Reducing unnecessary hospital days to improve quality of care through physician accountability: a cluster randomised trial

Table 3 Multivariate models

*Strategy’s effectiveness (N=3498)* *
Variable	OR	95% CI	P
Intervention	0.841	0.735 - 0.963	0.01
Age	1.004	0.998 - 1.010	0.18
Sex	0.921	0.803 - 1.057	0.24
General Medicine Unit	0.604	0.500 - 0.730	<0.001
Geriatrics Unit	0.739	0.592 - 0.924	0.008
Charlson	1.079	1.009 - 1.155	0.02
Follow-up (N=248) at 1 year *
Variable	OR	95% CI	P
Intervention	0.818	0.476 - 1.405	0.47
Age	0.998	0.976 - 1.020	0.83
Sex	0.877	0.514 - 1.494	0.63
General Medicine Unit	0.440	0.162 - 1.195	0.11
Geriatrics Unit	0.575	0.183 - 1.808	0.34
Charlson	1.324	0.997 - 1.759	0.05
*Overall LOS (* N=3498 ) †
Variable	IRR	95% CI	P
Intervention	0.829	0.718 - 0.958	0.01
Age	0.982	0.963 - 1.001	0.07
Sex	0.939	0.758 - 1.163	0.57
General Medicine Unit	0.221	0.174 - 0.282	<0.001
Geriatrics Unit	0.293	0.224 - 0.383	< 0.001
Charlson	1.020	0.958 - 1.087	0.53
Reason: “ Without any apparent reason ” ( *N=1682* ) * ‡
Variable	OR	95% CI	P
Intervention	0.672	0.539 - 0.837	0.001
Age	1.017	1.006 - 1.028	0.003
Sex	0.804	0.644 - 1.005	0.06
General Medicine Unit	0.253	0.188 - 0.341	<0.001
Geriatrics Unit	0.134	0.094 - 0.193	<0.001
Charlson	1.059	0.950 - 1.181	0.30

* Estimates of the relative rate (OR - Odds Ratio) were obtained from multivariate logistic model.
† Estimates of the relative rate (IRR - Incident Rate Ratio) were obtained from multivariate Poisson model.
‡ The number of observations is smaller because the estimates are based on the subset of patient - days showing delays due to problems with activities under medical staff control.

ISSN: 1472-6963