Agreement analysis of sleep patterns between self-reported questionnaires and actigraphy in adults

Abstract

Purpose: To investigate the agreement in sleep pattern recording by self-reported sleep questionnaires and actigraphy in adults.

Methods: This is a cross-sectional study. Men and women who met inclusion criteria were recruited for this study. The inclusion criteria were apparently healthy Omani nationals ages 19 to 50 years. Sleep questionnaires were randomly distributed in Muscat either directly or via electronic and paper announcements. Data were collected from the participants using the self-reported questionnaires with four piloted questions for sleep pattern identification and through the actigraphy wristband given to subjects to wear for a week. Cohen's kappa test was performed for agreement analysis.

Results: A total of 964 Omani subjects between ages 18 and 59 years of both genders were recruited and completed the questionnaires successfully. Out of these, only 321 subjects wore the actigraphy wristband for 1 week (response rate = 33%). Agreement analysis reported a mild level of agreement for the monophasic (41%), moderate level for biphasic (59%), and good level for polyphasic (70%) sleep patterns. The overall agreement level of sleep patterns between the two methods was 57%. There is a low specificity of self-reported assessment in reporting sleep pattern.

Conclusion: The average agreement level of subjective versus objective assessments of sleep patterns was moderate at 57% and self-reported sleep pattern is not specific. The study recommends the use of actigraphy along with sleep questionnaires for accurate assessment of sleep patterns in cohort studies.

Keywords: Agreement analysis; Sleep patterns; Sleep study; Validity.