How do GP practices and patient characteristics influence the prescription of antidepressants? A cross-sectional study

Table 3 Multivariate analysis: variables independently associated with the dichotomized AD prescription ratio (<0.42 for lower prescribers and ≥0.42 for higher prescribers)

Variable	Mean difference in AD prescription ratio [95% CI] ^b	OR [95% CI] ^c
GP’s age
≥60	0	1
50–59	0.026 [−0.023; 0.074]	1.87 [1.29; 2.71]
40–49	0.038 [−0.049; 0.080]	1.74 [1.14; 2.64]
<40	−0.068 [−0.14; 0.00]	0.65 [0.35; 1.2]
Practice location:
Rural	0	1
Urban	0.056 [0.016; 0.096]	1.954 [1.3; −2.80]
Number of consultations/year/GP*		0.93^a [0.90–0.97]
Number of patients with low income
≥80		1
40–79		1.69 [1.06; 2.67]
<39		2.23 [1.45; 3.40]
Class of prescription
<70% new ADs	0	1
>70% new ADs	0.12 [0.058; 0.18]	3.45 [1.94; 6.15]
Cumulative number of sick leave days prescribed per year
≥6,000		1
4,000–5,999		0.91 [0.55; 1.47]
2,000–3,999		1.12 [0.68; 1.84]
Fewer than 1,999		0.43 [0.24; 0.78]
Total volume of drugs reimbursed	−0.0053* [−0.0084; −0.00233]
Total volume of prescription orders	0.0462** [0.0184; 0.0740]
Number of consultations and home visits	−0.0128*** [−0.0179; −0.0077]

*Mean difference per 1,000 additional drugs reimbursed (quantitative variable).
**Mean difference per 1,000 additional prescription orders (quantitative variable).
***Mean difference per 100 additional consultations or home visits.
^aOdds ratio per 100 additional consultations (quantitative variable).
^bFrom multiple linear regression applied to the AD prescription ratio as a continuous variable.
^cFrom multiple logistic regression applied to the dichotomized AD prescription ratio using the lower prescription level as the reference level.

ISSN: 1744-859X