Table 2. Odds ratio (OR) and 95% confidence interval of diabetes* according to overtime work hours.
| Overtime work (hours per month) | |||||
| <45 | 45–79 | 80–99 | ≥100 | P for quadratic trend† | |
| 4 companies | |||||
| No. of subjects | 29,308 | 9,648 | 1,369 | 536 | |
| No. of cases (%) | 2418 (8.3) | 583 (6.0) | 64 (4.7) | 38 (7.1) | |
| Model 1‡ | |||||
| Multivariable-adjusted OR | 1.00 | 0.84 | 0.67 | 1.03 | 0.038 |
| (95% confidence interval) | (Ref) | (0.76, 0.93) | (0.52, 0.87) | (0.74, 1.45) | |
| Model 2§ | |||||
| Multivariable-adjusted OR | 1.00 | 0.85 | 0.69 | 1.05 | 0.037 |
| (95% confidence interval) | (Ref) | (0.77, 0.94) | (0.53, 0.89) | (0.75, 1.48) | |
| Model 3|| | |||||
| Multivariable-adjusted OR | 1.00 | 0.86 | 0.69 | 1.03 | 0.07 |
| (95% confidence interval) | (Ref) | (0.77, 0.95) | (0.53, 0.89) | (0.72, 1.46) | |
| 1 company | |||||
| No. of subjects | 23,094 | 9,116 | 1,136 | 461 | |
| No. of cases (%) | 1,927 (8.3) | 554 (6.1) | 51 (4.5) | 36 (7.8) | |
| Model 1¶ | |||||
| Multivariable-adjusted OR | 1.00 | 0.85 | 0.66 | 1.16 | 0.011 |
| (95% confidence interval) | (Ref) | (0.77, 0.94) | (0.49, 0.87) | (0.82, 1. 64) | |
| Model 2** | |||||
| Multivariable-adjusted OR | 1.00 | 0.89 | 0.70 | 1.17 | 0.045 |
| (95% confidence interval) | (Ref) | (0.80, 1.00) | (0.52, 0.94) | (0.81, 1.70) | |
| Model 3†† | |||||
| Multivariable-adjusted OR | 1.00 | 0.87 | 0.66 | 1.11 | 0.05 |
| (95% confidence interval) | (Ref) | (0.78, 0.97) | (0.49, 0.90) | (0.76, 1.61) | |
| Model 4‡‡ | |||||
| Multivariable-adjusted OR | 1.00 | 0.88 | 0.67 | 1.12 | 0.05 |
| (95% confidence interval) | (Ref) | (0.79, 0.98) | (0.49, 0.90) | (0.77, 1.63) | |
Abbreviations: OR, odds ratio; Ref, reference.
*Defined as fasting glucose ≥126 mg/dL (7.0 mmol/l), HbA1c ≥6.5% (48 mmol/mol), or current use of anti-diabetic drug.
P for quadratic trend obtained from multiple logistic regression analysis by assigning 23, 62, 90, and 100 to categories of overtime work.
Model 1 adjusted for age (continuous), sex, and company in 4 companies (n = 40,861).
Model 2 adjusted for factors in model 1 and smoking status (never, past, or current) in 4 companies (n = 40,861).
Model 3 adjusted for factors in model 2 and body mass index (kg/m2, continuous) in 4 companies (n = 40,861).
Model 1 adjusted for age (continuous) and sex in 1 company (n = 33,807).
**Model 2 adjusted for factors in model 1 plus smoking status (never, past, or current), body mass index (kg/m2, continuous), alcohol use (non-drinker, drinker consuming >0 to <23 g, 23 to <46 g, or ≥46 g of ethanol per day), family history of diabetes (yes or no), shift work (yes or no), department (field work or non-field work), and job position (high or low) in 1 company (n = 33,807).
Model 3 adjusted for factors in model 2 and sleep duration (<6 hours, 6 to <7 hours, or ≥7 hours per day) in 1 company (n = 33,807).
Model 4 adjusted for factors in model 3 and leisure time physical activity (<150 min or ≥150 min per week) in 1 company (n = 33,807).