| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
ORIGINAL ARTICLES |
From the Rush Alzheimer’s Disease Center (RSW, KRK) and the Rush Institute for Healthy Aging (LG, JLB, CFMdL, DAE), and the Departments of Neurological Sciences (RSW, DAE), Psychology (RSW, KRK), and Internal Medicine (JLB, CFMdL, DAE), Rush University Medical Center, Chicago, Illinois.
Address correspondence and reprint requests to Robert S. Wilson, PhD, Rush Alzheimer’s Disease Center, Rush University Medical Center, 600 S. Paulina, Suite 1038, Chicago, IL 60612. E-mail: rwilson{at}rush.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Methods: A census was taken of a geographically defined urban community in Chicago, and those aged 65 years or older were invited to participate in an in-home interview; 6158 (79% of those eligible) did so. The interview included brief measures of neuroticism and extraversion, medical history, and questions about current participation in cognitive, social, and physical activities. Vital status was subsequently monitored. The association of each trait with risk of death was examined in a series of accelerated failure-time models that controlled for age, sex, race, and education.
Results: During a mean of more than 6 years of observation, 2430 persons (39.5%) died. A high level of neuroticism (score = 27; 90th percentile) was associated with a 33% increase in risk of death compared with a low level of neuroticism (score = 9; 10th percentile). A high level of extraversion (score = 33; 90th percentile) was associated with a 21% decrease in risk of death compared with a low level (score = 18; 10th percentile). Adjustment for medical conditions and health-related variables did not substantially affect results, but adjusting for baseline levels of cognitive, social, and physical activity reduced the association of both traits with mortality.
Conclusions: The results suggest that higher extraversion and lower neuroticism are associated with reduced risk of mortality in old age and that these associations are mediated in part by personality-related patterns of cognitive, social, and physical activity.
Key Words: neuroticism • extraversion • mortality • cognitive activity • social activity • physical activity
Abbreviations: SD = standard deviation; SE = standard error; RR = relative risk; CI = confidence interval.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Health in old age has long been thought to depend on maintaining an active, socially engaged lifestyle. In particular, higher levels of participation in physical activity (9,10), social activity (11–13), and cognitive activity (14,15) have been associated with reduced risk of disease, disability, and death in old age. The extent to which characteristic patterns of activity in old age are related to personality is uncertain, however.
In the present study, we used data from the Chicago Health and Aging Project to examine the associations among personality, activity patterns, and longevity. Participants are more than 6000 older residents of a geographically defined biracial community. At baseline, we administered brief scales of neuroticism and extraversion and assessed frequency of physical, social, and cognitive activity. Vital status was monitored for a mean of approximately 6 years. In analyses, we first established the relation of each trait to mortality and then tested whether health-related factors or level of physical, social, and cognitive activity could account for the observed personality–mortality associations.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
2 [1] = 43.98, p < .001) but did not differ in gender (2 [1] = 1.29, p = .256). Follow-up interviews were conducted at approximately 3-year intervals. The study was approved by the Institutional Review Board of Rush University Medical Center.
Assessment of Neuroticism and Extraversion
As part of the in-home interview, neuroticism, an indicator of proneness to psychological distress, was assessed with four items (items 1, 21, 36, and 51) from the Neuroticism scale of the NEO Five-Factor Inventory (18); and extraversion, the tendency to be sociable, active, and optimistic, was assessed with four items (items 2, 7, 27, and 32) from the Extraversion scale of the same inventory. Persons rated agreement with each item on a five point scale (0–4), with higher scores indicating more of the trait. For each trait, scores on the four items were summed and multiplied by three to make the total scores (possible range, 0–48) more comparable to the original 12-item scales (18).
We examined the correspondence between the four-item measures of each trait and the standard 12-item score in a separate group of 903 older Catholic clergy members (mean age, 75.3 years; SD, 6.9) participating in the Rush Religious Orders Study (19). The correlation between the brief and standard measures was 0.83 for neuroticism and 0.84 for extraversion (both p < .001), supporting the validity of the brief measures. In addition, the mean level of each trait was similar in the two cohorts: 16.1 (SD, 6.3) for neuroticism and 25.8 (SD, 6.9) for extraversion in the Religious Orders Study compared with 16.8 (SD, 6.7) and 24.9 (SD, 6.4) in the present study. In previous analyses of the present population, the brief measure of neuroticism was associated with more rapid cognitive decline (20) and increased risk of incident dementia (21), consistent with previous research with the standard 12-item scale in the Religious Orders Study (22).
