Personality in women and associations with mortality: a 40-year follow-up in the Population Study of Women in Gothenburg

Background: The question of whether personality traits influence health has long been a focus for research and discussion. Therefore, this study was undertaken to examine possible associations between personality traits and mortality in women. Methods: A population-based sample of women aged 38, 46, 50 and 54 years at initial examination in 1968-69 was followed over the course of 40 years. At baseline, 589 women completed the Cesarec-Marke Personality Schedule (the Swedish version of the Edwards Personal Preference Schedule) and the Eysenck Personality Inventory. Associations between personality traits and mortality were tested using Cox proportional hazards models. Results: No linear associations between personality traits or factor indices and mortality were found. When comparing the lowest (Q1) and highest quartile (Q4) against the two middle quartiles (Q2 + Q3), the personality trait Succorance Q1 versus Q2 + Q3 showed hazard ratio (HR) = 1.37 (confidence interval (CI) = 1.08-1.74), and for the factor index Aggressive non-conformance, both the lowest and highest quartiles had a significantly higher risk of death compared to Q2 + Q3: for Q1 HR = 1.32 (CI = 1.03-1.68) and for Q4 HR = 1.36 (CI = 1.06-1.77). Neither Neuroticism nor Extraversion predicted total mortality. Conclusions: Personality traits did not influence long term mortality in this population sample of women followed for 40 years from mid- to late life. One explanation may be that personality in women becomes more circumscribed due to the social constraints generated by the role of women in society.