Psychometric properties of the social isolation and social network scale in community-dwelling older adults: Construct validity, reliability, and sensitivity

Abstract

Background Social isolation is associated with numerous negative health outcomes in older adults. Despite growing concerns about social disconnection, there remains a need for psychometrically sound instruments that comprehensively assess both social isolation and social networks, particularly in culturally diverse contexts. Objective This study aimed to evaluate the psychometric properties of the newly developed Social Isolation and Social Network (SISN) Scale among community-dwelling older adults in South Korea. Methods A total of 350 community-dwelling older adults aged >65 years completed the SISN and Korean version of the Lubben Social Network Scale (LSNS-K). Psychometric properties were assessed using reliability analysis (internal consistency and test-retest), construct validity (confirmatory factor analysis), and diagnostic accuracy (receiver operating characteristic [ROC] curve analysis). Results The SISN demonstrated excellent internal consistency (Cronbach’s α = 0.94) and test-retest reliability (intraclass correlation coefficient = 0.929, 95% confidence interval [CI]: 0.902–0.948). Confirmatory factor analysis supported the two-dimensional structure of social isolation (χ² (14) = 38.151, root mean square error of approximation [RMSEA] = 0.070, comparative fit index [CFI] = 0.983, Tucker–Lewis index [TLI] = 0.974) and social network (χ² (44) = 106.295, RMSEA = 0.064, CFI = 0.976, TLI = 0.964). The SISN showed a strong correlation with the LSNS-K (r = 0.785, p < .001). ROC curve analysis revealed gooddiscriminative ability (area under the curve = 0.900, 95% CI: 0.866–0.933), with an optimal cutoff score of 3.24 (sensitivity 81.8%, specificity 88.2%) for identifying social isolation risk. Conclusions The SISN provided strong preliminary evidence of reliability and validity as a measure of social isolation and networks among older Korean adults. While these findings support its potential for research and clinical applications, the limited representativeness of the sample warrants cautious interpretation. Future studies with more diverse populations are recommended to strengthen generalizability.