Table shows the distribution of characteristics of men and women in the WHS. There were statistically significant differences between men and women for all sociodemographic and health variables. The mean ages for men and women in the WHS were 38 and 40
years old respectively. Men were more likely to never have married than women while women were more likely to be widowed (P
0.01). Women were more likely to not have any formal education and to report all chronic conditions (P
In Table , we examined the characteristics of those with and without mobility difficulty. Women were more likely than men to report mobility difficulty (38% versus 27%, P
0.01). Those who were older, had less formal education, lived in rural areas, and had chronic health conditions were more likely to report mobility difficulty (P
0.01). Marital status was also associated with mobility difficulty such that those who never married had the least mobility difficulty (P
Characteristics of the study population with and without mobility difficulty
In Table , the prevalence rate ratios for female gender are given adjusting for different sets of covariates while holding the number of observations constant. The age-adjusted prevalence rate ratio for gender in Model 1 was 1.35 (95% CI 1.31-1.38). In other words, women had a 35% higher prevalence of mobility difficulty than men. The addition of education, marital status, and urban versus rural setting in Model 2 slightly reduced the prevalence rate ratio for female gender to 1.30 (95% CI 1.26-1.33). Adding the 5 health variables (back pain, arthritis, angina, depressive symptoms, and cognitive difficulties) in Model 3 further reduced the prevalence rate ratio for female gender to 1.12 (95% CI 1.09-1.15).
Poisson regression models showing the changes in the gender gap in mobility difficulty after adjustment
We then examined whether the gender gap in mobility difficulty differed across world regions by examining region-specific regression models (Table ). After age adjustment, the gender gap was the largest in the Eastern Mediterranean (prevalence rate ratio (PRR)
1.66, 95% CI 1.51-1.81) and was the smallest in the Western Pacific (PRR
1.12, 95% CI 1.06-1.18). In the fully adjusted model (Model 3), that pattern remained with the gender gap in the Western Pacific completely disappearing (PRR
1.03, 95% CI 0.98-1.08). Adding interaction terms for gender and world region to an age-adjusted model for all regions combined resulted in statistically significant interactions for the Western Pacific and the Eastern Mediterranean regions compared to Africa (data not shown in Table 5, P
0.001 and P
0.001 respectively). In other words, the prevalence rate ratio for female gender in the Eastern Mediterranean was significantly larger than in Africa while the prevalence rate ratio in the Western Pacific was significantly smaller than in Africa. Consistent with this result, of the 6 world regions in the WHS, countries in the Western Pacific region had the smallest difference between the HDI and the GDI while countries in the Eastern Mediterranean region had the largest.
Poisson regression models showing the changes in the gender gap in mobility difficulty by world region
The countries listed as part of the Eastern Mediterranean region by the WHS included Morocco, Pakistan, Tunisia, and the United Arab Emirates (UAE). These countries differ substantially in economic development, religious freedom, and geography. They share in common large Muslim populations. Despite the diversity of these 4 countries, there was not substantial heterogeneity in the age-adjusted prevalence rate ratios for female gender in each of these 4 countries as the prevalence rate ratios ranged from 1.54 (95% CI 1.12, 2.13) in the United Arab Emirates to 1.96 (95% CI 1.72, 2.23) in Tunisia. Furthermore, the differences between the human development index and the gender development index for 4 of the 5 countries were in the largest quartile of all the WHS countries (The UAE did not have data on the gender development index). At the other end of the gender equality spectrum, the countries listed as part of the Western Pacific by the WHS included Australia, China, Malaysia, Philippines, Lao People’s Democratic Republic, and Vietnam. Again, despite the diversity of these countries, the age-adjusted prevalence rate ratios for female gender were fairly similar ranging from 1.04 (95% CI 0.99, 1.10) in the Philippines to 1.39 (95% CI 1.15, 1.68) in China. The differences between the human development index and the gender development index for 4 of the 6 countries were in the smallest quartile of all the WHS counties while Malaysia and Lao were in the smallest 30th percentile.