figures are odds ratios (95% confidence intervals)

Network OR (95% CI) P value
Antwerp 0.22 (0.12 to 0.38) <0.001
Balatonfured 5.69 (2.88 to 11.26) <0.001
Barcelona 0.29 (0.16 to 0.51) <0.001
Bratislava 11.2 (6.20 to 20.27) <0.001
Cardiff 2.44 (1.42 to 4.19) <0.01
Helsinki 0.58 (0.31 to 1.09) 0.09
Jonkoping 0.25 (0.16 to 0.38) <0.001
Lodz 4.14 (2.4 to 7.16) <0.001
Mataro 0.66 (0.37 to 1.18) 0.16
Milan 6.81 (3.49 to 13.27) <0.001
Rotenberg 0.5 (0.27 to 0.92) 0.03
Southampton 0.84 (0.47 to 1.5) 0.55
Tromso 0.18 (0.11 to 0.30) <0.001
Utrecht 0.5 (0.29 to 0.85) 0.01

*Model controls for clinician rated symptom scores and clinical presentation. Clinician level variance component was 23.3%, using π2/3 estimator.