Public release of hospital quality data for referral practices in Germany: results from a cluster-randomised controlled trial

Table 3 Descriptive analysis and GEE based regression model predicting impact of NHQRS on hospital referrals

Parameter	Description (Overall 20.6%)	Regression analysis
Parameter	Description (Overall 20.6%)	Regression coefficient	Standard error	Odds ratio	95%-Wald CI lower bound	95%-Wald CI upper bound	p-value
Age		−0.184	0.142	0.83	0.63	1.77	0.196
Gender
Male	17.7%
Female	26.8%	0.072	0.429	1.08	0.46	2.49	0.866
Medical discipline
General practitioner	33.3%**
Specialist	10.8%	−0.721	0.899	0.49	0.08	2.83	0.422
Practice type
Single physician practice	19.1%
Otherwise	20.6%	0.339	0.431	1.40	0.60	3.27	0.431
Performing publicly reported procedure
No	24.2%
Yes	11.5%	0.048	0.734	1.05	0.25	4.42	0.948
Dissemination channel
Mass Media	13.2%*
Mass and Special Media	26.0%	0.546	0.443	1.73	0.72	4.12	0.219
Communication measure (survey)
Email	15.3%
Mail	23.1%	0.684	0.515	1.98	0.72	5.43	0.184
Time of survey
Before the last published report	31.5%*
After the last published report	9.1%	−1.155	0.880	0.32	0.06	1.10	0.189
Constant	–	−1.216	0.712	0.296	0.07	1.20	0.088

ISSN: 2191-1991