Network analyses reveal which symptoms improve (or not) following an Internet intervention (Deprexis) for depression

Abstract

Depression is a heterogeneous collection of symptoms. Prior meta-analyses using symptom sum scores have shown the Internet intervention, Deprexis, to be an efficacious treatment for depression. However, no prior research has investigated how Deprexis (or any other Internet intervention for depression) impacts specific symptoms of depression. The current study utilizes symptom-level analyses to examine which symptoms are directly, indirectly, or minimally influenced by treatment. Network analysis and mean-level approaches examined which symptoms, assessed by the Quick Inventory of Depression Symptoms (QIDS-SR), were affected by an 8-week course of Deprexis compared to a waitlist in a nationally recruited United States sample (N = 295). Deprexis directly improved the symptoms of sadness and indecision. Change in these symptoms, in turn, were associated with change in self-dislike, fatigue, anhedonia, suicidality, slowness, and agitation. All of these symptoms (except for agitation) show decreases with Deprexis compared to a waitlist after correcting for multiple comparisons. Six additional symptoms, particularly the somatic symptoms, were not impacted by Deprexis compared to waitlist. In this sample, the efficacy of Deprexis was due to its direct impact on sadness and indecision. Examining treatment-related change in specific symptoms may facilitate a more nuanced understanding of how a treatment works compared to examining symptom sum scores. Symptom-level approaches may also identify symptoms that do not improve and provide important direction for future treatment development.

Click the Slides button above to demo Academic’s Markdown slides feature.

Supplementary notes can be added here, including code and math.

Avatar
Michael Mullarkey
PhD Candidate in Clinical Psychology and Clinical Intern

My research interests include developing brief interventions, predicting treatment response, and symptom level analysis