Replication Data for: A brief but comprehensive three-item social connectedness screener for use in social risk assessment tools (PLOS One article)

Study variables and variable descriptions for variables used for the above named article. Abstract Background: The 2014 Institute of Medicine report “Capturing Social and Behavioral Domains and Measures in Electronic Health Records” described three subdomains of social relationships that affect patient health and well-being. However, most social risk screeners currently assess only one subdomain, frequency of social connections. We are proposing a three-item Brief Social Connectedness (SC) screener that additionally assesses risks in social/emotional support and loneliness/social isolation subdomains. Methods: We used data from 2244 Kaiser Permanente Northern California (KPNC) members ages 35-85y collected in a 2021 social risk survey. The survey included three validated questions that covered the three SC subdomains (frequencies of social contacts with people they care about, feeling lonely/socially isolated, and getting enough social/emotional support). Variables representing moderate/high versus low risk were created for each subdomain. Data weighted to the age-sex-racial/ethnic composition of the KPNC membership were used for bivariate analyses and modified log-Poisson regression models (MLPR) that adjusted for age, sex, and race/ethnicity that examined cross-sectional relationships among the three subdomain risks, as well as with two structural SC risks, living alone and not being in a committed relationship. We then used MLPR to study cross-sectional associations of these five SC variables with three single-item self-report measures of emotional health. Results: In models that included all five SC variables, loneliness/social isolation and social/emotional support risks were significantly associated with all three emotional health measures, while frequency of social contacts, living alone, and no committed relationship were not. However, low frequency of social contacts and no committed relationship significantly increased risk of often feeling lonely/socially isolated and lacking in social/emotional support. Conclusions: Without substantially lengthening social risk screening tools, a three-item screener assessing risks of loneliness/social isolation, inadequate social/emotional support, and low frequency of social contacts could provide more comprehensive information about risk in the social connectedness domain.