Table 5.
Stepped and full logistic regression models showing associations with CFSa
| Basic model (N = 3,794; R2 = 0.05) |
Basic model and social variables (N = 3,741; R2 = 0.14) |
Basic model and physical inactivity (N = 3,792; R2 |
Basic model and psychological variables (N = 3,794; R2 |
Full model (N = 3,739; R22 |
|||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| OR | 95% CI | P value | OR | 95% CI | P value | OR | 95% CI | P value | OR | 95% CI | P value | OR | 95% CI | P value | |
| Ethnicity | |||||||||||||||
| Irish | 2.12 | 0.8 to 5.71 | 0.13 | 1.96 | 0.70 to 5.44 | 0.2 | 1.76 | 0.64 to 4.80 | 0.27 | 2.25 | 0.83 to 6.09 | 0.108 | 1.72 | 0.58 to 5.1 | 0.32 |
| Black Caribbean | 2.33 | 0.90 to 5.98 | 0.078 | 1.48 | 0.55 to 4.02 | 0.44 | 1.80 | 0.70 to 4.60 | 0.22 | 2.62 | 1.02 to 6.74 | 0.045 | 1.65 | 0.58 to 4.68 | 0.34 |
| Bangladeshi | 3.24 | 1.17 to 9.01 | 0.02 | 1.54 | 0.55 to 4.34 | 0.41 | 1.89 | 0.62 to 5.68 | 0.26 | 4.5 | 1.61 to 12.4 | 0.004 | 1.87 | 0.61 to 5.75 | 0.27 |
| Indian | 2.51 | 0.97 to 6.53 | 0.057 | 2.01 | 0.73 to 5.50 | 0.17 | 1.77 | 0.65 to 4.84 | 0.26 | 2.33 | 0.86 to 6.3 | 0.093 | 1.51 | 0.5 to 4.6 | 0.45 |
| Pakistani | 4.09 | 1.6 to 10.44 | 0.003 | 3.34 | 1.27 to 8.75 | 0.01 | 2.71 | 1.01 to 7.30 | 0.048 | 3.74 | 1.39 to 10.05 | 0.009 | 2.49 | 0.83 to 7.45 | 0.10 |
| Sex | 0.98 | 0.59 to 1.65 | 0.96 | 1.01 | 0.59 to 1.73 | 0.95 | 1.02 | 0.61 to 1.70 | 0.92 | 0.88 | 0.51 to 1.51 | 0.645 | 0.91 | 0.53 to 1.56 | 0.74 |
| Age | 1.02 | 1.01 to 1.03 | 0.002 | 1.03 | 1.02 to 1.05 | <0.001 | 1.01 | 1.01 to 1.03 | 0.015 | 1.03 | 1.01 to 1.04 | 0.001 | 1.02 | 1.01 to 1.04 | 0.001 |
| Education | 0.95 | 0.52 to 1.74 | 0.87 | 0.85 | 0.45 to 1.59 | 0.61 | 0.98 | 0.53 to 1.82 | 0.96 | 0.91 | 0.48 to 1.75 | 0.795 | 0.85 | 0.44 to 1.66 | 0.65 |
| Employment | 1.18 | 0.66 to 2.12 | 0.56 | 1.20 | 0.66 to 2.18 | 0.55 | 1.07 | 0.59 to 1.93 | 0.82 | 0.90 | 0.49 to 1.68 | 0.760 | 0.86 | 0.44 to 1.68 | 0.67 |
| Type of employment | 1.27 | 0.74 to 2.17 | 0.38 | 1.24 | 0.71 to 2.17 | 0.44 | 1.22 | 0.71 to 2.09 | 0.46 | 1.26 | 0.71 to 2.24 | 0.41 | 1.18 | 0.65 to 2.14 | 0.57 |
| Strain (score) | 1.44 | 1.08 to 1.94 | <0.0001 | 1.22 | 1.02 to 1.47 | 0.03 | |||||||||
| Discrimination (score) | 1.45 | 1.08 to 1.95 | 0.01 | 1.28 | 0.92 to 1.78 | 0.15 | |||||||||
| Negative aspects of social support | 2.12 | 1.37 to 3.28 | 0.001 | 2.0 | 1.25 to 3.2 | 0.004 | |||||||||
| Physical in activity (moderate/vigorous vs. inactive/light) | 2.33 | 1.37 to 4.00 | 0.001 | 2.17 | 1.13 to 4.16 | 0.02 | |||||||||
| Depression score | 1.59 | 1.27 to 1.99 | 0.001 | 1.49 | 1.17 to 1.89 | 0.001 | |||||||||
| Anxiety score | 1.91 | 1.54 to 2.39 | 0.001 | 1.80 | 1.43 to 2.27 | <0.001 | |||||||||
aCFS, chronic fatigue syndrome; OR, odds ratio; 95% CI, 95% confidence interval. All variables were identified by likelihood ratio (LR) tests to contribute to the basic model, including age and socioeconomic status (SES) variables (education, none vs. some; employment of head of household, manual vs. nonmanual; employed, no vs. yes). Head of household and employment contributed significantly to the model when considered on their own, whereas education did not. However, when all three SES variables were considered together, they contributed significantly to the model. Sex did not contribute to models using the LR test and changed ORs only slightly. Age of migration and place of birth were tested after the demographic variables were entered, but did not contribute significantly to the model. They were also tested in the final model and made no difference in the results, so they were not retained. All analyses were weighted and adjusted for physical illnesses that might explain fatigue.