Examining varying number of intervention-modules and change in individual networks in young people remitted from depression or anxiety: Exploratory outcomes of the StayFine randomized clinical trial

In network theory, symptoms of mental disorders constitute networks of interconnected nodes. Interventions can target (aspects of) networks. Assuming that combined intervention-modules target multiple aspects of networks simultaneously, we examined on average whether receiving more intervention-modules was associated with more baseline to post-intervention change in network characteristics. Between Dec-2019 and May-2024, 14-to-22-year-old individuals (89.8% female) remitted from depressive or anxiety disorders were randomized (1:1) to relapse prevention intervention-modules plus care as usual (M + CAU; n = 69), or CAU ( n = 68). Participants completed baseline and post-intervention ecological momentary assessment (EMA; 2 weeks, 6 × /day) to construct contemporaneous partial correlation networks of 11 nodes ( anxiety, sadness, positive affect, anger, stress, fatigue, suppression, avoidance, loneliness, activity, social company ). Based on a data-driven advice, including the baseline network, six intervention-modules were offered in personalized varying combinations (psychoeducation, cognitive restructuring, positive affect, exposure, activation, sleep, wellness, relapse prevention plan). Using 45 univariate regression analyses ( α = .05), associations were assessed between the number of completed intervention-modules (M[SD] = 4.29[1.70]; range = 0-6) and change in network density, and in intensity, instability, and centrality of 11 nodes. Randomization groups were compared using two-sample t-tests. Results indicated that after multiple testing correction, a higher number of intervention-modules was not associated with significant change in network characteristics. Changes in networks were neither different between randomization groups. Some characteristics showed non-robust, small associations with a higher number of intervention-modules: centrality of fatigue decreased ( n = 69; β = −0.061, 95%CI [-0.122,0.000], p = .037, η p 2 = 0.03), and network density decreased ( n = 69; β = −0.008, 95%CI [-0.015,0.000], p = .015, η p 2 = 0.05). Change in centrality of loneliness was positive in the M + CAU-group and negative in the CAU-group (t[130.33] = -2.47, p = .015, d = −0.42). In conclusion, after multiple testing correction, more intervention-modules were not associated with more change in network characteristics across individuals. This initial examination provides a starting point for replication in confirmatory analyses, and for better mechanistic understanding of network change to prevent relapse over time.