Getting Under the Skin: Exploring Physiological Indicators of Program Engagement in the Early Adolescent Coping Power Program

Introduction: Program engagement, which includes uptake and responsivity of program components, has been linked with improved outcomes in multiple trials and implementation studies. However to date, most studies have used traditional participant or clinician rating scales as indicators of engagement in evidence-based programs. Advances in wearable technologies highlight the potential of physiological measures as an emerging, objective approach to assess intervention engagement. Yet scant research has examined how discrete indicators of physiological arousal, such as heart rate variability and galvanic skin response, correspond to engagement in preventive interventions. The current paper explored the extent to which these physiological measures of arousal corresponded with traditional measures of youth and clinician rated engagement across 10 sessions of an indicated preventive intervention.

Methods: Data for this study come from a subset of youth participating in a 40 middle school randomized controlled trial (62% male; 75% African American) testing the early adolescent version of Coping Power (Lochman & Wells, 2004). Coping Power is an indicated preventive intervention which targets youth with aggressive behavior problems using a clinician-facilitated small group format delivered weekly within the school across the course of the school year. During 10 of the 25 sessions, students reported perceived engagement using an engagement survey (alpha = .89), and psychological state using the Positive and Negative Affect Scales (PANAS; Watson et al., 1988; positive affect alpha = .81; negative affect alpha = .84); they also wore a Fitbit-like wristband device, which collected students’ blood volume pulse (used to generate indicators of heart rate variability) and electrodermal activity (used to assess galvanic skin response, including skin conductance). Clinicians also provided ratings of student engagement.

Results: Engagement and positive affect were significantly positively correlated (r = .656, p <.001) and engagement and negative affect were inversely correlated (r = -.137, p =.003). Heart rate variability is calculated using the MindWare software and will be examined in relation to the self and clinician rated engagement and PANAS data. We hypothesize that skin conductance will be associated with student self-reported positive affect and engagement; we further hypothesize that one indicator of heart rate variability (LF/HF) will be associated with negative affect and another (NN50) will be associated with a domain of engagement related to comfort with the intervention (Pietro et al., 2014).

Conclusions: This study aims to enhance the field’s understanding of the potential utility of physiological measures to assess youth engagement in preventive interventions. Implications for best practice in using wearable measures as an engagement indicator will be considered.