We examined the association between social isolation and loneliness, increasingly recognised but neglected social determinants of health, with being unvaccinated against COVID-19.

This was a cross-sectional study.

A representative cohort of 22 756 individuals (aged 15–81 years) from the general Japanese population who responded to both the Japan COVID-19 and Society Internet Survey 2021 and Japan Society and New Tobacco Internet Survey 2022.

We calculated the ORs of remaining unvaccinated against COVID-19 in 2022, attributable to social isolation as assessed by the Lubben Social Network Scale, or loneliness as evaluated by the University of California, Los Angeles Loneliness Scale version 3. Reasons for abstaining from vaccination were solicited from the unvaccinated respondents. A multivariable logistic regression model was conducted with adjustments for demographic variables. Propensity score-matched comparisons were conducted as part of the sensitivity analysis.

Individuals with social isolation were more likely to be unvaccinated (OR 1.48, 95% CI 1.37 to 1.60), while individuals with loneliness were not (OR 0.96, 95% CI 0.88 to 1.05). Socially isolated individuals were significantly less likely to receive information from people who had been vaccinated (11% vs 15%) and less likely not to trust the vaccine approval process (19% vs 27%) compared with those who were not socially isolated.

Despite not harbouring negative perceptions of the vaccine, socially isolated individuals exhibited lower rates of COVID-19 vaccination. Socially isolated individuals are important targets to reach to increase the number of vaccinated individuals.

The COVID-19 pandemic has highlighted the long-term consequences of SARS-CoV-2 infection, termed long COVID. However, in the absence of comparative groups, the differentiation of disease progression remains difficult, as COVID-19 symptoms become indistinguishable from symptoms originating from alternative etiologies. This study aimed to longitudinally investigate the association between COVID-19 exposure and the somatic symptoms in the Japanese general population.

This was a longitudinal cohort study with 1-year follow-up.

Longitudinal data from 19 545 individuals who participated in the Japan Society and New Tobacco Internet Survey (JASTIS) 2022 and 2023 were included. In this study, we used data from the 2022 JASTIS as baseline data and the 2023 JASTIS as follow-up data. Based on questionnaire responses, respondents were classified into three categories of exposure to COVID-19.

The somatic symptoms were assessed by the Somatic Symptom Scale-8 (SSS-8). Using generalised linear models adjusted for baseline covariates, we calculated the ORs of having very high somatic symptoms assessed by SSS-8, attributable to COVID-19 exposure (no COVID-19 cases as the reference group).

Follow-up completers were divided into three groups according to COVID-19 exposure (no COVID-19, n=16 012; COVID-19 without O2 therapy, n=3201; COVID-19 with O₂ therapy, n=332). After adjusting for all covariates, COVID-19 cases with O₂ therapy had a significant positive association (OR 7.60, 95% CI 5.47 to 10.58) with a very high somatic symptoms burden while other COVID-19 exposure groups did not. Pre-existing physical and psychological conditions were also associated with increased risk of somatic symptoms.

The findings of our study suggest that the severity of COVID-19 symptoms requiring O₂ therapy in the acute phase led to high somatic symptoms. Pre-existing conditions were also associated with a subsequent risk of somatic symptoms.