How do the factors associated with frailty change with sex? An exploratory network analysis
Abstract Background It is important to study multiple social, physical and psychosocial factors associated with frailty in populations characterized by social and health disparities, such as men and women. Methods This was a cross-sectional population-based study with older adults ≥65 years from the FIBRA (Frailty in Brazilian Older Adults) 2008-2009 study. We carried out a comparative analysis of the factors associated with the frailty phenotype in older men (N=706) and women (N=1.251) using networks based on mixed graphical models (MGM) according to sex. Results In the male network, frailty was most strongly associated with years of schooling, overall satisfaction with life and falls; in the female network, the syndrome was associated with satisfaction with problem solving, depression and diabetes in addition to years of schooling. Furthermore, permutation tests showed that the networks for males and females were statistically different in terms of their structure, the global strength of the relationships and the strength of the relationships between frailty and diabetes; frailty and falls; frailty and depression; frailty and overall satisfaction with life; and frailty and satisfaction with problem solving (p<0.05). The walktrap network cluster detection algorithm revealed that in men, frailty was in a physical and social dimension while in women the syndrome was in a cardiometabolic and psychosocial dimension. Conclusions Network analysis showed that different factors are associated with frailty for each sex. The findings suggest that different strategies for dealing with frailty should be adopted for men and women so that care and prevention efforts can be directed appropriately.