Table 4.

Univariate logistic regression model is firstly used to select important variables and then multivariate logistic regression model with variables of age, average working time with digital screens, screen size of these devices, history of use of screens (years), cell phone use for calls (talk mode), stand-by time (cell phone use when it is switched on but there are no calls, internet use, etc.), cellular internet use, cordless phone use, sleep disorders is done.

Predictor		Odds Ratio (OR)	CI (95%)	Significance
All Users	The history (years) of screen use	0.725	0.595–0.883	P=0.001
	The daily use of mobile phone in standby mode	0.380	0.244–0.592	P<0.001
Heavy Users	Using digital screens >30 min/day in dim light before sleep in breast cancer patients compared to controls	1.156	0.6307 to 2.1202	NS P=0.630

CI: Confidence Interval, NS: Non-significant