Local determinants of COPD exacerbations (after PSM) | Final model conditional logistic regression for matched pairs |
OR (95% Wald confidence limits), (point estimate) | |
Patient level | |
Demographic | |
Patients with COPD (sum) (%) total | n=1492 pairs |
Patient’s place of residence in gmina (localisation) | |
1-urban vs 3 | 1.044 (0.673–1.620) |
2-rural vs 3 | 0.897 (0.630–1.277) |
3-urban–rural (Ref. Cat.) | 1 Ref. Cat. |
Healthcare uses | |
Specialisation of physician working in PHC | |
Family doctors vs other | 1.076 (0.920–1.257) |
Other (Ref. Cat.) | 1 Ref. Cat. |
No of consultations in PHC | |
High (>12) vs low (<5) | 1.261 (0.974–1.633) |
Medium<5–12> vs low (<5) | 0.823 (0.643–1.053) |
Low (<5) (Ref. Cat.) | 1 Ref. Cat. |
Gmina/commune, postal codes level | |
Socioeconomic status | |
Total personal income of residents per PHC postal codes (tax office) | |
High (>21 056,52) vs low (<17 079,50) | 0.601 (0.385–0.939) |
Medium <17 079,50–21 056.52> vs low (<17 079,50) | 0.824 (0.581–1.169) |
Low (<17 079,50) (Ref. Cat.) | 1 Ref. Cat. |
Environmental factors | |
Forest cover of the patient’s place of residence in gmina (GUS) | |
High (>20,60%) vs low (<7,00%) | 0.897 (0.605–1.331) |
Medium <7.00%%–20,60%> vs low (<7,00%) | 0.925 (0.648–1.322) |
low (<7,00%) (ref. Cat.) | 1 Ref. Cat. |
The LOGISTIC procedure, model binary logit, 1:1 matched case–control conditional logistic regression for matched pairs. Sex and age variables were included as stratification variables (the 1:1 matched case–control), thus they are not explicitly included in the conditional logistic regression for matched pairs models.
COPD, chronic obstructive pulmonary disease; PHC, primary healthcare clinic; PSM, propensity score matching.