Subjective reports of children’s sleep duration: Does the question matter? A literature review

Abstract

Questionnaires are widely used to determine children’s sleep duration. While authors have been encouraged to move on from the question: ‘how many hours of sleep do you usually get?’, the validity of different reports remains unclear, and there do not appear to be any evidence-based recommendations for how to best enquire about children’s sleep.

Our study aims to determine the validity of subjective reports of children’s sleep duration and if elements of phrasing influence the accuracy of reports.

A systematic literature review was conducted to identify studies that report the subjective sleep duration of children ages five to 18 years, inclusive. A thematic analysis was used to identify ‘elements’ of phrasing that may influence the accuracy of reports, and a descriptive analysis was conducted to determine the validity of subjective reports.

Our review identified five elements of phrasing that may affect the accuracy of reports. These include, specifying sleep parameters that define sleep duration, the recall period, the day type, the method of data collection, and the respondent. The validity of subjective reports varied across studies and in accordance with identified elements of phrasing.

Future studies that assess subjective reports of children’s sleep should define sleep duration, the recall period, the days constituting schooldays and nonschooldays, the method of data collection, and the respondent of the questionnaire.

Introduction

There have been an increasing number of studies that have investigated the relationship between subjective reports of children’s sleep duration and health outcomes. These studies have identified strong associations between short sleep duration and cognitive [1], [2], [3], [4], behavioural [3], [5], and physiological [6], [7], [8], [9] impairments. Some of these findings have been further supported in studies [3], [10], [11], [12], [13] that use objective measures of sleep (i.e., polysomnography [PSG] and actigraphy [acti]). Given the importance of subjective sleep reports in both the clinical and the research setting as an indicator of impaired health and cognitive function, the need to verify subjective reporting has become increasingly important, especially for school-aged children (ages 5–18 years, inclusive) who are prone to sleep deprivation [14], [15], [16] and often are involved in school-based studies [17], [18], [19], [20] and sleep interventions [21] that require children to provide subjective sleep reports.

Although there are many advantages of collecting subjective reports over objective measures of children’s sleep duration such as ease of use and minimal time and costs involved in data collection, there do not appear to be any gold-standard instruments or questionnaires to assess children’s sleep duration. As Spruyt and Gozal [22] noted in a recent systematic review of all published and unpublished instruments used to investigate or evaluate children’s sleep, ‘too many sleep instruments are being used that have not undergone careful and methodological psychometric evaluation and therefore may be fraught with bias or invalid findings’.

Questions that determine children’s sleep duration constitute just one of the components contained within children’s sleep questionnaires. However, the validity and accuracy of these questions may be of particular importance, as a growing body of evidence suggests that children’s sleep duration may predict various adverse physical, behavioural, and academic outcomes. In addition, it has been argued that interventions are needed to increase children’s sleep time [23]. If subjective reports of children’s sleep are to inform public policies, guidelines, interventions, and preventative measures, it is essential that reports have established validity.

Few studies have comprehensively reviewed the evidence for the validity of subjective sleep reports of school-aged children. Although several reviews [24], [25], [26] have discussed different methods of assessing children’s sleep and have examined psychometric characteristics of children’s sleep questionnaires, only one study [27] appears to have systematically reviewed the evidence for the accuracy of subjective sleep reports. This study by Bauer and Blunden [27], suggested that subjective reports of children’s total sleep time are not as strongly correlated to objective sleep measures as other sleep parameters such as bedtime and wake time. Although the authors of this study suggested that discrepancies likely are multifactorial and are related to differences in methodological approaches, studies were not further examined to verify if methodological factors did influence reports. Thus, there is a need to systematically review the literature to determine the validity of children’s subjective sleep reports and to better understand if methodological approaches such as the phrasing of questions, influence the accuracy of reports.

To date, there do not appear to be any guidelines for the best way to enquire about children’s sleep duration. Because there are several ways that sleep duration may be asked about, interpreted, calculated and reported (see Fig. 1), it could be argued that validity has only been established for a select number of questions. Indeed, given that some sleep parameters such as night time awakenings and sleep onset latency are poorly reported [27], [28], [29], it would seem plausible that the sleep parameters selected to define sleep duration could influence the accuracy of subjective reports. Although there do not appear to be any studies to support this, unpublished evidence (Partinen, oral communication [30]) exists to suggest that the phrasing of sleep questions does influence the validity of reports; it has been suggested that “authors should move on from single-phrased questions” to improve the accuracy of reports [31].

