Longitudinal Associations Between Loneliness and Depressive Symptoms in Adolescents with ASD

Background: Adolescents with ASD report high rates of loneliness (Bauminger & Kasari, 2000; Bauminger et al., 2003), suggesting that, as a group, these adolescents are sensitive to their peer difficulties. Research with typical adolescents has consistently found a relationship between peer difficulties and symptoms of depression. 

Objectives: Given that adolescents with ASD experience high rates of loneliness and more peer difficulties than typically developing adolescents, we examined whether loneliness predicts depressive symptoms in adolescents with ASD.

Methods: Participants were drawn from an ongoing longitudinal study of individuals with ASD first diagnosed at age 2.  Between the approximate ages of 18 and 22, participants capable of completing self-report measures (n=40) and their parents filled out yearly questionnaire packets. Packets at three of these time points included the Asher loneliness scale and a measure of depressive symptoms (either the Child Depression Inventory or the Beck Depression Inventory). Across these three waves, measures of depressive symptoms were highly correlated (r = .84-.67, p < .001).

Results: Exploratory principal components analysis was performed on all 24 Asher items for data from each of the three time points.  For each of these analyses, only one factor was consistently found across each of the time points. Confirmatory principal components analysis was conducted for the three items that comprised this factor and found a consistent factor at each time point (For each time point, eigenvalue > 2.15, > 71.7% of the total variance explained, and all items loadings > .82, and all alphas > .80). Analyses were conducted on the average of these items for each time point (i.e. I feel alone at school or work, I feel left out of things, I am lonely). Correlations indicated that time 1 loneliness was correlated with depressive symptoms at all three time points (r_s > .43, p_s < .001), time 2 loneliness was correlated with depressive symptoms at times 1 and 2 (r_s > .49, p_s < .001), and time 3 loneliness was not correlated with depressive symptoms at any time point.  Multiple regression analyses tested the direction of these associations and found that loneliness predicted increases in depressive symptoms from time 1 to time 3 (ΔR² = .19, F = 12.58, p < .003), but depressive symptoms did not predict changes in loneliness over the same period (ΔR² = .01, F = 0.20, p ns).  Finally, LCGM found a two class model to be the best fit for the data with one group reporting low levels of loneliness over all three years and the other group reporting high levels of loneliness over all three years. The consistently high loneliness group had higher levels of depressive symptoms than the consistently low loneliness group for each of the three years (t_s > 2.34, p_s < .03). 

Conclusions: Across multiple types of analyses, the findings suggest a predictive link between loneliness and depressive symptoms in adolescents/young adults with ASD.  This underscores the importance of addressing loneliness and other aspects of the peer world in interventions for internalizing issues in this group.