Abstract
Purpose: :
This study ascertained the independent association of visual impairment (VI) with sleep duration using the National Health Interview Survey (NHIS) data.
Methods: :
Analysis was based on the 2009 NHIS, providing valid sleep and vision data for 19,985 volunteers (ages: 18-85). Of the sample, 87% was white and 13% was black; 55% were female and 70% had a family income <$35,000. The NHIS is a cross-sectional household interview survey utilizing a multistage area probability design. Probability samples of the civilian population of all 50 states and DC were obtained. Trained personnel from the US census bureau gathered data during face to face interviews. Volunteers provided socio-demographic and subjective data; physician-diagnosed chronic conditions were also elicited.
Results: :
Prevalence of VI was 13% (cataract=11%, glaucoma=3%, macular degeneration=2%, and diabetic retinopathy=1%); 31% were short sleepers (<=6hrs) and 15%, long sleepers (>=9hrs); 33% were obese (>=30kg/m² ); and 34% reported negative moods (depression and/or anxiety). Long sleep was more common among blacks [18% vs. 14%, p<0.001], whereas VI was more common among whites [13% vs. 10%, p<0.001]. Multivariate adjusted logistic regressions showed stronger independent associations of VI with long, rather than, short sleep [OR=1.51, p<0.001, OR=1.07, p<0.001, respectively]. Since our first regression model indicated that obesity and mood also had strong associations with long sleep, we ascertained interactions of those factors with VI on long sleep. Among blacks, greater odds of being a long sleeper were observed for individuals with VI and obesity, rather than with VI and mood [OR=1.43, p<0.001, OR=1.01, p<0.001, respectively]. However, among whites greater odds were noted for individuals with VI and mood, rather than with VI and obesity [OR=1.59, p<0.001, OR=1.20, p<0.001, respectively]. Effects of sociodemographic and medical factors were adjusted in the model.
Conclusions: :
Findings suggest that analysis of epidemiologic sleep data should consider visual impairment in assessing odds of long sleep. Plausibly, individuals with visual impairment spend more time at home, providing a greater opportunity to spend time in bed, likely increasing sleep time. Race/ethnicity is an important mediator in the link of visual impairment to long sleep duration.
Keywords: clinical (human) or epidemiologic studies: biostatistics/epidemiology methodology