. 2017 Mar 30;12(3):e0174831. doi: 10.1371/journal.pone.0174831

Table 3. Results of the multinomial regression analysis (independent variable: Recommendation for use for the categories “yes” vs. “no”).

Parameter	Estimate	Standard error	Wald chi-square	P-value	Adjusted p-value (sig. < 0.05)	OR^a	95% confidence interval for OR^a
Parameter	Estimate	Standard error	Wald chi-square	P-value	Adjusted p-value (sig. < 0.05)	OR^a	Lower bound	Upper bound
Intercept (recommended)	-9.744	0.856	129.729	< 0.001
Domain 1 (scope and purpose)	0.013	0.009	2.059	0.151	0.227	1.140	0.954	1.367
Domain 2 (stakeholder involvement)	0.013	0.010	1.603	0.206	0.247	1.135	0.933	1.381
Domain 3 (rigour of development)	0.109	0.011	93.824	< 0.001	< 0.001	2.963	2.395	3.719
Domain 4 (clarity of presentation)	0.046	0.010	20.521	< 0.001	< 0.001	1.581	1.301	1.934
Domain 5 (applicability)	0.022	0.009	6.026	0.014	0.028	1.250	1.048	1.498
Domain 6 (editorial independence)	0.003	0.006	0.200	0.657	0.657	1.029	0.909	1.166

^a: The OR corresponds to the change in the respective domain score by 10 percentage points.

Dependent variable: recommendation for use; Reference category: “no”.