Table 2.

Meta-regression analysis for the possible sources of heterogeneity

Variances	Coefficient standard	Standard error	P value	RDOR	95 % CI
Inverse variance weights 1
Cte	3.344	0.9147	0.0353	–	–
S	0.228	0.083	0.5130	–	–
Setting	0.341	0.2697	0.2952	1.41	0.60–3.32
Instrument	−0.945	0.4349	0.1180	0.39	0.10–1.55
Cut-off	0.046	0.0938	0.6558	1.05	0.78–1.41
Inverse variance weights 2
Cte	3.560	0.8027	0.0114	–	–
S	0.282	0.2884	0.3838	–	–
Setting	0.295	0.2531	0.3082	1.34	0.67–2.71
Instrument	−0.978	0.4297	0.0851	0.38	0.11–1.24
Inverse variance weights 3
Cte	4.244	0.5481	0.0006	–	–
S	0.458	0.2458	0.1215	–	–
Instrument	−1.200	0.3856	0.0265	0.30	0.11–0.81

The RDOR means the DOR for studies that lacked a particular methodologic feature divided by the DOR for studies without the flaw

—, not available, Cte constant term in the equation, RDOR relative diagnostic odds ratio, S indicator of threshold

P < 0.05 indicated the significant relationship between the characteristics of studies and the diagnostic odds ratio