Sleep and Aging in Population Research

Sleep is very important to older adults, and research studies about sleep, particularly those that identify problems with sleep and consequent ill health, receive significant media attention. However, making sure that research findings about sleep are accurate and informative is quite challenging. Incorporating sleep into studies of health and aging is complicated both conceptually and practically. Sleep is not a single entity, and no single approach to measuring sleep can capture all the dimensions that may be relevant for health, nor is there a summary number that describes it. Sleep is multi-faceted with behavioral, social, psychological, and biological components. Adding to the complexity of learning about sleep from individual respondents is the problem that they do not observe many aspects of sleep. Thus, the reporting challenges for sleep are different and perhaps even greater than health behaviors well recognized to be challenging for individuals to report accurately such as diet and physical activity. In addition, sleep research has multiple disciplinary origins, and they approach the conceptualization and measurement of sleep in different ways. Psychological approaches often rely on instruments with multiple questions that are designed to tap into latent constructs; epidemiological approaches query quantifiable characteristics; and clinical sleep medicine focuses on diagnosing defined sleep disorders or “diseases.”

This paper will first sketch out why sleep is so complicated by describing some of the dimensions of sleep that may be relevant for population studies and the ways sleep has been most often measured in population research. As with any variable employed in research about groups of respondents, random noise is much less of a problem for inferences than is biased measurement. Finally, there will be a summary of outstanding questions about sleep and health that studies of older adults might investigate.