Association of Social Isolation With Disability Burden and 1-Year Mortality Among Older Adults With Critical Illness

Abstract

Importance: Disability and mortality are common among older adults with critical illness. Older adults who are socially isolated may be more vulnerable to adverse outcomes for various reasons, including fewer supports to access services needed for optimal recovery; however, whether social isolation is associated with post-intensive care unit (ICU) disability and mortality is not known.

Objectives: To evaluate whether social isolation is associated with disability and with 1-year mortality after critical illness.

Design, setting, and participants: This observational cohort study included community-dwelling older adults who participated in the National Health and Aging Trends Study (NHATS) from May 2011 through November 2018. Hospitalization data were collected through 2017 and interview data through 2018. Data analysis was conducted from February 2020 through February 2021. The mortality sample included 997 ICU admissions of 1 day or longer, which represented 5 705 675 survey-weighted ICU hospitalizations. Of these, 648 ICU stays, representing 3 821 611 ICU hospitalizations, were eligible for the primary outcome of post-ICU disability.

Exposures: Social isolation from the NHATS survey response in the year most closely preceding ICU admission, which was assessed using a validated measure of social connectedness with partners, families, and friends as well as participation in valued life activities (range 0-6; higher scores indicate more isolation).

Main outcomes and measures: The primary outcome was the count of disability assessed during the first interview following hospital discharge. The secondary outcome was time to death within 1 year of hospital admission.

Results: A total of 997 participants were in the mortality cohort (511 women [51%]; 45 Hispanic [5%], 682 non-Hispanic White [69%], and 228 non-Hispanic Black individuals [23%]) and 648 in the disability cohort (331 women [51%]; 29 Hispanic [5%], 457 non-Hispanic White [71%], and 134 non-Hispanic Black individuals [21%]). The median (interquartile range [IQR]) age was 81 (75.5-86.0) years (range, 66-102 years), the median (IQR) preadmission disability count was 0 (0-1), and the median (IQR) social isolation score was 3 (2-4). After adjustment for demographic characteristics and illness severity, each 1-point increase in the social isolation score (from 0-6) was associated with a 7% greater disability count (adjusted rate ratio, 1.07; 95% CI, 1.01-1.15) and a 14% increase in 1-year mortality risk (adjusted hazard ratio, 1.14; 95% CI, 1.03-1.25).

Conclusions and relevance: In this cohort study, social isolation before an ICU hospitalization was associated with greater disability burden and higher mortality in the year following critical illness. The study findings suggest a need to develop social isolation screening and intervention frameworks for older adults with critical illness.

Figure 1.. Assembly of the Analytic Sample From the 2011 National Health and Aging Trends Study (NHATS) Cohort

Participants included for the primary analysis of disability count and the secondary outcome of 1-year mortality are shown. ADL indicates activity of daily living; ICU, intensive care unit.

Figure 2.. Unadjusted Count of Disabilities for Older Survivors of Critical Illness Across Levels of Prehospitalization Social Isolation

The blue markers represent the mean count of disabilities reported during the survey preceding intensive care unit (ICU) admission, and the orange markers represent the count of disabilities reported during the survey following ICU discharge. The error bars represent 95% CIs.

Figure 3.. Mortality Rates Over 1 Year Following Admission for Critical Illness

Kaplan-Meier survival curves for each level of social isolation (0 to 6) are shown, with higher scores indicating greater social isolation.