Assessing the sleeping habits of patients in a sleep disorder centre: a review of sleep diary accuracy

Background: Excessive daytime sleepiness (EDS) is a complaint common to many aspects of medicine. There are primary and secondary causes for EDS, with secondary causes including a large number of common conditions. Primary causes, such as narcolepsy, are much rarer. When assessing for primary hypersomnia, restricted or fragmented sleep must be ruled out. This process involves assessment of sleeping habits using a sleep diary and/or actigraphy. Clinicians are suspicious of the accuracy with which patients use the former. This review aims to evaluate the accuracy of a sleep diary study against the ‘objective gold standard’ actigraphy report. Methods: Data from 35 patients at a Sleep Disorder Centre who underwent both a sleep diary and actigraphy study for suspected primary hypersomnia in 2016 was collected. Mean values of four variables were calculated: ‘time of lights out’, ‘time to fall asleep’, ‘time of waking’ and ‘sleep time’. The ‘similarity’ was assessed. This was a term defined in three different ways: if sleep diary values are accurate to within 20, 30 and 60 min respectively. Percentage ‘similarity’, mean time differences and standard deviations (SDs) were calculated for each variable. A paired t-test was also performed to assess the significance of the time differences between the two modalities. Results: Least accurate was ‘sleep time’, with 14.7%, 23.5% and 58.8% of patients within 20, 30 and 60 min of the actigraphy respectively. Mean time difference for this variable was 66 min (versus 33, 15 and 22). ‘Time to fall asleep’ was most accurate, with 76.5%, 82.4% and 100% ‘similarity’ respectively. Conclusions: The clinically acceptable accuracy has no universal definition, so clinicians must use experience and reasoning to determine this level to interpret this data. The review suggests that some variables are entered with high accuracy, and the diary is low cost and adds subjective information that cannot be gathered from actigraphy. Therefore, use is recommended to continue alongside actigraphy.