Skip to main content
. 2018 Jan 25;13(1):e0190522. doi: 10.1371/journal.pone.0190522

Table 3. Results of the multivariate regression analyses.

Non Adjusted (R2 0.04) Ajusted (R2 0.14)
Coef. CI 95% p value Coef. CI 95% p value
GP Trainer
Yes -6.43 [-8.43; -4.42] <0.001 -6.62 [-8.55; -4.69] <0.001
No Ref. Ref.
Gender
Male - - - 0.34 [-1.05; 1.73] 0.63
Female - - - Ref.
Age
≥ 55yo - - - 3.67 [2.31; 5.02] <0.001
< 55yo - - - Ref.
Location of the GP’s practice
≥ 15,000 inhabitants - - - 0.36 [-1.01; 1.75] 0.61
< 15,000 inhabitants - - - Ref.
Level of activity (Number of visits)
<3,000 - - - -3.81 [-5.28; -2.35] <0.001
3,000–4,999 - - - Ref.
≥5,000 - - - 3.52 [1.86; 5.17] <0.001
Proportion of attendances by patients
< 16yo - - - 0.13 [0.02; 0.25] 0.02
with long-term disease - - - 0.09 [-0.07; 0.23] 0.291
with medical fee exemption status for perceiving low incomes - - - -0.19 [-0.30; -0.08] 0.001

Reading grid: GP for General Practitioner, CI for Confidence Interval, yo for years old.

The adjusted model showed a significant average of 6.62 percentage points lower for the GP trainers’ antibiotic prescribing rate compared to non-trainers. (23.4% in relative terms)