Table 3.

Association between change in sleep duration and subsequent incident diabetes across four data cycles using all available glycemic data and participant report of doctor-diagnosed diabetes

			Confounder adjustments
Change in sleep duration	No. events	N*	Age, sex	Age, sex, ethnic group	Age, sex, ethnic group, employment grade	Age, sex, ethnic group, employment grade and BMI at the beginning and end of each exposure period
			OR (95% CI)	OR (95% CI)	OR (95% CI)	OR (95% CI)
≥2 hours decrease in sleep	32	533	1.44 (0.97, 2.12)	1.24 (0.83, 1.84)	1.22 (0.82, 1.81)	1.16 (0.78, 1.73)
No change in sleep†	204	4875	1.00 (ref‡)	1.00 (ref‡)	1.00 (ref‡)	1.00 (ref‡)
≥2 hours increase in sleep	40	566	1.83 (1.28, 2.60)	1.69 (1.18, 2.42)	1.65 (1.15, 2.37)	1.50 (1.04, 2.16)

^*Number of person-observations

^†7 hours at each data cycle

^‡Odds ratios compared to those who had 7 hours sleep on both occasions