Depression Consultant Better treatment through informed communication

CAN CLINICIANS IMPROVE ANTIDEPRESSANT REMISSION RATES WITH BETTER TREATMENT ALGORITHMS?
Richard Metzner, M.D. & Andrew P. Ho, M.D.

SUMMARY:

According to the STAR*D ("Sequenced Treatment Alternatives to Relieve Depression") study group nearly 7 in 10 Major Depressive Disorder patients fail to remit after first-step antidepressant treatment, with few predictors to guide second-step choices. We contrast those findings with the more positive results we obtained in a 33 month study of depressed patients treated in 80 naturalistic primary care and psychiatric settings throughout the United States. Participating clinicians volunteered to receive guidance from an online program called the TTDI ("Targeted Treatment Depression Inventory"), which suggested antidepressants on the basis of patient responses to a 17 item printed questionnaire. The TTDI program was accessed successfully 1,164 times. A sample of 117 patients repeated the TTDI questionnaire between 10 days and 6 months after initial evaluation. Of those 61 were first-step patients. After a mean of 5.7 weeks, they evidenced a 59% remission rate on the TTDI (convergent validity between TTDI and both Hamilton Depression Rating Scale and Zung Self Rating Depression Scale: p<.01). TTDI score improvements were also highly significant (p< .001; paired t tests). Despite the high remission rate achieved by weighing combinations of TTDI responses in recommending antidepressants, analysis of those items using logistic regression did not demonstrate significant predictive value for any of them alone. The conclusions of this study were limited by the fact that that most of the patients were treated without follow-up TTDI testing. Nonetheless, the fact that the TTDI algorithm helped produce a 59% remission rate in a first-step sample without demonstrating predictive value for any specific item, suggests that researchers may need to find other statistical methods to evaluate predictors of antidepressant effectiveness.