Concordance between subjective and objective measures of infant sleep varies by age and maternal mood: Implications for studies of sleep and cognitive development

Abstract

Infant habitual sleep has been proposed as an important moderator of development in domains such as attention, memory or temperament. To test such hypotheses, we need to know how to accurately and consistently assess habitual sleep in infancy. Common assessment methods include easy to deploy but subjective parent-report measures (diary/sleep questionnaire); or more labour-intensive but objective motor movement measures (actigraphy). Understanding the degree to which these methods provide converging insights is important, but cross-method agreement has yet to be investigated longitudinally. Moreover, it is unclear whether concordance systematically varies with infant or maternal characteristics that could represent confounders in observational studies. This longitudinal study (up to 4 study visits/participant) investigated cross-method concordance on one objective (7-day actigraphy) and three commonly used subjective (7-day sleep diary, Brief Infant Sleep Questionnaire, Sleep & Settle Questionnaire) sleep measures in 76 typically developing infants (age: 4-14 months) and assessed the impact of maternal characteristics (stress, age, education) and infant characteristics (age) on cross-method concordance. In addition, associations between objective and subjective sleep measures and a measure of general developmental status (Ages & Stages Questionnaire) were investigated. A range of equivalence analyses (tests of equivalence, correlational analyses, Bland-Altman plots) showed mixed agreement between sleep measures. Most importantly, cross-method agreement was associated with maternal stress levels and infant age. Specifically, agreement between different measures of night waking was better for mothers experiencing higher stress levels and was higher for younger than older infants; the reverse pattern was true for day sleep duration. Interestingly, objective and subjective measures did not yield the same patterns of association with developmental domains, indicating that sleep method choice can influence which associations are found between sleep and cognitive development. However, results converged across day sleep and problem-solving skills, highlighting the importance of studying day sleep in future studies. We discuss implications of sleep method choice for investigating sleep in the context of studying infant development and behaviour.

Keywords: Actigraphy; Cross-method concordance; General development; Infant sleep; Maternal stress; Parent-report.

Fig. 1

Illustration of correlations between the four different sleep parameters of objective and subjective sleep measures. Note. BISQ = Brief Infant Sleep Questionnaire, NWNr = Night waking number, WASO = Wake after sleep onset, SSQ = Sleep & Settle Questionnaire, diary = sleep diary, r = Pearson’s correlation coefficient.

Fig. 2

Bland-Altman plots for main sleep parameters. Note. Plotting mean of the two measures (x) against the difference of the two measures (y). Blue line = proportional bias line i.e., mean difference between the two measures. Dotted lines = ± 1.96 *SD margins of the mean difference between the measures. If methods show good agreement the bias line should coincide with x = 0 line. If bias line is parallel to x = 0 it means that the bias is the same across participants.

Fig. 3

Illustration of cross-method difference of actigraph vs. diary in several sleep variables by infant age and maternal stress/anxiety level. Notes. r = Pearson's correlation coefficient, maternal stress/anxiety score = z-scored composite score of PSS and STAI-T/-S.

Fig. 4

Day sleep duration (in minutes) for Actigraphy (A), Diary (B), Brief Infant sleep questionnaire (BISQ; C) and for Sleep and Settle Questionnaire (SSQ; D). Note. Figures for night waking parameters in SM.

Fig. 5

Night sleep duration (in minutes) for Actigraphy (A), Diary (B), Brief Infant sleep questionnaire (BISQ; C) and for Sleep and Settle Questionnaire (SSQ; D). Note. Figures for night waking parameters in SM.

Fig. 6

Illustration of the developmental changes in the association between objectively (A) and subjectively (B – C) measured day sleep and parent-reported problem-solving skills. Note. ASQ = Ages & Stages Questionnaire.

Fig. 7

Summary graph of the association of development with sleep parameters. Note. Grey = no association between sleep and development, blue = age-related changes in the association between sleep and development, red = negative association between sleep and development. COM = Communication, GM = Gross motor skills, FM = Fine motor skills, PS = Problem-solving skills, SOC = social skills. For detailed age-related changes see SM and above Section 3.5.

Fig. 8

Illustration of questions to consider when choosing actigraphy vs. parent-report measures. Note. & = both actigraph and questionnaires can be chosen. / = either actigraph or parent-report can be chosen.