Poor reporting quality of observational clinical studies comparing treatments of COVID-19 – a retrospective cross-sectional study

Table 2 Multiple linear regression model for the percentage adherence to the STROBE checklist

Independent variable	Point estimate of change in percentage adherence (Unstandardised coefficient β)			Standardised coefficient Beta	P-value
Independent variable	Difference	95% CI	SE	Standardised coefficient Beta	P-value
Month of publication	.847	-.177 to 1.872	.518	.130	.104
STROBE mentioned	4.304	-2.808 to 11.415	3.594	.096	.233
STROBE in author guidelines	4.760	.181 to 9.338	2.314	.171	.042
Impact factor ^a
2^nd quartile	8.633	2.745 to 14.521	2.976	.278	.004
3^rd quartile	6.658	.772 to 12.543	2.974	.215	.027
4^th quartile	11.072	5.121 to 17.022	3.007	.357	< .001
Country of origin ^b
China	-6.653	-12.230 to -1.075	2.819	-.224	.020
Italy	-6.574	-13.383 to .236	3.441	-.168	.058
Spain	1.566	-6.020 to 9.153	3.834	.035	.684
France	-1.927	-12.635 to 8.781	5.411	-.029	.722
Great Britain	.0420	-11.451 to 11.535	5.808	.001	.994
Other countries	-7.838	-14.869 to -.807	3.553	-.195	.029

The present table shows the effects of the prespecified independent variables (predictors) on the percentage adherence to the STROBE checklist according to a multiple linear regression model. Overall regression model: R2 = 0.285; adjusted R2 = 0.217; F (12,127) = 4.214; p = < .001. Dataset n = 140; missing values for IF n = 7
^a Estimates to be interpreted in relation to 1^st quartile. Quartile boundaries: 1^st: 0.717 to 2.739, 2^nd: 2.740 to 3.639, 3^rd: 3.656 to 5.893, 4^th: 6.407 to 74.669
^b Estimates to be interpreted in relation to the reference country USA

ISSN: 1471-2288