Table 6.
Mixed-effects linear regression model of sleep onset latency time in minutes by type of assessment (habitual, AM estimated, and PSG) and by predictive variables
| Variables | Regression coefficient | p-value | 95% CI* |
|---|---|---|---|
| Habitual | 0.98 | 0.552 | 0.93–1.04 |
| AM Estimated | 1.29 | <0.0001 | 1.22–1.37 |
| Female sex | 1.14 | <0.0001 | 1.07–1.22 |
| Ethnicity, other than Caucasian | 1.16 | 0.001 | 1.06–1.27 |
| 12–16 yr of education | 0.93 | 0.179 | 1.18–1.03 |
| >16 yr of education | 0.83 | <0.002 | 0.74–0.94 |
| Any Heart Disease | 1.15 | 0.001 | 1.06–1.24 |
| Intercept | 15.5 | <0.0001 | 13.8–17.4 |
CI = confidence interval. Note: Values for habitual, am estimated, and polysomnogram sleep onset latency time were log transformed for this test, anti-log values are presented here. PSG is the reference category for type of assessment, male is the reference category for sex, Caucasian is the reference category for ethnicity, <12 years of education is the reference category for education, no heart disease is the reference category for any heart disease, and Pacific/Eastern is the reference category for time zone.