New research moves toward personalized treatment for depression

Depression involves a complex interplay of psychological patterns, biological vulnerabilities and social stressors, making its causes and symptoms highly variable. Equally complex is the treatment of depression, which requires a highly individualized approach that may involve a combination of medication, psychotherapy and lifestyle changes.

In a decade-long multi-institutional study, U of A psychologists teamed up with Radboud University in the Netherlands to develop a precision treatment approach for depression that gives patients individualized recommendations based on multiple characteristics, such as age and gender. Their findings are published in the journal PLOS One.

First-line treatment for depression should not be a one-size-fits-all approach, said Zachary Cohen, senior author on the paper and assistant professor in the U of A Department of Psychology. Unfortunately, he said, the current standard of care largely involves a trial-and-error approach, in which different medications or therapies are tried until an intervention or combination that effectively alleviates symptoms is found. 

About 50% of people don’t respond to first-line treatments for depression. There’s a lot of heterogeneity of treatment response, meaning that there are some people who respond really well and some people who don’t.”

Zachary Cohen, senior author on the paper and assistant professor in the U of A Department of Psychology

The study focused specifically on depression in adults. The research team brought together patient data from randomized clinical trials conducted worldwide that have assessed the efficacy of five widely used depression treatments.

Before treatment, patients were evaluated on a variety of dimensions, including for associated psychiatric conditions such as anxiety and personality disorders, said Ellen Driessen, the study’s lead researcher and assistant professor of clinical psychology at Radboud University. 

“We examined whether people with certain features, like the presence of a comorbid condition, might benefit from one treatment method over the other,” Driessen said. 

The researchers hope their results will lead to the creation a clinical decision support tool, an algorithm that simultaneously considers many variables, such as age, gender and comorbid conditions and the relationships among the variables to create a single recommendation. Once the patient’s variables are fed into the tool, it will generate a personalized recommendation as opposed to a guideline that provides a list of generalized recommendations.

The data that the team generated looked at the patients’ outcomes from clinical trials of antidepressant medications, cognitive therapy, behavioral therapy, interpersonal therapy, and short-term psychodynamic therapy, a form of in-depth talk therapy. 

“Much of the prior work on treatment selection has relied on data from single trials whose sample sizes limit their ability to develop powerful, reliable clinical prediction models,” Cohen said.

The research group spent around 10 years collecting and processing data from over 60 trials involving almost 10,000 patients. Researchers from different parts of the world participated in the initiative by sharing data from their studies. The research group also brought together an international group of scientists from different disciplines to develop the strategy for analyzing the data.

“It has taken about five years just to clean and combine the existing data so we can build a model that’s informed by all the available evidence,” Cohen said. 

“This paper is a protocol, which lays out our plans in detail, but the actual building of the tool is something that we will work on the next year or two,” Driessen said. 

In the future, the team plans to conduct a clinical trial evaluating the benefits of using a clinical decision support tool to help match patients to their optimal treatment. If the results are favorable, the tool could be scaled up and implemented in real-world clinical contexts. The researchers envision the tool to be a simple computer program or web application in which patient information can be entered. 

The team hopes to provide clinicians, people with depression, and society a means to make more efficient use of existing treatment resources and help reduce the immense personal and societal costs associated with depression.

“If the results generalize, this tool has the potential to be globally applicable,” Cohen said. “What’s exciting about the variables that go into this is that they’re relatively straightforward to obtain by self-report questionnaires or clinical demographic features. The cost of implementing this will also be relatively low.”