Associations Between Comorbidity and Depressive Symptoms During COVID-19: Variation by Social Isolation?

Abstract Adults with comorbidities are at high COVID-19 risk and may experience elevated depressive symptoms during the pandemic. We aimed to investigate the associations between comorbidity at pandemic onset and subsequent depressive symptoms and whether social isolation modified this association. Data were from monthly online questionnaires in the COVID-19 Coping Study of US adults aged ≥55 from April/May-September/October 2020 (n=4,383). Depressive symptoms were measured by the 8-item CES-D, and social isolation as “high” vs. “low” based on contact with family, friends, social organizations, and living alone. In multivariable mixed-effects models, comorbidity (≥2 vs. <2 chronic conditions) was associated with greater depressive symptoms at baseline (β=0.50; 95% CI: 0.36-0.64), this association varied negligibly by social isolation. Differences in depressive symptoms by comorbidity status at pandemic onset were consistent over the six-month follow-up. This study indicates that middle-aged and older US adults with comorbidities experienced persistently elevated depressive symptoms during the COVID-19 pandemic.

this gap, a total of 112 members from four Village programs in a Midwest Metropolitan area completed a cross-sectional pilot survey on their social well-being between January and March 2020. The age of participants ranged from 51 to 92 years old (M=72.30, SD=8.38), over 74% of participants were female and over 88% of participants identified as White/Caucasian. The relationship between the 20-item UCLA loneliness scale and length of Village membership was roughly linear according to the Loess Curve. The scores of the UCLA scale range from 20-80 and higher scores indicate higher loneliness. The Cronbach's alpha of the UCLA loneliness scale was 0.86 in the sample, indicating good internal consistency. The average loneliness score of the sample was 38.45, resembling the average of community-living older adults. Regression results suggested that a one-year increase in village membership was associated with approximately two points reduction in loneliness, holding all else constant. Being female, a racial/ethnic minority, retired, a driver, and having higher frequencies of socializing with friends and neighbors were associated with lower levels of loneliness among Village members. This pilot study provides initial support for the social impact of Villages and informs future larger sample longitudinal studies.

SOCIAL ISOLATION AND WELL-BEING AMONG MIDDLE-AGE AND OLDER ADULTS: BEFORE AND DURING THE COVID-19 OUTBREAK Chair: Lydia Li
This symposium brings together five studies that examined the relationship between social isolation and well-being. Two used pre-COVID data from the Health and Retirement Study (HRS). One aimed to identify patterns of social isolation trajectory in a 9-year period, where social isolation was conceptualized as a multidimensional construct. It identified four distinct patterns, and the pattern had a gradient relationship with health outcomes. Another examined the association between self-perceptions of aging (SPA) and social well-being among older adults. It found that positive SPA predicted increased social connectedness and reduced loneliness in four years. Two other studies were based on a longitudinal survey (COVID-19 Coping Survey) that began in April 2020. One reports that adults 55+ with comorbidity at pandemic onset had persistently elevated depressive symptoms in a 6-month period, regardless of their social isolation level. Another paper suggests that physical isolation at pandemic onset was associated with elevated symptoms of depression, anxiety, and loneliness throughout the following six months. The fifth paper was based on two-wave data-2019 survey and 2020 COVID supplement-from the National Aging and Health Trend Study (NAHTS). It found that older adults who were very socially isolated and completely homebound before the pandemic experienced less psychological distress during the outbreak than those who were very socially integrated and not homebound. The five studies highlight the multiple dimensions of social isolation, their antecedents and development over time, and their role in shaping mental health in a pandemic context.

SOCIAL ISOLATION PATTERNS AND HEALTH OUTCOMES IN MIDLIFE AND LATER LIFE Meng Sha Luo, Zhejiang University, Hangzhou, Zhejiang, China (People's Republic)
Conceptualizing social isolation as a multidimensional construct encompassing social networks, social contacts, perceived support and loneliness, this research aims to: (1) identify patterns of social isolation trajectory among middleaged and older adults in the U.S.; (2) investigate how different patterns of social isolation trajectory are related to adults' physical, mental, cognitive, and overall health. Latent class growth modeling was used to examine social isolation trajectory patterns over nine years in a national sample of 6,457 adults aged 51+. Four patterns of social isolation trajectory were identified: severely isolated, moderately isolated, subjectively integrated, and objectively integrated. The objectively integrated group reported the best physical, mental, cognitive, and overall health, whereas the severely isolated group reported the worst. The moderately isolated and subjectively integrated groups fell in between, with the latter displaying relatively better health outcomes. Findings support a close relationship between poor health and long-term social isolation.

