Table 3

Incidence and adjusted sub-HRs of community engagement controlling for competing risk of death

n=9550; 76 876 person-years Dementia cases: 893Adjusted sub-HRs (95% CI)P for trend
Model 1
Community cultural engagement
 Less than once a year1 (Ref)<0.001
 Once or two times per year0.30 (0.24 to0.38)
 Every few months or more0.23 (0.18 to0.30)
Community group engagement
 Less than once a year1 (Ref)<0.001
 Once or two times per year0.54 (0.41 to0.70)
 Every few months or more0.61 (0.49 to0.76)
Model 2
Community cultural engagement
 Less than once a year1 (Ref)<0.001
 Once or two times per year0.37 (0.29 to0.47)
 Every few months or more0.33 (0.25 to0.43)
Community group engagement
 Less than once a year1 (Ref)0.003
 Once or two times per year0.64 (0.50 to0.84)
 Every few months or more0.79 (0.63 to1.00)
Model 3
Community cultural engagement
 Less than once a year1 (Ref)<0.001
 Once or two times per year0.41 (0.32 to0.52)
 Every few months or more0.38 (0.29 to0.50)
Community group engagement
 Less than once a year1 (Ref)0.012
 Once or two times per year0.69 (0.53 to0.89)
 Every few months or more0.82 (0.65 to 1.04)
Model 4
Community cultural engagement
 Less than once a year1 (Ref)<0.001
 Once or two times per year0.46 (0.36 to0.59)
 Every few months or more0.44 (0.33 to0.57)
Community group engagement
 Less than once a year1 (Ref)0.30
 Once or two times per year0.81 (0.62 to 1.07)
 Every few months or more0.94 (0.74 to 1.20)
  • Community cultural engagement and group engagement were entered simultaneously into the models so are mutually adjusted. Model 1 unadjusted. Model 2 adjusted for demographic covariates (age, sex, educational attainment, wealth, employment status). Model 3 additionally adjusted for health covariates (depression, eyesight, hearing, cardiovascular conditions, physical activity). Model 4 additionally adjusted for social covariates (marital status, living status, social contact, social network size, perceived loneliness, perceived positive social support, perceived negative social support).