Introduction: Epidemiological studies often utilize self-reports to characterize total sleep time (TST). Studies provide discordant information on the association between self-reported and objective TST with some finding weak or no correlations and bias in reporting based on sex and body mass index.
Objective: Our goal was to investigate the correlation between self-reported and objective measures of TST in a diverse sample of adult men and women volunteering to participate in sleep studies and determine whether sex, race/ethnicity, and weight status influence this association.
Methods: Participants were individuals who screened for participation in sleep studies for whom we had >7 d of wrist actigraphy sleep data (TSTobj) and subjective total sleep time (TSTPSQI) reported using the Pittsburgh Sleep Quality Index (PSQI) questionnaire (n=53 men and 60 women; n=67 minority/Hispanics; age 30.9±9.7y; BMI 26.2±3.2kg/m2).
Results: Average TSTPSQI was greater than TSTobj (0.63±0.99h, p<0.0001). In univariate regression analyses, the difference between TSTPSQI and TSTobj did not vary by sex (ß=0.12, p=0.52), race/ethnicity (ß=0.15, p=0.48), age (ß=-0.01, p=0.27), or BMI (ß=0.04, p=0.13). Overall, there was no correlation between TSTobj and TSTPSQI (r=0.12, p=0.22), which did not vary when sex was analyzed separately (men: r=0.20, p=0.15; women: r=0.046, p=0.73). Interestingly, there were moderate and highly significant correlations between self-reported and objective assessments of bedtime (r=0.42, p<0.0001) and wake time (r=0.35, p<0.0001). We therefore used bedtime, waketime, and time to fall asleep information on the PSQI to calculate a new TSTbed/wake. Average TSTbed/wake was greater than TSTobj (0.79±0.76h, p<0.0001). This measure was correlated with TSTobj (r=0.57, p<0.0001); with strong correlation in women (r=0.76, p<0.0001) and moderate correlation in men (r=0.36, p=0.0082). In univariate regression, the difference between TSTbed/wake and TSTobj varied by age (ß=-0.020, p=0.0051), BMI (ß=0.054, p=0.0021), race/ethnicity (ß=0.36, p=0.021), and sleep efficiency (ß=-0.089, p<0.0001).
Conclusions: Self-reported bedtimes and wake times were more accurately reported than estimated TSTPSQI and provided a more reliable estimate of TST. In general, the degree of over-reporting was greater in younger and heavier individuals, non-White/Hispanics, and those with low sleep efficiency. Although self-reported assessments of sleep duration over-estimated objective measures by approximately 45 min, using information on habitual bedtimes, wake times, and time to fall asleep improved the relation in both men and women.