학술논문
The association between major depression prevalence and sex becomes weaker with age
Original Paper
Original Paper
Document Type
Academic Journal
Author
Source
Social Psychiatry and Psychiatric Epidemiology. February 2016, Vol. 51 Issue 2, p203, 8 p.
Subject
Language
English
ISSN
0933-7954
Abstract
Author(s): Scott B. Patten [sup.1] [sup.2] [sup.3] [sup.4], Jeanne V. A. Williams [sup.1], Dina H. Lavorato [sup.1], Jian Li Wang [sup.1] [sup.2] [sup.3], Andrew G. M. Bulloch [sup.1] [sup.2] [sup.3], [...]
Background Women have a higher prevalence of major depressive episodes (MDE) than men, and the annual prevalence of MDE declines with age. Age by sex interactions may occur (a weakening of the sex effect with age), but are easily overlooked since individual studies lack statistical power to detect interactions. The objective of this study was to evaluate age by sex interactions in MDE prevalence. Methods In Canada, a series of 10 national surveys conducted between 1996 and 2013 assessed MDE prevalence in respondents over the age of 14. Treating age as a continuous variable, binomial and linear regression was used to model age by sex interactions in each survey. To increase power, the survey-specific interaction coefficients were then pooled using meta-analytic methods. Results The estimated interaction terms were homogeneous. In the binomial regression model I.sup.2 was 31.2 % and was not statistically significant (Q statistic = 13.1, df = 9, p = 0.159). The pooled estimate (-0.004) was significant (z = 3.13, p = 0.002), indicating that the effect of sex became weaker with increasing age. This resulted in near disappearance of the sex difference in the 75+ age group. This finding was also supported by an examination of age- and sex-specific estimates pooled across the surveys. Conclusions The association of MDE prevalence with sex becomes weaker with age. The interaction may reflect biological effect modification. Investigators should test for, and consider inclusion of age by sex interactions in epidemiological analyses of MDE prevalence.
Background Women have a higher prevalence of major depressive episodes (MDE) than men, and the annual prevalence of MDE declines with age. Age by sex interactions may occur (a weakening of the sex effect with age), but are easily overlooked since individual studies lack statistical power to detect interactions. The objective of this study was to evaluate age by sex interactions in MDE prevalence. Methods In Canada, a series of 10 national surveys conducted between 1996 and 2013 assessed MDE prevalence in respondents over the age of 14. Treating age as a continuous variable, binomial and linear regression was used to model age by sex interactions in each survey. To increase power, the survey-specific interaction coefficients were then pooled using meta-analytic methods. Results The estimated interaction terms were homogeneous. In the binomial regression model I.sup.2 was 31.2 % and was not statistically significant (Q statistic = 13.1, df = 9, p = 0.159). The pooled estimate (-0.004) was significant (z = 3.13, p = 0.002), indicating that the effect of sex became weaker with increasing age. This resulted in near disappearance of the sex difference in the 75+ age group. This finding was also supported by an examination of age- and sex-specific estimates pooled across the surveys. Conclusions The association of MDE prevalence with sex becomes weaker with age. The interaction may reflect biological effect modification. Investigators should test for, and consider inclusion of age by sex interactions in epidemiological analyses of MDE prevalence.