Adjusted ORs and regression coefficients (β) for variables included in a more comprehensive risk score
| OR* | (95% CI) | β | (95% CI) | |
|---|---|---|---|---|
| Smoking status | ||||
| Never | 1 | 0 | ||
| Former | 4.36 | (4.00 to 4.75) | 1.47 | (1.39 to 1.56) |
| Current | 12.0 | (10.97 to 13.12) | 2.48 | (2.40 to 2.57) |
| Missing | 2.87 | (2.52 to 3.26) | 1.05 | (0.92 to 1.18) |
| Asthma | 1.89 | (1.71 to 2.08) | 0.64 | (0.54 to 0.73) |
| LRTI† | ||||
| 0 | 1 | 0 | ||
| 1 | 1.81 | (1.64 to 1.99) | 0.59 | (0.49 to 0.69) |
| >1 | 2.23 | (1.96 to 2.54) | 0.80 | (0.67 to 0.93) |
| Presentations with cough† | ||||
| 0 | 1 | 0 | ||
| 1 | 1.42 | (1.30 to 1.56) | 0.35 | (0.26 to 0.44) |
| >1 | 1.77 | (1.59 to 1.97) | 0.57 | (0.46 to 0.68) |
| Presentations with dyspnoea† | ||||
| 0 | 1 | 0 | ||
| 1 | 3.17 | (2.82 to 3.57) | 1.16 | (1.04 to 1.27) |
| >1 | 4.53 | (3.89 to 5.28) | 1.51 | (1.36 to 1.66) |
| Wheeze† | 1.86 | (1.60 to 2.17) | 0.62 | (0.47 to 0.77) |
| Sputum production† | 1.49 | (1.17 to 1.90) | 0.40 | (0.16 to 0.64) |
| Unintended weight loss† | 1.75 | (1.33 to 2.31) | 0.56 | (0.29 to 0.84) |
| Antibiotic courses for a LRTI† | ||||
| 0 | 1 | 0 | ||
| 1 | 1.33 | (1.23 to 1.44) | 0.29 | (0.21 to 0.37) |
| 2 | 1.53 | (1.38 to 1.70) | 0.43 | (0.32 to 0.53) |
| >2 | 1.80 | (1.62 to 2.01) | 0.59 | (0.48 to 0.70) |
| Salbutamol† | 4.19 | (3.81 to 4.61) | 1.43 | (1.34 to 1.53) |
| Prednisolone† | 1.53 | (1.38 to 1.69) | 0.42 | (0.32 to 0.52) |
As this model was developed using case-control data, the intercept term is not applicable and has therefore not been presented. The c statistic for this model in the external validation sample was 0.87 (95% CI 0.86 to 0.87).
*Estimated using a multivariable conditional logistic regression model.
†Within 3 years of COPD diagnosis or equivalent matched time point for controls.
Risk score=(former smoker×1.47)+(current smoker×2.48)+(unknown smoking status×1.05)+(asthma×0.64)+(1 episode of LRTI×0.59)+ (>1 episode of LRTI×0.80)+(1 episode of cough×0.35)+(>1 episode of cough×0.57)+(1 episode of dyspnoea×1.16)+(>1 episode of dyspnoea×1.15)+(wheeze×0.62)+(sputum×0.40)+(unintended weight loss×0.56)+(1 antibiotic course×0.29)+(2 antibiotic course×0.43)+ (>2 antibiotic courses×0.59)+(salbutamol×1.43)+(prednisolone×0.42).
NB. Each variable can either take the value 0 (not present) or 1 (present).For example, A former smoker with a history of asthma who presented with more than one lower respiratory tract infection and episode of cough in the past 3 years, reported of unintended weight loss and received salbutamol and 2 course of antibiotics for a LRTI in the past 3 years would have the following risk score:
(1×1.47)+(0×2.48)+(0×1.05)+(1×0.64)+(0×0.59)+(1×0.80)+(0×0.35)+(1×0.57)+(0×1.16)+(0×1.15)+(0×0.62)+(0×0.40)+(1×0.56)+(0×0.29)+(1×0.43)+(0×0.59)+(1×1.43)+(0×0.42)=5.9.
COPD, chronic obstructive pulmonary disease; LRTI, lower respiratory tract infection.