Table 1. Univariate and multivariate regression models.

Model: # Google scholar cites		Estimate (SE)	t value	p value	OR [CI95%]
Methodological quality	High	-	-	-
	Moderate	-0.237 (0.299)	-0.793	ns
	Low	-0.417 (0.396)	-1.052	ns
Funding sources	Academic	-0.542 (0.291)	-1.863	ns
	Pharma	-0.289 (0.357)	-0.808	ns
Author′s conflict of interest		0.009 (0.038)	0.253	ns
Impact factor		0.029 (0.059)	0.500	ns
Year of publication	2011	-	-	-
	2012	0.545 (0.944)	0.577	ns
	2013	0.629 (0.873)	0.720	ns
	2014	-0.383 (0.893)	-0.429	ns
	2015	-2.156 (0.903)	-2.388	<0.016	0.115 [0.019; 0.679]
	2016	-3.645 (0.930)	-3.917	<0.001	0.026 [0.004; 0.161]
Number of tweets		0.015 (0.020)	0.754	ns
Mendeley′s readers		0.033 (0.007)	4.263	<0.001	1.033 [1.018; 1.049]
Model: # Tweets		Estimate	SE	t value	OR [CI95%]
Methodological quality	High	-	-	-
	Moderate	0.627 (0.548)	1.142	ns
	Low	0.321 (0.629)	0.510	ns
Funding sources	Academic	0.750 (0.557)	1.346	ns
	Pharma	-0.271 (0.687)	-0.394	ns
Author′s conflict of interest		0.221 (0.120)	1.833	ns
Impact factor		0.184 (0.051)	3.584	<0.001	1.202 [1.087; 1.329]
Year of publication	2011	-	-	-
	2012	1.316 (0.863)	0.271	ns
	2013	1.053 (0.807)	0.406	ns
	2014	1.139 (0.811)	-2.661	ns
	2015	1.360 (0.804)	-2.436	ns
	2016	0.690 (0.823)	-4.26	ns

…