Although avoiding single-phrased questions may improve the accuracy of reports by reducing the risk for construct underrepresentation and construct-irrelevant variance [32], there neither appear to be any evidence-based recommendations regarding the additional detail that should be asked for, nor any evidence-based recommendations regarding how the phrasing of specific questions could improve the accuracy of reports. Thus, although it may be important to move on from single-phrased questions, such as, “how many hours of sleep do you usually get?”, further research is needed to determine if there are specific elements within questions that enquire about sleep that may influence the accuracy of reports. For example, although recall bias is recognised as an important source of bias in epidemiological studies [33], it remains unclear as to how different recall periods influence the accuracy of subjective sleep reports. Similarly, although several studies [34], [35], [36], [37] have reported that significant differences exist in the sleep of children across different days of the week, it remains unclear if the accuracy of sleep reports differs according to day type (i.e., school, nonschool days). Furthermore as identified in a literature review [27], while parents tend to overestimate their child’s sleep the accuracy of these reports remains unclear, as does the age of the child that parents are best to report their child’s sleep.

The method of data collection (i.e., diary vs. questionnaire reports) also may be of importance. Although there has been little research in this area, evidence exists to suggest that the method of data collection could indeed influence study results. For example, Knutson and Lauderdale [38] noted that although questionnaire reports of children’s sleep duration were associated with body mass index, time-diary sleep reports were not. Beebe et al. [39] also noted that the relationship between weight status and sleep duration was dependent on the method used to assess children’s sleep time, with objective measures revealing no association and subjective reports revealing notable inverse associations. Similarly, Short et al. [40] noted that while diary-reported sleep durations were associated with feelings of sleepiness and fatigue, actigraphy measures were not. Thus, the measure of sleep duration may be a confounding factor underlying associations between sleep and health outcomes. Interestingly Guidolin and Gradisar [41] in a literature review that determined the association between children’s sleep duration and obesity, reported significant associations across all studies that assessed sleep duration using a non validated instrument but no association for studies that used the validated School Sleep Habits Survey [2]. We acknowledge a need for future research to determine the most sensitive and accurate manner of calculating self-reported sleep duration estimates.

There are many different methods to determine subjective estimates of children’s sleep including, questionnaires, sleep diaries, time use diaries, and sleep logs. Questionnaires are available in a variety of different formats (e.g., handwritten, electronic, and computer-assisted telephone interview), may enquire about children’s sleep duration in a number of different ways (e.g., multiple-choice response, open-ended response, and time-use diaries) and be a question within a general health and lifestyle questionnaire or a sleep-specific questionnaire (e.g., the Children’s Sleep Habits Questionnaire [42] and the Sleep Timing Questionnaire [43]). Subjective reports of children’s sleep duration also may be collected via diary methods. Time-use diaries require participants to recall all daily activities (often in 5-, 10-, or 15-min time slices) over a 24-hour period, while sleep diaries require participants to recall only sleep information (i.e., the time that they went to bed, fell asleep, woke up during the night and in the morning). Although similar to sleep diaries, sleep logs often are presented graphically as a timeline, whereby the participants are required to use symbols to mark the time they went to bed, fell asleep, woke in the morning, and so on. Some sleep diaries and sleep logs also ask participants to rate their sleep quality and report information such as whether or not they exercised and/or consumed caffeine during the day. The format and level of detail required for sleep diaries and sleep logs vary across instruments. This raises the interesting question of whether or not such detail influences the accuracy of reports.

Given the growing number of studies that examine subjective reports of children’s sleep duration, further research is needed to determine if the style of question and/or the method of data collection could influence the accuracy of subjective sleep reports. The primary aim of our study was to conduct a systematic literature review to determine the evidence for the validity of subjective sleep reports (sleep diary, questionnaire and sleep logs) of children ages five to 18 years (inclusive) by comparing these estimates with objective (PSG and acti) measures, with particular focus on determining if elements of phrasing influence the accuracy of reports. Elements of phrasing were identified via a thematic analysis of studies that provided subjective reports of children’s sleep. The identified elements were then used to guide a descriptive analysis of validity studies and to determine if elements of phrasing influenced the accuracy of subjective sleep reports. The final aim of our study was to examine how sleep duration was measured and reported in the literature in an effort to provide recommendations for how to improve current methods of investigation.