Table 4.

Models for percent completeness as a function of administrative functions and hospital characteristics^a,b

Predictor	AMI percent completeness			Stroke percent completeness			Hip percent completeness			Delivery percent completeness
	B	se	p	b	se	p	b	se	p	b	se	p
Data is systematically abstracted from the medical record into the hospital database	1.27	11.38	0.911	8.43	11.32	0.457	0.13	9.69	0.989	11.36	11.92	0.342
The hospital database is part of a national database	12.42	5.30	0.020	−2.37	5.12	0.644	3.57	4.29	0.406	−5.34	5.43	0.327
The hospital database is used for reimbursement purposes	4.64	8.18	0.571	14.77	8.08	0.069	3.90	6.88	0.572	4.12	8.52	0.63
Public hospital	−7.82	7.21	0.280	2.26	6.67	0.735	−5.19	5.43	0.340	−3.86	7.15	0.589
Teaching hospital	−0.56	5.79	0.923	3.38	5.74	0.557	2.13	4.90	0.664	6.86	6.05	0.258
Hospital size
<200	Ref	Ref	Ref	Ref	Ref	Ref	Ref	Ref	Ref	Ref	Ref	Ref
200–500	−5.87	7.85	0.456	13.86	10.03	0.169	−0.63	8.48	0.940	14.24	10.60	0.181
501–1000	−3.03	8.48	0.721	7.68	7.43	0.303	2.51	6.30	0.691	−1.24	7.85	0.874
>1000	−9.57	10.26	0.352	9.57	7.43	0.199	3.90	6.29	0.536	0.59	7.85	0.939

^aData completeness defined here as the proportion of hospitals with complete data on the items required for the calculation of the indicator.

^bMultivariate linear mixed model with random intercept by country; all adjustment variables displayed.