Methods: This cross-sectional study used extant program evaluation data from a community mental health center in a Midwestern US city. Participants (n=100) were aged 50 or older and experiencing serious mental illness and chronic physical health conditions. Self-report measures were used to collect depressive symptoms (Geriatric Depression Scale-Short Form [GDS-SF]), loneliness (De Jong Gierveld Loneliness Scale [DJGL]), and to screen for trauma history (Primary Care PTSD Screen). Four hierarchal linear regression models analyzed the relationship of loneliness to depressive symptoms.
Results. The mean age of the sample was 59.26 years (SD=6.48) and was primarily African-American (56%) and male (54%). The GDS-SF had acceptable internal consistency (α=.86), as well as the social (α=.78) and emotional (α=.76) loneliness subscales of the DGLS (α=.80). The social and emotional loneliness subscales correlated modestly with each other (r=.38, p<.001), and correlated moderately with overall depressive symptoms (r=.55, p<.001 and r=.59, p<.001, respectively). Model one included demographics (age, education, race, gender) and trauma symptoms (R2adj=.234). The second model included self-rated health (R2adj=.399). The third model included social loneliness (R2adj=.466), and the final model included emotional loneliness (R2adj=.593). The final model revealed emotional loneliness as having the strongest association with depressive symptoms (B=.396, p<.001), compared to social loneliness (B=.184, p=.026). Significant covariates include self-rated health (B=-.315, p<.001), and emotional numbing symptoms associated with trauma exposure (B=.266, p=.004).
Conclusions and Implications: Emotional loneliness is significantly associated with depressive symptoms in middle-aged and older adults experiencing serious mental illness. Several treatment implications arise from: (a) the modest correlation of social and emotional loneliness, and (b) the weaker association of social loneliness with depressive symptoms compared to emotional loneliness. Persons with SMI may be “lonely but not alone”—they may experience social integration, but experience weak attachment and intimate connection with others. Interventions to ameliorate this may target maladaptive social cognition that may be common to both emotional loneliness and depressive symptoms. This study contributes to recent work on loneliness and depressive symptoms by examining their relationship among persons with SMI, who experience unique risks—the interplay of biological factors and stigmatizing expectations and perceptions—which may contribute to more intense experiences of emotional loneliness and depressive symptoms.