Table 4. Multivariate Bayesian analysis showing posterior distributions for the regression coefficients associated with recruitment hospital performance.

	Mean posterior probability	25% Limit of credibility	75% Limit of credibility	OR
Response variable: ratio of interim recruited to expected COPD cases
Intercept	−2.50	−3.18	−1.79
Total patients admitted in 2007 (log N)	0.29	0.19	0.38	1.33
Access to electronic/digital information (Yes)	0.23	0.13	0.34	1.26
COPD patients admitted in 2007 (log N)	0.20	0.12	0.28	1.23
Access to pulmonologist in the ED (Yes)	0.13	0.03	0.25	1.15
Written protocol for COPD (Yes)	0.09	−0.01	0.19	1.09
Early discharge scheme/day hospital/hospital at home (Yes)	−0.31	−0.42	−0.21	0.73
Total hospital beds (log N)	−0.17	−0.28	−0.03	0.85
Hospital case-mix index (units)	−0.14	−0.26	0.00	0.87
University hospital (Yes)	−0.13	−0.25	−0.01	0.88
Pulmonary medicine staff (N)	−0.04	−0.06	−0.02	0.96
Hospital random effects (standard deviation)	0.70	−0.67	0.73
Response variable: proportion of interim COPD cases that become definite
Intercept	1.71	0.97	2.45
Hospital Case-mix (units)	0.24	−0.00	0.49	1.27
Lung function laboratory (Yes)	0.40	0.15	0.65	1.49
Access to electronic/digital information (Yes)	0.34	0.15	0.54	1.40
Pulmonary physicians on duty on site (Yes)	0.18	−0.06	0.41	1.20
Pulmonary medicine staff (N)	0.03	0.00	0.05	1.03
University Hospital (Yes)	−0.72	−0.92	−0.52	0.49
Residents in training (Yes)	−0.34	−0.54	−0.14	0.71
Written protocol for COPD (Yes)	−0.31	−0.48	−0.12	0.74
COPD patients admitted for COPD in 2007 (log N)	−0.26	−0.38	−0.12	0.77
Non-invasive ventilation (Yes)	−0.26	−0.53	−0.00	0.77
Hospital random effects (standard deviation)	1.29	1.23	1.36

OR: Odds Ratio.