Association between sleep quality and central obesity among southern Chinese reproductive-aged women

BMC Women's Health

Table 4 Cluster logistic regression models explaining central obesity by variables in three clusters

Predictor variable^a	OR^b (95% CI)	P	Nagelkerke R^2c	Independent contribution (%)
Cluster 1
Age group (years) (15–25)	Reference
26–35	2.04 (1.40–2.98)	< 0.001
36–45	2.47 (1.65–3.68)	< 0.001
46–49	2.58 (1.59–4.19)	< 0.001
Marital status (unmarried)	Reference
Married	1.90 (1.31–2.74)	0.001
Educational level (primary school or lower)	Reference
High school or above	0.48 (0.30–0.77)	0.002
Occupational status (employed)	Reference
Unemployed	1.40 (1.11–1.77)	0.004
Total			0.074	73.3%
Cluster 2
Hypertension, yes versus no	4.17 (1.78–9.75)	0.001
Diabetes mellitus, yes versus no	5.32 (1.07–26.45)	0.041
Hospitalization in the last year, yes versus no	1.88 (1.33–2.65)	< 0.001
Total			0.091	16.8%
Cluster 3
Global PSQI score, > 7 versus ≤ 7	2.20 (1.28–3.78)	0.004
Subjective sleep quality	0.81 (0.73–0.90)	< 0.001
Sleep disturbance	1.11 (1.01–1.22)	0.042
Total			0.101	9.9%

The forward stepwise method was used in the logistic regression analysis
^aOnly variables with P ≤ 0.05 were included in the model
^bFor age group, marital status, educational level, occupational status, hypertension, diabetes mellitus, hospitalization in the last year, and sleep quality domain scores, odd ratios per standard deviation increase were presented
^cNagelkerke R² in this study is the variance of the dependent variable (central obesity), which could be explained by variables in three clusters included in the regression model

ISSN: 1472-6874