Table 6. Subgroup analyses of methodological quality assessment (n/%).

Items (Yes)	All NMAs (n = 102)	Journal impact factor*		Year of publication#		Funding source		Country of corresponding author		Type of NMAs
		Low (<5.00) (n = 55) vs. High (≥5.00) (n = 32)	P-value	Older (n = 58) vs. Recent (n = 44)	P-value	Funding (n = 46) vs. Non-funding (n = 56)	P-value	China (n = 29) vs. Others (n = 73)	P-value	Bayesian NMAs (n = 61) vs. Adjusted indirect comparisons (n = 43)&	P-value
Was the research question (i.e., research purpose, inclusion and exclusion criteria) clarified?	62/60.78	31/56.36 vs. 20/62.50	0.568	33/56.90 vs. 29/65.91	0.358	26/56.52 vs. 36/64.29	0.426	20/68.97 vs. 42/57.53	0.288	38/62.30 vs. 25/58.14	0.671
Was there duplicate study selection and data extraction?	74/72.55	38/69.09 vs. 24/75.00	0.547	41/70.69 vs. 33/75.00	0.631	32/69.57 vs. 42/75.00	0.543	26/89.66 vs. 48/65.75	0.015	50/81.97 vs. 26/60.47	0.015
Was a comprehensive literature search performed?	58/56.86	25/45.45 vs. 25/78.13	0.002	31/53.45 vs. 27/61.36	0.426	26/56.52 vs. 32/57.14	0.950	16/55.17 vs. 42/57.53	0.829	43/70.49 vs. 17/39.53	0.002
Was the status of publication (i.e. grey literature) used as an inclusion criterion?	81/79.41	43/78.18 vs. 26/81.25	0.728	46/79.31 vs. 35/79.55	0.977	35/76.09 vs. 46/82.14	0.454	26/89.66 vs. 55/75.34	0.109	53/86.89 vs. 29/67.44	0.017
Was a list of included studies provided?	94/92.16	50/90.91 vs. 29/90.63	0.965	53/91.38 vs. 41/93.18	0.739	44/95.65 vs. 50/89.29	0.236	28/96.55 vs. 66/90.41	0.300	56/91.80 vs. 40/93.02	0.819
Were the characteristics of the included studies provided?	100/98.04	54/98.18 vs. 31/96.88	0.715	58/100.00 vs. 42/95.45	0.103	45/97.83 vs. 55/98.21	0.889	29/100.00 vs. 71/97.26	0.370	59/96.72 vs. 43/100.00	0.233
Was the scientific quality of the included studies assessed and documented?	70/68.63	34/61.82 vs. 25/78.13	0.098	32/55.17 vs. 38/86.36	0.001	29/63.04 vs. 41/73.21	0.273	25/86.21 vs. 45/61.64	0.016	48/78.69 vs. 24/55.81	0.013
Was the scientific quality of the included studies used appropriately in formulating conclusions?	31/30.39	11/20.00 vs. 8/25.00	0.584	12/20.69 vs. 19/43.18	0.015	9/19.57 vs. 22/39.29	0.032	20/68.97 vs. 11/15.07	0.000	21/34.43 vs. 10/23.26	0.222
Were the methods used to combine the findings of studies appropriate?	66/64.71	32/58.18 vs. 26/81.25	0.019	37/63.79 vs. 29/65.91	0.826	28/60.87 vs. 38/67.86	0.465	18/62.07 vs. 48/65.75	0.727	39/63.93 vs. 29/67.44	0.713
Was the likelihood of publication bias assessed?	16/15.69	3/5.45 vs. 8/25.00	0.017	10/17.24 vs. 6/13.64	0.622	5/10.87 vs. 11/19.64	0.228	6/20.69 vs. 10/13.70	0.383	8/13.11 vs. 9/20.93	0.291
Was the conflict of interest stated?	62/60.78	38/69.09 vs. 22/68.75	0.974	37/63.79 vs. 25/56.82	0.477	29/63.04 vs. 33/58.93	0.673	11/37.93 vs. 51/69.86	0.001	35/57.38 vs. 27/62.79	0.581

^*6 studies published in journals with no associated impact factor.

^#Based on the median division of number of included NMAs, December 31st 2013 is the cut-off point.

^&2 adjusted indirect comparisons also were conducted using Bayesian framework.