PT  - JOURNAL ARTICLE
AU  - C Kim
AU  - J Dunn
TI  - OP29 Interaction between socioeconomic position and social integration in suicide mortality: a nationally representative cohort study
AID  - 10.1136/jech-2019-SSMabstracts.29
DP  - 2019 Sep 01
TA  - Journal of Epidemiology and Community Health
PG  - A14--A14
VI  - 73
IP  - Suppl 1
4099  - http://jech.bmj.com/content/73/Suppl_1/A14.1.short
4100  - http://jech.bmj.com/content/73/Suppl_1/A14.1.full
SO  - J Epidemiol Community Health2019 Sep 01; 73
AB  - Background A low level of social integration and lower socioeconomic position (SEP) are well-known risk factors for suicidal behavior. Most literature has suggested that the effects of family types as a proxy of social integration and SEP measures on suicide are merely additive. However, as social integration is not dependent on SEP, the association of these effects could be interactive. Since the protective effects of social integration vary by gender, so could this interaction. The aim of this study was to examine the interaction between social integration, SEP and gender on suicide mortality in the Canadian context.Methods Using data from the 1991 Canadian Census Health and Environment cohort (CanCHEC) —which included 2.5 million Canadians over a 20-year follow up period - we applied Cox proportional hazards regression models to examine the association among living arrangements (lives alone versus single parent family versus others), education (secondary versus non-secondary), income (low income versus non-low income), and employment status (unemployed versus else) by gender. Models were developed to observe how living arrangements attenuated the association between SEP measures and suicide. In the full model, we added interaction terms between living arrangements and employment status.Results In model 1, which was only adjusted for age and three SEP measures, both men and women with low income (Hazard Ratio (HR): 1.846 [women], HR: 1.278 [Men]) and who were unemployed (HR: 1.501 [Women], HR: 1.677 [Men]) were more likely to be exposed to completed suicide. In all models, lower education was associated with suicide risk among men, but not among women. In model 2, when living arrangements were added, the association between SEP measures and suicide was much attenuated among women, but not among men. In the full model, an interactive effect between unemployment and living arrangements (living alone) was not shown among men. However, there was a significant interactive effect for women, demonstrating that unemployed women who do not live alone were 1.429 times more likely to complete suicide than employed women living with others, but women living alone and unemployed were 2.125 times more likely to do so.Conclusion While SEP had more independent impacts from social integration on suicide among men, there were significant synergetic effects on suicide mortality among women in Canada.