Sex‐specific associations between self‐reported sleep duration, depression, anxiety, fatigue and daytime sleepiness in an older community‐dwelling population
For measuring sleep disturbances, several approaches can be used. The golden standard for objective registration of sleep is polysomnography (PSG) in a sleep laboratory 7. However, a PSG, which includes recording the subjects’ electroencephalography, electrooculography (i.e. eye movements) and surface electromyography (i.e. muscle tone), is unfeasible due to high costs, low availability and need for specific competence. Self‐assessment from patients based on a questionnaire is another option. There are several questionnaires to measure different aspects of sleep and associated problems 8, but despite established validity and reliability, and frequent use in research, these instruments are not commonly used by nurses in clinical practice. Therefore, a less burdensome and straight‐forward indicator to identify older individuals with sleep complaints in need of an intervention could be self‐reported sleep duration. Previous research 2 has used various cut‐offs, and a common way to subdivide samples according to sleep duration is short sleep (≤6 hours), normal sleep (7–8 hours) and long sleep (≥9 hours) 2. Sleep duration can be used by nurses as a suitable topic for the initial question to identify elderly with sleep disturbances. How to focus potential ‘follow‐up questions’ (e.g. questions to explore gender‐based associates to sleep problems) might be difficult, but still important.
There are sex differences both in disturbed sleep and depressive symptoms, with higher prevalence among older women than men 10. Women tend to report more sleep complaints 11, and men tend to have somewhat longer sleep duration according to some 12, but not according to other studies 10. Also, older men seem to report more daytime sleepiness in some studies 12, but not in other 13. Sex differences are well documented in various scientific disciplines. One aspect of sex differences is that the sizes may vary across the adult age span. Several theoretical accounts exist at the biological level, but also explanations including differential life experiences and social expectations. Importantly, although less well investigated, the impact of a causative factor may vary between the two sexes, irrespective of differences at the mean level in an outcome of interest 14. Such sex‐specific effects may be examined by statistical models including sex X predictor interaction terms. In both sexes, an increased number of medical disorders and lower social support are associated with psychological distress.