I also suspect that the STARD participants were not "tested" for compliance, in other words, blood levels of the drug were not taken.  Dr. Watson

New York Times, October 2, 2007

Talk Therapy Pivotal for Depressed Youth , By BENEDICT CAREY

A talking cure for depression called cognitive behavior therapy appears to cancel the risk of suicidal thinking or behavior associated with taking antidepressant medication, according to the most comprehensive and long-running study to date of depression treatment among adolescents.

The study, which followed for a year more than 600 adolescents being treated for chronic depression, found that four in five recovered entirely, or nearly so, when treated over nine months with medication, talk therapy or a combination of the two.

Patients taking medication showed significant signs of improvement up to six weeks earlier than those who received talk therapy alone, but were about twice as likely to report feeling suddenly suicidal. The combination of the two therapies, the authors found, produced the most rapid recovery and protected against sudden suicidal urges.

For several years experts have been debating the risks to children and adolescents who take antidepressants like Prozac and Paxil. In 2004, health regulators required that all labels for antidepressants carry prominent warnings that the drugs were associated with increased risks of suicidal thinking and behavior in young patients, a link that many psychiatrists say has been blown out of proportion, scaring off patients who could benefit from drug treatment.

In this study, antidepressants lowered the risk of suicidal thoughts and actions over all, but significantly less so than talk therapy.

“What this study shows, convincingly and for the first time, is that there are very good options for a child who is thought to be at risk for suicidal thinking,” said Kevin Stark, a psychologist at the University of Texas, who was not involved with the research. “Psychosocial therapies do work on their own, with time. But they also help prevent relapse, and this shows that they can help make drug treatment safer.”

In the study, which began in 1999, researchers recruited 654 youths ages 12 to 17 who had been moderately to severely depressed for up to a year or longer. The adolescents were randomly assigned to be treated with Prozac, the antidepressant made by Eli Lilly; cognitive behavioral therapy for a weekly hourlong session; placebo pills; or a combination of Prozac and talk therapy.

After 12 weeks, about three in four of the patients receiving both talk therapy and medication were rated as “much better” or “very much better,” and two-thirds taking just the drug fared just as well. Talk therapy by itself was no better than the placebo.

After four months, about two-thirds of those receiving any treatment were rated as much or very much improved — significantly better than a typical response to placebo pills.

By nine months, 8 in 10 adolescents had shaken off their depression, entirely or almost entirely, no matter the treatment.

Talk therapy was a safer alternative. Almost 15 percent of the patients taking just Prozac reported what were described as “suicidal events,” mainly talk and thoughts of suicide so alarming that doctors called in the patients and, often, altered dosages.

The rate of such events for those receiving just cognitive behavior therapy was 6 percent. The results for combination therapy were about the same.

“The message is that medication accelerates recovery, but cognitive therapy protects against these bad reactions, and the combination is the best option,” said Dr. John March, chief of child and adolescent psychiatry at the Duke University Medical Center and the principal investigator for the study.

The talk therapy promoted changes in behavior like getting patients out of bed and doing something that they enjoy, like playing basketball or going to a party. It also provided cognitive therapy, in which patients are taught to diffuse poisonous assumptions like “I’m a loser” or “I’ll never get a girlfriend.”

Experts say it is not easy to find specialists in this therapy outside large cities. The techniques have been widely published in manuals and books, and Dr. March said a good therapist could usually work such techniques into a treatment plan.

“The trick,” he said, “is to be an intelligent consumer and find a skilled therapist who’s willing to work with you on these methods.”

Benedict Carey writes,
"The rate of such [suicidal] events for those receiving just cognitive behavior therapy was 6 percent. The results for combination therapy were about the same."

Not really "about the same."  The actual percentage of patients experiencing suicidal events were:
14.7% of those receiving Prozac alone.
8.4% of those receiving Prozac and talk therapy.
6.3% of those receiving talk therapy alone.

Treatment WITHOUT medication is safer, even according to a biased study like this one!
looked at another way: 

15% of the patients taking Prozac alone were suicidal during treatment (not BEFORE treatment)

6%  of those receiving talk therapy alone.

 for those receiving Prozac + talk therapy,  the rate of suicidality fell from 15% to 8%.

 conclusion:

1)  Prozac doubled the risk of suicide in this study

2) talk therapy may partly (but not completely) offset the suicide-enhancing effects of Prozac.

On the latest Medscape (Feb. 2009), this highlight came out of some new TADS data...

"The most recent results from the Treatment for Adolescents with Depression Study (TADS) show that the overall remission rate at 36 weeks was about 60%. The rates were similar in each of the 3 treatment groups: antidepressant fluoxetine alone (55%), cognitive behavior therapy (CBT) alone (64%), or a combination of these 2 therapies (60%)"

CBT alone did much better than prozac in remission rates, but its effectiveness was slightly diminished when combined with prozac.  It would be nice if the sample had been bigger to see if this effect would have been enhanced.