Personality data were missing for 3% of the population. Those with missing personality data had substantially lower Mini-Mental State Examination scores than those with personality data (mean, 5.2; SD, 9.2 versus mean, 25.7; SD, 4.9; t [125] = 24.8, p < .001); they were also older and less educated but did not differ in sex or race.
Assessment of Health-Related Variables
We evaluated multiple indicators of health and health-related behaviors at baseline. We assessed five cardiovascular conditions based on self-report that a physician had previously identified the condition: myocardial infarction, congestive heart failure, stroke, hypertension, and diabetes. We used the number of cardiac conditions present in analyses. Tobacco use was classified as current, former, or never. Alcohol use was expressed as the mean number of drinks per day currently and in the past when consumption was highest. An index of body mass was formed by dividing weight by height squared (kg/m2).
Assessment of Activity Patterns
At baseline, we quantified levels of cognitive, social, and physical activity with previously established measures. Persons were asked to rate frequency of participation in seven cognitively stimulating activities (e.g., reading a magazine) from one (once a year or less) to 5 (everyday or about everyday) with the mean activity score used in analyses, as previously described (14,15,23). Social activity was assessed with four questions about involvement in socially engaging activities (e.g., attending religious services, part-time or full-time employment). The sum of the item scores, which could range from zero to 8, was used in analyses, as described elsewhere (11). Participation in physical activity was assessed with questions adapted (24) from the 1985 Health Interview Survey (25). Persons were asked if they had participated in nine physical activities (e.g., walking for exercise) in the past 2 weeks, and if so, the number of occasions and minutes per occasion. The total number of minutes per week in the nine activities was used in analyses (15).
Ascertainment of Vital Status
Information about the death of study participants was secured through two mechanisms. First, when study personnel tried to schedule a follow-up interview, conducted at approximately 3-year intervals, they might be told of the participant’s death and its date by family, friends, or neighbors. Second, we regularly scan local newspapers for obituaries and search web sites (e.g., www.ancestry.com) using the participant’s name, social security number, or both. In addition, we attempt to verify all deaths by obtaining a death record from the National Death Index that corresponds with a high degree of probability to data on the participant (e.g., social security number, Medicare number). For those who died, the censoring date was the date of death; for survivors, it was the date on which the third population interview was completed.
Data Analysis
To examine the relation of each personality trait to risk of death, we constructed a series of survival models adjusted for age, sex, race, and education. Our initial plan was to use the flexible Cox proportional hazards models. We found, however, that the essential assumption of proportional hazards was not met for key covariates (i.e., age and sex) or for a key predictor, neuroticism, as evidenced by testing time-by-predictor terms in the model as well as by Schoenfeld residual plots. By contrast, the Weibull accelerated failure-time model (26,27) was able to fit the data well for all of our analyses, as confirmed by residual plots, and so we used it in lieu of the Cox models. The Weibull accelerated failure-time model permits estimation of relative risk ratios when the assumption of proportional hazards is not tenable. We first examined each trait separately and then included both traits in the same model. In subsequent analyses of each trait, we added terms for health-related variables and for activity measures. All models were validated graphically and analytically. Programming was done in SAS (28).
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Neuroticism, Extraversion, and Mortality
During a mean of 6.2 years of observation, 2430 (39.5%) people died. Those who died were older, less educated, and more apt to be male and less likely to be black than those who survived (Table 1). Both neuroticism and extraversion were related to mortality in crude analyses. To further investigate these associations, we constructed separate accelerated failure-time models for each trait adjusted for the potentially confounding effects of age, sex, race, and education (Table 2). For each point on the neuroticism scale, risk of death increased by approximately 1.6% (relative risk [RR], 1.016; 95% confidence interval [CI], 1.010–1.022) (model A, Table 2). Thus, as shown in the upper panel of Figure 1, a person with a high neuroticism score (27, 90th percentile) was 33% more likely to die during the observation period than a person with a low score (9, 10th percentile) (model B, Table 2). For each point on the extraversion scale, risk of death decreased by approximately 1.6% (RR, 0.984; 95% CI, 0.978–0.991). Therefore, as shown in the lower panel of Figure 1, a person with a low level of extraversion (18, 10th percentile) was 21% more likely to die during the study compared with a person with a high score (33, 90th percentile). When we repeated the analysis with both neuroticism and extraversion in the model (model C, Table 2), results were comparable, suggesting that the associations of neuroticism and extraversion with mortality were substantially independent.
|
|
|
Health, Personality, and Mortality
Survival was associated with a number of health-related variables at baseline (Table 1). To determine whether personality was related to mortality resulting from an association with current health, we repeated the initial analysis of each trait separately, adding terms for chronic cardiac-related conditions, for current and past use of tobacco and alcohol, and for body mass index and body mass index squared (because both very low and very high body mass are associated with poor health), first separately and then simultaneously. Even in models that controlled for all of these health-related variables, the associations of neuroticism (estimate, –0.008; standard error [SE], 0.004; RR, 1.010; 95% CI, 1.003–.017) and extraversion (estimate, 0.012; SE, 0.004; RR, 0.985; 95% CI, 0.978–0.992) with mortality were not substantially changed, suggesting that current health probably does not mediate that association of personality with subsequent risk of death.
