Table 1.

Logistic regression results

	Antidepressant prescription filled within 30 days of initial diagnosist			Paroxetine within 30 days of diagnosis, conditional on AD receipt			Fluoxetine within 30 days of diagnosis, conditional on AD receipt			At least two outpatient visits within 30 days of AD initiation, conditional on AD receipt
	Beta	Robust standard error		Beta	Robust standard error		Beta	Robust standard error		Beta	Robust standard error
Age	0.139	0.007	***	0.017	0.021		−0.038	0.015	**	0.011	0.013
Male	−0.227	0.033	***	0.033	0.092		−0.279	0.073		−0.078	0.062
Region:
North east	omitted			omitted			omitted			omitted
North central	0.137	0.060	**	−0.209	0.160		−0.157	0.125		0.120	0.110
South	0.234	0.060	***	−0.337	0.161	**	−0.250	0.126	*	−0.112	0.111
West	−0.167	0.069	**	−0.223	0.193		0.063	0.145		−0.177	0.133
Episode Start:
5/1/01-6/18/03	omitted			omitted			omitted			omitted
6/19/03-10/27/03	0.084	0.056		−1.003	0.154	***	−0.168	0.134		−0.027	0.101
10/28/03-3/22/04	0.055	0.064		−1.277	0.193	***	0.092	0.140		−0.051	0.113
3/23/04-9/12/04	−0.030	0.054		−1.626	0.195	***	0.447	0.114	***	0.023	0.099
9/13/04-10/30/05	−0.172	0.039	***	−1.560	0.125	***	0.807	0.083	***	0.061	0.073
Constant	−3.099	0.127	***	−1.437	0.357	***	−0.932	0.266	***	−1.013	0.234	***

NOTE: Significance levels on episode start date coefficients noted above indicate whether there are significant differences compared to omitted category. Significance levels noted in figures indicate whether there are significant differences from preceding time period using Wald test of equality of coefficients.

^tRegression predicting antidepressant prescription filled includes dummy variables to control for seasonal effects.

^*p<.10;

^**p<.05;

^***p<.01