Association of social isolation with overall survival in cancer patients.

12011 Background: Social isolation has been associated with increased all-cause mortality, particularly in cardiovascular and mental health conditions. Prior studies show that larger social networks, greater perceived social support, and marital status are associated with improved survival. However, the impact of social isolation on cancer patients remains incompletely characterized. This study examines the association between social isolation and overall survival (OS) in a large cancer cohort. Methods: 47,562 adult cancer patients treated at British Columbia Cancer between April 1, 2011, and December 31, 2016, who completed the Psychosocial Screen for Cancer (PSSCAN-Revised) questionnaire at their initial visit were included. Patients with prior malignancies were excluded. OS was estimated using Kaplan-Meier methods and compared using log-rank tests. Multivariable Cox proportional hazards regression was used to identify independent associations with OS. Results: The mean age was 64.5 years (SD 13.9). 55.2% were female. At their initial visit, 67% had non-metastatic disease, 17% had metastatic disease, and metastatic status was unknown in 16%. Social isolation was prevalent: 24.7% lived alone, 14.1% had lost a partner, 11.6% lacked assistance with instrumental activities of daily living (IADLs), 4.7% lacked regular social contact, 6.9% lacked emotional support, and 29.3% had at least one domain of social isolation. Median OS was shorter across all five domains of social isolation: patients living alone (37 vs 58 months), those who lost a partner (34 vs 56 months), those without regular social contact (32 vs 52 months), those without IADL support (49 vs 53 months), and those lacking emotional support (45 vs 53 months; all p<0.05). Anxiety (36.8%) and depression (25.6%) were also associated with worse survival (p<0.001). In multivariable Cox regression, independent predictors of increased mortality included metastatic disease (HR 4.41), depression (HR 1.44), anxiety (HR 1.29), lack of regular social contact (HR 1.26), male sex (HR 1.22), living alone (HR 1.18), and older age at presentation (HR 1.04 per year; all p<0.001). Conclusions: Multiple domains of social isolation were independently associated with shorter OS in cancer patients. These findings highlight social isolation as a clinically relevant prognostic factor and support the incorporation of psychosocial risk assessments into oncology care to inform risk stratification and supportive interventions.