Activity Patterns, Personality, and Mortality
We next considered the possibility that characteristic patterns of cognitive, social, and physical activity might mediate the association of personality with mortality given the simple correlations of each trait with mortality (Table 1) and with cognitive (r = –0.22, p < .001 for neuroticism; r = 0.16, p < .001 for extraversion), social (r = –0.18, p < .001 for neuroticism; r = 0.21, p < .001 for extraversion), and physical (r = –0.13, p < .001 for neuroticism; r = 0.09, p < .001 for extraversion) activity. To test this idea, we repeated the initial analysis of neuroticism with terms added for current levels of cognitive, social, and physical activity, first in separate analyses and then together in a single model. As shown in Table 3, higher level of each kind of activity was associated with a reduced risk of death. In separate models, the association of neuroticism with mortality was reduced by 31% after adjusting for cognitive activity (model A), by 38% after adjusting for social activity (model B), by 8% after adjusting for physical activity (model C), and by 54% after adjusting for all forms of activity together, with p = .032 (model D).
|
We repeated this series of analyses for the measure of extraversion with similar results (Table 4). Thus, the extraversion-mortality association was reduced by 15% after adjusting for cognitive activity (model A), by 46% after adjusting for social activity (model B), and by 8% after adjusting for physical activity (model C). Adjustment for all three forms of activity reduced the association by 54%, with p = .032 (model D).
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
Previous research on the relation of neuroticism and extraversion to mortality has been sparse and inconsistent. Thus, in community-dwelling older adults, higher level of neuroticism has been associated with increased mortality in some studies (1,2), consistent with the present results, but other studies have found no effect (29–31) or the opposite result (32). Extraversion was inversely related to mortality in a cohort of older Catholic clergy members (1), consistent with these results. Other studies have not observed an association between extraversion and mortality, however (30–32). Some of the inconsistency from study to study may reflect differences in how the traits were assessed. Thus, some studies like this one have used short forms (30,32) or been based on archival data (31). In addition, neuroticism and extraversion are multidimensional, and some facets may be more strongly related to mortality than others. Thus, one study found depressive and anxious affect, but not angry affect, to be associated with mortality (33). Similarly, optimism, a facet of extraversion, has been associated with reduced risk of death (34), although optimism in childhood has been associated with increased mortality in adulthood and old age (35).
We found that levels of physical activity, social activity, and cognitive activity each had independent inverse associations with mortality (33). Similar effects have previously been reported for physical (9) and social (12) activity, but not to our knowledge for cognitive activity. We also found that neuroticism was inversely related to frequency of each type of activity and that extraversion was positively correlated with each form of activity. Prior research has not suggested strong associations of these personality traits with cognitive, social, or physical activity in older people (36) or with physical activity in younger people (37,38), although sample sizes were small compared with the present study, suggesting limited statistical power. A novel finding of this study is that controlling for levels of physical, cognitive, and social activity substantially reduced the associations of both neuroticism and extraversion with mortality, with most of the effect resulting from cognitive and social activity. This suggests that much of the association of these traits with mortality is mediated by their association with characteristic patterns of social and cognitive activity.
The principal strength of this study is that participants, who had a mean of approximately 12 years of education, are from a geographically defined population as opposed to a clinic or some other specialized group, making it more likely that a wide spectrum of each personality trait was represented and that the results are generalizable. In addition, the large number of participants made it possible to control for a number of covariates while maintaining adequate statistical power. The main limitation is that the personality traits were assessed with brief measures. Although the brief measures were strongly related to standard measures of each trait, their use may have led us to underestimate their associations with mortality and to overestimate the mediating effects of physical, social, and cognitive activity. In addition, reliance on self-report assessment of health-related variables may have led us to underestimate the effects of these variables in analyses.
The authors thank Ms. Ann Marie Lane for community development and oversight of project coordinating; Ms. Michelle Bos, Ms. Holly Hadden, Mr. Flavio La Morticella, and Ms. Jennifer Tarpey for coordination of the study; Mr. George Dombrowski and Mr. Greg Klein for data management; and Ms. Valerie J. Young for preparing the manuscript.
|
NOTES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
This research was supported by National Institute on Aging grants R01 AG11101, R01 AG09966, and P30 AG10161 and by National Institute of Environmental Health Sciences grant R01 ES10902.
