The relationship of loneliness with mental health trajectories in bereaved older adults.

Losing a loved one often results in acute grief and depressive symptoms. The vast majority of older individuals are resilient and return to normal functioning within a year after a loss. However, a significant minority develop mental health complications, including prolonged grief disorder and bereavement-related depression. This can lead to adverse consequences, including physical and cognitive decline, impairment in global functioning, and early mortality. The contributing factors that increase the likelihood of adverse mental health outcomes in bereaved individuals have not been fully elucidated. This study explores the relationship of loneliness with grief and depressive symptom trajectories in older adults after an attachment loss. 45 individuals over age 50 and within 13 months following the death of a loved one completed a comprehensive clinical assessment at the baseline visit (week 0), including the UCLA Loneliness Scale-3, the Inventory of Complicated Grief (ICG), and the Hamilton Depression Rating Scale (HAM-D). All participants completed a 6-month longitudinal study; ICG and HAM-D were completed at weeks 8, 16, and 26 (endpoint). Statistical analyses : Two classes of mixed-effects models, where covariates are fixed effects and patient-specific dependents are random effects, were considered. Cross-sectional models were used to investigate relationships between loneliness and grief and depressive symptoms at baseline. Longitudinal models were used to examine the relationship of baseline loneliness scores with grief and depressive symptoms across the 6-month study duration. In longitudinal models, the interactions between time elapsed since baseline and dynamic measurements (complicated grief and depressive symptom scores) were examined. A hierarchical variance-covariance structure for random effects, informed by clustering of clinical trajectories, was used to determine participant trajectories. Age, gender, days since the loss, and cumulative medical illness rating scale for geriatrics scores were included as covariates in all models. The R statistical software (version 3.6.3) was utilized for data analyses. Cross-sectionally, loneliness was positively associated with scores on the ICG (p<0.01) and HAM-D (p<0.01) in separate models, after adjusting for covariates. However, the loneliness-ICG relationship was no longer significant when HAM-D was included as a covariate in an extended model (p=0.32). Mediation analyses revealed that the loneliness-ICG relationship was mediated by baseline HAM-D scores. In contrast, the loneliness-HAM-D relationship remained significant even after including ICG as a covariate (p=0.01). Longitudinally, baseline loneliness scores were associated with worsening HAM-D scores over time (p<0.01); this association persisted over 6 months (p<0.01). The loneliness-ICG longitudinal relationship was not significant. Of additional interest were the co-movement patterns expressed by HAM-D and ICG: adults following a C-shaped trajectory for HAM-D scores were likely to demonstrate a similar trajectory with ICG scores. Loneliness was independently associated with grief and depression symptom severity in older adults following bereavement. However, the baseline loneliness-grief symptom association was mediated by depression. Since loneliness could predict depressive symptom trajectories in older grieving individuals, interventions targeting loneliness may also improve depression outcomes in this study population. Although the co-movement analysis was exploratory (i.e., this was used to inform the structure of the covariance matrix for random effects in analytic models), it has implications for future treatment intervention study designs. This research was funded by NIMH grants R01 MH122490 and R21 MH109807 (Goveas); Costigan Family Foundation (Goveas); and UL1TR001436 [ABSTRACT FROM AUTHOR]