Use of antidepressant medications: are there differences in psychiatric visits among patient treatments in the Veterans Administration?

BACKGROUND

Information on the effectiveness of newer antidepressants like serotonin-norepinephrine reuptake inhibitors in terms of healthcare utilization is limited. Treatment guidelines affect evaluation. Second-line medications are usually prescribed to patients with higher utilization.

OBJECTIVES

The objective of this study was to compare antidepressants within the Veterans Affairs (VA) healthcare system on the basis of the number of outpatient psychiatric visits for each class of antidepressants.

RESEARCH DESIGN AND SUBJECTS

We conducted a retrospective cohort design using precollected information from VA national databases from 1999 and 2000. The study identified 92,537 patients on serotonin specific reuptake inhibitors (SSRIs), serotonin-norepinephrine reuptake inhibitors (SNRIs), or tricyclic antidepressants (TCAs).

MEASURES

We stratified individual patients by the number of visits in the baseline year for each medication class. For each stratum, we created a dichotomized variable YK: Yk = 1 if there is a reduction of K visits or more and 0 if there is no reduction. We calculated the odds of reduction of psychiatric visits among the 3 classes of antidepressants.

RESULTS

TCAs and SSRIs were associated with greater odds of reduction compared with SNRIs at the level of 1 through 10 or 11 visits, respectively. SNRIs were associated with greater odds of reduction in visits at the level of 14, 16, or more visits compared with SSRIs and TCAs, respectively (P <0.05). SSRIs were associated with greater odds of reduction compared with TCAs at the level of 1 to 11 visits (P <0.05); there were no significant differences between the 2 classes above 11 visits.

CONCLUSION

Effectiveness research using databases should consider how medications are prescribed within systems. Treatment guidelines result in differences in severity and utilization among users of different medications.