DOI:10.1097/01.psy.0000190615.20656.83
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONNOTESREFERENCES |
|---|
This article has been cited by other articles:
 |
|
 |
|
  B. P. Chapman, K. Fiscella, I. Kawachi, and P. R. Duberstein Personality, Socioeconomic Status, and All-Cause Mortality in the United States Am. J. Epidemiol., January 1, 2010; 171(1): 83 - 92. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. D. Taylor, M. C. Whiteman, G. R. Fowkes, A. J. Lee, M. Allerhand, and I. J. Deary Five Factor Model Personality Traits and All-Cause Mortality in the Edinburgh Artery Study Cohort Psychosom Med, July 1, 2009; 71(6): 631 - 641. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. S. Wilson, N. T. Aggarwal, L. L. Barnes, J. L. Bienias, C. F. Mendes de Leon, and D. A. Evans Biracial Population Study of Mortality in Mild Cognitive Impairment and Alzheimer Disease Arch Neurol, June 1, 2009; 66(6): 767 - 772. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. R. Grossardt, J. H. Bower, Y. E. Geda, R. C. Colligan, and W. A. Rocca Pessimistic, Anxious, and Depressive Personality Traits Predict All-Cause Mortality: The Mayo Clinic Cohort Study of Personality and Aging Psychosom Med, June 1, 2009; 71(5): 491 - 500. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Weiss, C. R. Gale, G. D. Batty, and I. J. Deary Emotionally Stable, Intelligent Men Live Longer: The Vietnam Experience Study Cohort Psychosom Med, May 1, 2009; 71(4): 385 - 394. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 P. J. Rentfrow, S. D. Gosling, and J. Potter A Theory of the Emergence, Persistence, and Expression of Geographic Variation in Psychological Characteristics Perspectives on Psychological Science, September 1, 2008; 3(5): 339 - 369. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Yong Yu, T. Chamorro-Premuzic, and S. Honjo Personality and Defense Mechanisms in Late Adulthood J Aging Health, August 1, 2008; 20(5): 526 - 544. [Abstract] [PDF]
|
|
|
|
|
|
A. Terracciano, C. E. Lockenhoff, A. B. Zonderman, L. Ferrucci, and P. T. Costa Jr Personality Predictors of Longevity: Activity, Emotional Stability, and Conscientiousness Psychosom Med, July 1, 2008; 70(6): 621 - 627. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 H. Nabi, M. Kivimaki, M. Zins, M. Elovainio, S. M Consoli, S. Cordier, P. Ducimetiere, M. Goldberg, and A. Singh-Manoux Does personality predict mortality? Results from the GAZEL French prospective cohort study Int. J. Epidemiol., April 1, 2008; 37(2): 386 - 396. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. H. Ironson, C. O'Cleirigh, A. Weiss, N. Schneiderman, and P. T. Costa Jr Personality and HIV Disease Progression: Role of NEO-PI-R Openness, Extraversion, and Profiles of Engagement Psychosom Med, February 1, 2008; 70(2): 245 - 253. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. W. Roberts, N. R. Kuncel, R. Shiner, A. Caspi, and L. R. Goldberg The Power of Personality: The Comparative Validity of Personality Traits, Socioeconomic Status, and Cognitive Ability for Predicting Important Life Outcomes Perspectives on Psychological Science, December 1, 2007; 2(4): 313 - 345. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. A. Shipley, A. Weiss, G. Der, M. D. Taylor, and I. J. Deary Neuroticism, Extraversion, and Mortality in the UK Health and Lifestyle Survey: A 21-Year Prospective Cohort Study Psychosom Med, November 1, 2007; 69(9): 923 - 931. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. P. Chapman, J. M. Lyness, and P. Duberstein Personality and Medical Illness Burden Among Older Adults in Primary Care Psychosom Med, April 1, 2007; 69(3): 277 - 282. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. S. Wilson, S. E. Arnold, J. A. Schneider, Y. Li, and D. A. Bennett Chronic Distress, Age-Related Neuropathology, and Late-Life Dementia Psychosom Med, January 1, 2007; 69(1): 47 - 53. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R E Rhodes and N E I Smith Personality correlates of physical activity: a review and meta-analysis. Br. J. Sports Med., December 1, 2006; 40(12): 958 - 965. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
C. R. Cloninger How Does Personality Influence Mortality in the Elderly? Psychosom Med, November 1, 2005; 67(6): 839 - 840. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
