Table A2.
Results of multivariate proportional hazards regression
| Variable | Variable estimate | SE | Hazards ratio (95% CI) |
|---|---|---|---|
| Serotonin reuptake inhibition increased to next group | 0.10 | 0.04 | 1.10 (1.02 to 1.19) |
| Year entered study | −0.02 | 0.04 | 0.98 (0.91 to 1.05) |
| Age when entered study | 0.05 | 0.00 | 1.05 (1.04 to 1.06) |
| Female patient | −0.51 | 0.07 | 0.60 (0.53 to 0.68) |
| Diabetic patient | 0.23 | 0.08 | 1.26 (1.07 to 1.48) |
| Patient had previous upper gastrointestinal bleeding | 1.36 | 0.10 | 3.89 (3.17 to 4.77) |
| Patient taking non-steroidal anti-inflammatory drugs | 1.05 | 0.08 | 2.85 (2.43 to 3.34) |
| Patient taking acetylsalicylic acid | 0.37 | 0.10 | 1.45 (1.18 to 1.77) |
| Patient taking glucocorticoid | 0.30 | 0.16 | 1.35 (0.99 to 1.84) |
| Patient taking anticoagulant | 0.64 | 0.13 | 1.90 (1.47 to 2.46) |
| Patient taking H2 receptor antagonist or proton pump inhibitor | 0.48 | 0.07 | 1.62 (1.40 to 1.87) |
Factor associated with increased risk of upper gastrointestinal bleeding when hazards ratio >1 and decreased risk when <1. Association significant at P=0.05 level if 95% confidence interval excludes 1.