Are Network Characteristics Prognostic for Long-Term Outcomes in Depression?

No significant association between higher network strength and poorer outcomes
No significant association between higher network strength and poorer outcomes

HealthDay News — Adolescents with depression who are poor responders to treatment do not have more densely connected networks, according to a research letter published online November 29 in JAMA Psychiatry.

Among adolescents with depression who participated in a psychological treatment trial, Lizanne Schweren, Ph.D., from the University of Cambridge in the United Kingdom, and colleagues replicated a study conducted in adults that showed that adult patients with depression who continue to experience problems in subsequent years have more densely connected networks at baseline. A total of 465 adolescents completed the 33-item Mood and Feelings Questionnaire (MFQ) prior to treatment, with regular additional assessments up to 12 months after treatment. Relatively good and poor responders were differentiated by the median percentage change from baseline to final follow-up of the MFQ summary score. 

Related Articles

The researchers found that good responders had higher mean MFQ summary scores at baseline and higher mean levels of suicidality than poor responders. A higher global network strength was seen in poor responders, although the difference was not significant (good responders, 3.6; poor responders, 4.3; P=0.15). Local node strength did not differ, except for that of concentration problems, which was more connected to other nodes in poor versus good responders. The findings were similar when treatment response was defined as below clinical threshold at the final follow-up.

"With network analyses taking an astonishing flight in psychiatry, we recommend cautious application of group-level network density as a prognostic marker," the authors write. "Crucial steps to be taken by the field include further replication studies as well as in-depth psychometric evaluation of the reliability and clinical correlates of network parameters."

One author disclosed financial ties to Lundbeck.

Abstract/Full Text (subscription or payment may be required)