Regression coefficients of the random effect model used to calculate annual decline in lung function
| Dependent variable: FEV1 % predicted | β-coefficient | p-value |
| Intercept | −20.95 | 0.26 |
| Year | 0.24 | 0.33 |
| Smoking cessation (0 = no, 1 = yes) | 14.32 | <0.0001 |
| Sex (0 = male, 1 = female) | 7.32 | 0.10 |
| Age | 1.11 | 0.13 |
| Baseline FEV1 % pred | 1.36 | <0.0001 |
| Yr×smoking cessation | 0.45 | <0.0001 |
| Yr×sex | −0.16 | <0.0001 |
| Yr×age | −0.031 | <0.0001 |
| Yr×baseline FEV1 % pred | 0.006 | <0.01 |
| Smoking cessation×sex | 1.73 | <0.0001 |
| Smoking cessation×baseline FEV1 % pred | −0.12 | <0.0001 |
| Sex×age | −0.40 | <0.05 |
| Sex×baseline FEV1 % pred | 0.027 | <0.05 |
| Age×baseline FEV1 % pred | −0.018 | <0.05 |
| Age2 | −0.012 | 0.10 |
| Age2×smoking cessation | −8.6×10−4 | <0.0001 |
| Age2×sex | 4.29×10−3 | <0.05 |
| Age2×baseline FEV1 % pred | 1.97×10−4 | <0.05 |
FEV1: forced expiratory volume in one second; % pred: % predicted; Age2: age×age. Annual decline was calculated by subtracting the calculated FEV1 % pred in year 0 (given certain sex, age, smoking status and baseline FEV1 % pred) of the FEV1 % pred in year 1 (given certain sex, age+1, smoking status and baseline FEV1 % pred)