Extended Data Table 4 Sensitivity analysis of the longitudinal association between social capital and incidence of severe mental illnesses in final multivariable models, after regenerating social capital scores excluding 280 SPHC respondents in 2002 later diagnosed with SMI

From: Longitudinal association between neighborhood-level social capital and incidence of major psychiatric disorders in a cohort of 1.4 million people in Sweden

  1. MV: Multivariable; HR: Hazard ratio; CI: confidence interval: LR: likelihood ratio; df: degrees of freedom
  2. aFrom Table 3 (main effects of trust variables on risk) or Table 4 (region-specific personal trust estimates)
  3. bExcluding social capital scores from N = 280 of 23 510 (1·2%) respondents who took part in the SPHC survey in 2002, but who later became diagnosed with SMI
  4. cWald p-value for HR associated with parameter of interest
  5. dLikelihood ratio test for statistical interaction between each trust measure and region-of-origin. Stratified estimates provided where LRT p < 0·05. For equivalent p-values for the final MV model, see Table 4
  6. eAdjusted for age group, sex, their interaction, parental history of SMI, parental region-of-origin (except where stratified results are presented), family disposable income quintile at cohort entry, and time-varying deprivation quintile, own-group migrant density, and other social capital domains. For non-affective psychotic disorders, we additionally adjusted for time-varying population density quintile. See Supplementary Figs. 3 and 4
Back to article page