Table 3.

GMDR models of sleep quality

Model	Prediction accuracy	Sign test (P)	Cross-validation consistency
Working shifts	0.55	5 (0.62)	7/10
Working years, working shifts	0.65	9 (0.01)	10/10
Working years, occupation, working shifts	0.55	7 (0.17)	8/10
Working years, occupation, working shifts, supporting Wuhan	0.56	7 (0.17)	7/10
Gender, working years, occupation, working shifts, supporting Wuhan	0.61	8 (0.05)	10/10

P was adjusted for age, educational level, type of hospital, and working position using logistic regression in GMDR analysis