EFFECTS OF SELF-PERCEPTIONS OF AGING ON SOCIAL DISCONNECTEDNESS AND LONELINESS IN OLDER ADULTS Rita Xiaochen Hu, Social Work, Ann Arbor, Michigan, United States
Research shows that self-perceptions of aging (SPA) predict physical, mental, cognitive, and emotional well-being in older adults. Few studies have examined SPA's effects on social well-being. Using data from the 2014-2018 Health and Retirement Study, we examined SPA's effects on older adults' social connectedness and loneliness (age 65+, N = 3,808). SPA was measured by eight items. Social connectedness was operationalized by network size, social contact, and social participation. The UCLA Loneliness Scale assessed loneliness. Linear regression results show that more positive SPA is correlated with increased social connectedness (b = 0.05 SE = 0.01 p = 0.0003) and decreased loneliness (b = -0.09 SE = 0.02 p < 0.0001) in four years, controlling for sociodemographic and health characteristics. Further, loneliness is a mediator between SPA and social connectedness. Findings suggest that older adults with negative SPA are at risk of both objective and subjective social isolation. Adults with comorbidities are at high COVID-19 risk and may experience elevated depressive symptoms during the pandemic. We aimed to investigate the associations between comorbidity at pandemic onset and subsequent depressive symptoms and whether social isolation modified this association. Data were from monthly online questionnaires in the COVID-19 Coping Study of US adults aged ≥55 from April/May-September/October 2020 (n=4,383). Depressive symptoms were measured by the 8-item CES-D, and social isolation as "high" vs. "low" based on contact with family, friends, social organizations, and living alone. In multivariable mixed-effects models, comorbidity (≥2 vs. <2 chronic conditions) was associated with greater depressive symptoms at baseline (β=0.50; 95% CI: 0.36-0.64), this association varied negligibly by social isolation. Differences in depressive symptoms by comorbidity status at pandemic onset were consistent over the six-month follow-up. This study indicates that middle-aged and older US adults with comorbidities experienced persistently elevated depressive symptoms during the COVID-19 pandemic.

. University of Michigan, Ann Arbor, Michigan, United States
The ongoing COVID-19 pandemic has set an urgent need to understand the impact of physical isolation on mental health. We aimed to investigate the relationships between physical isolation during the period when many US states had shelter-in-place orders (April-May 2020) and subsequent longitudinal trajectories of mental health in middleaged and older adults (aged 55+, N=3,978) over a six-month follow-up (April to October 2020). We used population and attrition-weighted multivariable linear mixed-effects models. At baseline, 7 days/week of physical isolation (vs. 0 days/ week) was associated with elevated depressive symptoms (β=0.82; 95% CI: 0.04-1.60), and all of 1-3, 4-6, and 7 days/ week of physical isolation (vs. 0 days/week) were associated with elevated anxiety symptoms and loneliness. Physical isolation was not associated with changes in mental health symptoms over time. These findings highlight the need to prioritize opportunities for in-person connection for middleaged and older adults when safe to do so.

PRE-PANDEMIC SOCIAL ISOLATION: PROTECTION OR VULNERABILITY IN THE TIME OF COVID? Lydia Li, University of Michigan, Ann Arbor, Michigan, United States
How do older adults at risk of social isolation before the pandemic fare during the COVID-19 outbreak? Using data from two waves (Round 9 [2019] and COVID-19 Supplement) of the National Health and Aging Trend Study (NHATS), we examined the relationship between pre-pandemic social isolation and psychological distress during the outbreak among community-living older adults (age 65+). Results show that the most socially integrated respondents had more PTSD (β=1.47, SE=.37, p<.001) and depression/anxiety (β=.34, SE=.11, p=.002) symptoms than the most isolated. Older adults who were not homebound had more PTSD (β=2.0, SE= .76, p=.01) and depression/anxiety (β=1.05, SE=.20, p<.001) than the completely homebound. With shelter-in-place and social distancing requirements, older adults who have been socially active and integrated may experience high-stress levels and may need extra support to adjust to the changes. Relatively, those who have been very isolated and homebound may experience fewer changes in their lives. Social isolation and loneliness are pressing health concerns in older adults, likely exacerbated by social distancing guidelines enacted during COVID-19. Creating effective interventions to address health issues is challenging. Design is an alternative approach to create innovative interventions and to test their preliminary potential. In the present case study, we describe the processes and outcomes of a four-week project in a graduate design studio. Students were asked to develop a prototype for an intervention using digital technologies to increase social connectedness among older adults. This was an interdisciplinary process guided by faculty with expertise in design (Mejía), healthcare redesign (Doebbeling), and gerontology (Der Ananian). In the first week, the faculty helped the students understand the design goals, the implications of social isolation and loneliness, and technology use in older adults. In the second week, students conducted user interviews. In the third week, students set the problem by defining a specific potential audience and context. They also prototyped two preliminary concepts using storyboards and received feedback from the faculty. In the last week, students presented refined prototypes with storyboards, user flows, and interface mockups. Student design ideas included an audio story-sharing app that facilitates conversations and new friendships, an assistance digital service for immigrant older adults that need support with language or cultural challenges, and an art and crafts subscription service with a sharing platform to connect older adults with similar interests. The students' design projects provided innovative technological approaches for improving social connections and could be used in future R&D.

LATE WORKING LIFE PATTERNS IN SWEDEN
Gülin Öylü, 1 Susanne Kelfve, 2 and Andreas Motel-Klingebiel, 3 1. LinkÃ ¶ping University,Norrköping,Ostergotlands Lan,Sweden,2. Linköping University,Norrkoping,Ostergotlands Lan,Sweden,3. Linköping University,Linkoping University,Ostergotlands Lan,Sweden Late working life patterns differ across different social groups and birth cohorts. The mechanisms of these participation differences and role of working life policies can be understood better by using additional working life indicators and historical perspective. This paper investigates how late working life patterns of different age, gender, education groups and birth cohorts are structured in Sweden using participation, employment type, employment break and exit trajectories of different groups. Using Swedish National Innovation in Aging, 2021, Vol. 5, No. S1