Harm Avoidance and Disability in Old Age

Abstract

The relation of personality to disability in old age is not well understood. The authors examined the relation of harm avoidance, a trait indicating a tendency to worry, fear uncertainty, be shy, and tire easily, to disability in a group of 474 older persons without dementia. Participants completed the 35-item Harm Avoidance scale. Disability was assessed with the Rosow-Breslau scale, a self-report measure of physical mobility. Performance-based tests of lower limb functions were also administered from which composite measures of gait, balance, and strength were derived. In a logistic regression model controlled for age, sex, education, and lower limb function, persons with high levels of harm avoidance were nearly three times as likely to report mobility limitations as persons with low levels, and these effects largely reflected fatigability and fear of uncertainty. The association of harm avoidance with disability was not explained or modified by frailty, physical activity, depressive symptoms, neuroticism, extraversion, or cognition. The results suggest that harm avoidance is associated with disability in old age.

This research was supported by National Institute on Aging grants R01 AG17917, R01 AG024480, and R01 AG022018, and the Illinois Department of Public Health. The authors thank the many Illinois residents for participating in the Rush Memory and Aging Project; Traci Colvin, MPH, and Tracy Hagman for coordinating the study; Woojeong Bong, MS, for statistical programming; George Dombrowski, MS, and Greg Klein for data management; and Sherry A. Carroll for preparing the manuscript.

Notes

∗ p < .05 for correlations with an absolute value of .10 or greater. Educ., education; Phys. Act., physical activity; CES-D, Center for Epidemiologic Studies Depression Scale; neur., neuroticism; ext., extraversion; cog., global cognition.

Note. Data are presented as mean (SD) or %. CES-D, Center for Epidemiologic Studies Depression Scale.

∗ Estimated from logistic regression models that controlled for age, sex, and education. Results show the effect of 1-unit of a given model term on likelihood of disability.

∗ Estimated from separate logistic regression models that controlled for age, sex, education, gait, balance, and strength. Results show the effect of 1-unit of a given harm avoidance component on likelihood of disability.

∗ Estimated from separate logistic regression models that controlled for age, sex, education, gait, balance, and strength; in addition, Model A controlled for frailty, Model B for physical activity, Model C for depressive symptoms, Model D for neuroticism, and Model E for extraversion. Results show the effect of 1-unit of a given trait on the likelihood of disability.