TABLE 4

OR (95% CI) for unemployment for chronic airflow obstruction (CAO) and other risk factors, stratified by country income category, with increasing degree of adjustment (demographic characteristics, comorbidities and forced vital capacity (FVC))

Model 2Model 3Model 4
High incomeLow to middle incomeHigh incomeLow to middle incomeHigh incomeLow to middle income
Spirometric CAO1.71 (1.17–2.49)1.16 (0.78–1.73)1.63 (1.16–2.28)1.18 (0.79–1.76)1.68 (1.16–2.45)1.15 (0.77–1.71)
FVC 10 percentage points decrease in % pred1.09 (1.03–1.15)1.08 (1.02–1.14)
Female1.36 (1.16–1.59)3.34 (2.76–4.04)1.43 (1.23–1.68)3.25 (2.68–3.94)1.44 (1.23–1.68)3.23 (2.66–3.91)
Age 10-year increment4.28 (3.77–4.86)2.31 (2.06–2.59)4.04 (3.55–4.59)2.21 (1.96–2.48)4.02 (3.53–4.57)2.20 (1.96–2.47)
Smoking status
 Current smoker1.34 (1.09–1.65)0.88 (0.72–1.09)1.36 (1.11–1.68)0.89 (0.72–1.10)1.36 (1.10–1.67)0.89 (0.72–1.10)
 Ex-smoker1.31 (1.09–1.56)1.15 (0.90–1.47)1.28 (1.07–1.53)1.13 (0.88–1.44)1.29 (1.08–1.54)1.12 (0.88–1.43)
Education
 Some college1.85 (1.45–2.38)0.96 (0.60–1.52)1.83 (1.43–2.36)0.95 (0.60–1.52)1.83 (1.42–2.35)0.97 (0.61–1.54)
 High school2.30 (1.84–2.88)1.26 (0.92–1.73)2.27 (1.81–2.84)1.28 (0.93–1.76)2.24 (1.79–2.81)1.30 (0.94–1.78)
 Middle school3.65 (2.73–4.87)1.23 (0.89–1.70)3.54 (2.64–4.74)1.26 (0.91–1.74)3.49 (2.60–4.67)1.26 (0.91–1.75)
 Primary school4.14 (3.02–5.66)1.52 (1.10–2.10)3.90 (2.84–5.37)1.56 (1.13–2.16)3.86 (2.80–5.30)1.56 (1.13–2.16)
 No education1.96 (0.61–6.32)1.61 (1.13–2.30)2.01 (0.62–6.43)1.64 (1.15–2.34)1.98 (0.61–6.43)1.65 (1.15–2.35)
Comorbidities
 Hypertension1.25 (1.05–1.49)1.25 (1.02–1.53)1.22 (1.02–1.45)1.22 (0.99–1.50)
 Heart disease1.56 (1.22–2.00)1.20 (0.86–1.67)1.53 (1.19–1.96)1.18 (0.85–1.64)
 Diabetes1.53 (1.14–2.04)1.39 (1.01–1.92)1.46 (1.09–1.95)1.38 (1.00–1.91)
 Stroke2.17 (1.14–4.13)1.58 (0.83–3.02)2.15 (1.13–4.10)1.56 (0.82–2.99)
 Lung cancer2.87 (0.89–9.30)1.01 (0.04–28.54)2.86 (0.89–9.26)1.03 (0.04–28.37)

Adjustment variables: no fixed effects (model 1); age, sex, education and smoking (model 2); model 2 adjustment + comorbidities (model 3); model 3 adjustment + FVC (model 4). All five models were fit using a multilevel mixed-effects generalised linear model with study site included as random effect to account for within-site clustering. Separate analyses were performed for high-income countries and low- to middle-income countries. Reference values for categorical variables: no CAO, male, never-smoker, university education, no hypertension, no heart disease, no diabetes, no stroke and no lung cancer. n=11 675. Retirees (age limit defined as ≥65 years) and homemakers/caregivers were excluded from the analysis.