Sleepless in Inequality: Results from the 2018 Behavioral Risk Factor Surveillance System, a cross-sectional study | BMC Public Health

0

The 2018 Behavioral Risk Factor Surveillance System dataset included 425,712 respondents from 50 states and the District of Columbia. All respondents with missing data on sleep behavior and other covariates were excluded, resulting in a complete dataset of 350,929 (82.4%). Withdrawn participants were more likely to be non-Hispanic black, male, young, and rural.

Table 1 presents the characteristics and corresponding weighted percentage of respondents with complete data. Of the respondents, 50.0% were women. The majority of participants were white (64.6%), followed by Hispanics (15.8%) and blacks (11.6%). Among the sample, 36.1%, 28.2% and 35.7% belonged to high, middle and low income families respectively. Most respondents lived in urban areas (93.5%).

Table 1 Characteristics of U.S. adults participating in the 2018 Behavioral Risk Factor Surveillance System (BRFSS) (n=350,929) and U.S. states (50 states and the District of Columbia)

The characteristics of the 50 states and Districts of Columbia are also depicted in Table 1. The Gini index had a mean of 0.468, a standard deviation of 0.02, a median of 0.468, and ranged from 0.427 to 0.524. Median state income was $58,143 (SD=9,820), mean proportion black was 10.9% (SD=10.7), mean proportion poor was 22.5% ( SD = 13.1) and the mean population was 6,332,183 (SD = 7,235,904).

The intercept-only model indicated that the overall predicted probability was 36.2% and 4.6% for insufficient and very insufficient sleep, respectively. Additionally, the intercept-only model confirmed significant variability in the percentage of the population getting less than 7 h and less than 5 h of sleep consistently each day. For example, the overall predictive probability was 30.7-42.2% and 3.2-5.5% for insufficient and very insufficient sleep in US states.

The crude bivariate and adjusted associations are shown in Table 2. In the adjusted analyses, compared to men, women were less likely to get insufficient sleep (OR = 0.94, 95% CI: 0.92, 0.96) and very insufficient (OR = 0.91, 95% CI: 0.84 to 0.98) (Table 2). Additionally, those from low-income households were more likely to have insufficient (OR=1.14, 95% CI: 1.07, 1.21) and very insufficient (OR=2.10, CI at 95%: 1.88, 2.34), compared to those from high-income families. household income.

Table 2 Cross-sectional associations between income inequality and the likelihood of getting insufficient (

Table 2 shows the association between the Gini coefficient and the likelihood of getting insufficient and very insufficient sleep. Crude analyzes indicated that an increase in the Gini standard deviation was associated with both an increased likelihood of getting insufficient sleep (OR = 1.06, 95% CI = 1.01, 1.11 ) and very poor sleep (OR = 1.08, 95% CI = 0.99, 1.17). Associations between income inequality and the likelihood of getting insufficient (OR = 1.06, 95% CI = 1.00, 1.13) and very insufficient (OR = 1.11, 95% = 1.03, 1.20) after adjusting for individual and regional covariates. When testing whether the associations varied by sex (male vs. female), there was no heterogeneity when insufficient sleep was the outcome (1.01, 95% CI = 0.99; 1 ,03). However, the interaction term between the levels indicated that a one standard deviation increase in the Gini coefficient was associated with an increased likelihood of sleeping less than 5 h in women (OR = 1.07, 95% CI: 1.01, 1.13). In other words, the estimated proportion of women getting less than 5 h of sleep is higher than the estimated proportion of men, especially at higher levels of income inequality (Fig. 1).

Fig. 1

Association between the Gini index and obtaining very insufficient sleep (

Addition of mediators (Table 3) resulted in a slight attenuation of the state-level Gini coefficient estimate for insufficient sleep (OR = 1.06, 95% CI = 0.99, 1, 12) and very poor sleep (OR = 1.09, 95% CI = 1.01, 1.17). Table 4 presents the results of the Baron-Kenny mediation analyzes examining the bivariate associations. A standard deviation increase in the state-level Gini coefficient was associated with an increased likelihood of getting insufficient sleep (OR = 1.06, 95% CI = 1.01, 1.11) and very poor sleep (OR = 1.08, 95% CI = 0.99, 1.17) . Additionally, a one-standard deviation increase in the state-level Gini coefficient is associated with an increased odds of experiencing 14 or more days of poor mental health in the previous month (OR = 1.03, CI at 95% = 1.02, 1.04). Finally, compared to the experience of 0 days when mental health was not good, those who had experienced 1 to 13 days and more of and equal to 14 days when mental health was not good, had more chances of getting insufficient and very insufficient sleep. . Figure 2 illustrates the mediating associations observed. Although direct and indirect pathways are described, these are proposed mechanisms and are not necessarily causal.

Table 3 Multilevel regression analyzes with adjustment for mediator: number of days with poor mental health among participants in the 2018 Behavioral Risk Factor Surveillance System (BRFSS)
Figure 2
Figure 2

Associations between income inequality, days of poor mental health (mediator), and

Share.

Comments are closed.