Teenage depression treatment: Medication, therapy or both? It may depend on marital conflict.

By Nestor Lopez-Duran PhD

Recently I discussed a study that examined the rates of psychiatric conditions in children and adolescents. I mentioned how by age 18, about 16% of girls and 8% of boys will experience a depressive disorder (major depressive disorder or dysthymia). Depression is one of the most prevalent psychiatric conditions in adolescents and unfortunately it is not just a temporary phase. Research suggests that depression that starts in childhood or adolescence is actually more chronic and more impairing than adult onset depression.

Although we have some effective interventions for depression, including medications, not all teens respond to these interventions. Because of this finding, researchers have attempted to identify factors that may make a child more or less likely to respond to the intervention. This month the Journal of Abnormal Child Psychology published a large study examining whether parental conflict negatively impacted the effectiveness of interventions for depressed adolescents. Given that parental conflict is one factor associated with depression in teens, it was sensible to think that depressed kids living in families with high parental conflict would be less likely to respond to interventions than depressed kids living in low conflict families.

Is this the case?

This study included 260 adolescents with a depressive disorder living in two parent households (122 girls and 138 boys). These teens were randomly assigned to one of four treatment groups:

1. Fluoxetine (Prozac)

2. Cognitive Behavior Therapy (CBT)



3. Fluoxetine + CBT

4. A placebo (fake pill)

Below you can see the basic results across all groups.

As you can see, the combination of CBT and Fluoxetine was just as effective as Fluoxetine alone, and both of these treatments were more effective than CBT alone and the placebo. Sadly, CBT alone was not any better than the fake pill. Thus across all participants, only Fluoxetine made a significant difference. Parental marital discord did not change these results.

However, a slightly different pattern emerged when looking at the effect of parental marital discord separately for boys and girls.

The next graph shows the effect of the interventions for girl with high levels of parental marital discord.

The results for this group were identical to those for the entire sample. The combination of CBT and Fluoxetine or Fluoxetine alone was very effective. CBT alone was not significantly different from the placebo.

Now here we can see the effect of the therapy for girls living in families with low parental marital discord.

As you can see, the results are a bit striking. Only 20% of those receiving CBT alone actually improved. That compares to 36% of those receiving a fake pill. The fluoxetine alone and the combination of CBT/Fluoxetine were significantly better than CBT alone.

What does this mean? Before we try to interpret these findings lets talk about what is truly happening in the placebo (fake pill) condition. Most people interpret the response to the fake pill as the placebo effect. That is, we get better because we think or hope that the intervention (fake pill) is supposed to help even though there is no real mechanism by which the intervention (fake pill) could actually help. We always include a placebo condition when examining the efficacy of an intervention because we want to know if the intervention is actually working BEYOND what is expected from the patients high hopes and beliefs.

But in the case of depression the response to the fake pill is not only due to the placebo. It is actually due to the placebo effect PLUS the effect of time. Depression episodes come in cycles and although adolescent depression is chronic (many depressed adolescents will continue to experience depressive episodes throughout their lives), many people will eventually come out of an acute episode given enough time.  Thus, when we see that 35% of depressed kids get significantly better after 12 weeks of taking the fake pill we have to remember than some of these kids would have improved significantly even if they had just continued with their normal life without taking the fake pill.

So given these thoughts, in this study the story about CBT as treatment of adolescent depression is actually quite sad. Not only was CBT not effective in the treatment of depressed teens, but actually it was a very poor placebo among girls in low conflict households. Thus, depressed teens, and especially girls in low parental conflict households, seem to respond to Prozac but not to CBT.

Now, the discussion among clinical researchers on this issue is that this finding is not really representative of CBT because in this study the clinicians used a specific version of CBT called TADS CBT. Some argue that other CBT versions have actually been found to be effective in treating depressed adolescents, and by effective I mean better than a placebo (but not necessarily better than fluoxetine). This is true. Some studies have shown that some versions of CBT are actually very effective in treating depressed teens. However, in practice this is bad news because it means that details matter. It means that not ANY version of CBT would work and the minor nuances between one manual and another (one version of CBT and another) actually make the difference between being effective or being potentially worse than a sugar pill!

Why bad news?

Because even among clinicians who say they use CBT, the use of a manualized approach is rare at best (I say most likely non-existent). Instead, most CBT clinicians use their own flexible version of CBT that borrows from many different CBT versions. The problem with this method is that parents have no way of knowing if the hybrid version of CBT that their kids therapist is using is actually the one that works. Clinicians have argued that details do not matter and that as long as they use principles of CBT the work they do is effective. But the research clearly suggests that this is not true.

There is one last caveat though. This study examined the effect of the interventions at 12 weeks after starting treatment. Thus, these results dont tell us much about whether these interventions prevented relapse or at least shortened the time to the next depressive episode. Thats a discussion for another day, but sadly the data on this last issue are not encouraging at all.

Disclaimer: I do not prescribe medications in my clinical work (Im a clinical researcher) and do not receive any royalties or payments from any pharmaceutical company.

The reference: Amaya, M., Reinecke, M., Silva, S., & March, J. (2010). Parental Marital Discord and Treatment Response in Depressed Adolescents Journal of Abnormal Child Psychology DOI: 10.1007/s10802-010-9466-2