Table 4. Factors associated with the prescription of antidepressant treatment combined with psychotherapy versus an antidepressant alone (versus antidepressant therapy only, case-vignette, French nationwide panel of general practitioners, weighted multivariate logistic model, N = 720).
| GPs' treatment choices | ||||
| ATD without psycho-therapy (%) | ATD with psycho-therapy (%) | Odds ratio [95% confidence interval] | ||
| All | 33.8 | 66.2 | ||
| GPs' personal characteristics | ||||
| Age† | <49 | 22.1 | 36.2 | __ |
| 49–56 | 37.7 | 37.2 | 0.67 [0.45;1.00] | |
| >56 | 40.2 | 26.6 | 0.49 [0.32;0.76] | |
| Gender† | Male | 83.3 | 69.8 | __ |
| Female | 16.7 | 30.2 | 1.76 [1.17;2.63] | |
| GPs' opinion about depression treatments | ||||
| Thinks that she/he is sufficiently trained to manage depression | Totally disagrees/partially disagrees | 10.3 | 18.9 | __ |
| Partially agrees/totally agrees | 89.7 | 81.1 | 0.57 [0.36;0.91] | |
| Score of perception of the effectiveness of psychotherapy in treating depression‡ | Low (0–6) | 38.4 | 22.7 | __ |
| Intermediate (7–9) | 44.7 | 53.2 | 1.88 [1.29;2.73] | |
| High (10–12) | 16.9 | 24.1 | 1.73 [1.09;2.77] | |
| GPs' opinion on cooperation with mental health specialists | ||||
| Satisfaction about the cooperation with mental health specialists | Yes | 29.3 | 40.6 | __ |
| No | 70.7 | 59.4 | 0.63 [0.45;0.89] | |
| Score of perception of the utility of cooperation with mental health specialists§ | Low (0–8) | 43.1 | 28.0 | |
| Intermediate (9–10) | 36.1 | 29.7 | 1.28 [0.87;1.88] | |
| High (11–12) | 20.8 | 42.3 | 3.08 [2.06;4.62] | |
| Patient's personal characteristics and depression severity | ||||
| Intensity of depression | Mild | 54.6 | 41.8 | __ |
| Severe | 45.4 | 58.2 | 1.82 [1.31;2.52] | |
| Patient's SES | Blue-collar | 54.8 | 43.1 | __ |
| White-collar | 45.2 | 56.9 | 1.58 [1.14;2.18] | |
Sample stratification variables were forced in the model even if they did not meet the p<0.05 criterion; among them, the number of visits and house calls in 2008 and the place of practice were not significantly associated to the dependent variable (results not presented).
Score transformed into a three categorical variable according to its first and third tertiles values (min = 0; Q1 = 6; Q3 = 9; max = 12).
Score transformed into a three categorical variable according to its first and third tertiles values (min = 0; Q1 = 6; Q3 = 11; max = 12).
Hosmer-Lemeshow test: p = 0.60. 8 GPs excluded because of missing values.
The following variables were rejected during the univariate selection process (p>0.15): type of practice, occasional practice of alternative medicine, personal history of depression/psychotherapy/antidepressant treatment and participation in a formal mental health care network, density of private-practice psychiatrists in each GPs' local area.
The following variables were rejected by the backward process in the final multivariate model: CME on depression diagnosis and management, family history of depression, personal history of consultation of a mental health specialist, thinks that antidepressant are over-prescribed in France, and GPs' opinions regarding barriers to access to psychotherapy in their practice area.