Associations between parent gender and young children’s sleep in heterosexual couples in the Guelph Family Health Study: a preliminary analysis

Introduction

Sleep is widely recognized as essential for physical, cognitive, and socio-emotional development of young children. According to the Family Systems Theory [1], family members are interdependent given their shared physical and emotional environment [2]. Investigating sleep hygiene within the family unit may provide a deeper understanding into sleep behaviors of young children, especially as parents often serve as the external organizers of their child’s sleep patterns [3].

Previous research has examined family sleeping behaviors in relation to parent self-identified gender. A review by Varma et al. (2021) analyzed 30 studies of children aged 2–18 years, in which sleep was assessed using either subjective or objective measures. The review found strong evidence for associations between mother–child sleep quality and quantity, whereas little to no associations were observed for fathers. A limitation of existing research is its reliance on subjective sleep measures (e.g. questionnaires), which may introduce bias as mothers are typically the primary reporters of their children’s sleep. A systematic review by Coles et al. (2022), which examined the impact of children’s sleep on fathers’ health and well-being, found significant father–child sleep associations when fathers actively reported on their children’s sleep. The review included studies of children aged 0–19 years, with sleep measured using either subjective or objective methods. To enhance accuracy, researchers have emphasized the need to implement objective sleep assessment tools such as accelerometers (also known as actigraphy).

The current study objectively assessed sleep timing, quality, and duration in toddlers and preschool-aged children (1.5 to <6 years) to determine if these variables were associated with their mothers’ and fathers’ sleep quality, duration, and timing. Elucidating associations between parent–child sleep patterns may inform the design of more effective sleep interventions and enhance our understanding of how sleep is experienced within young families. Based on previous literature [4,5] demonstrating strong associations between mothers’ and children’s sleep, and limited evidence for such associations between fathers and children, we hypothesized that sleep quality, duration, and timing would be related between mothers and their children, but not between fathers and their children.

Materials and Methods

Our investigation included baseline data of mother–father–child triads enrolled in the Guelph Family Health Study (GFHS), a family-based obesity prevention intervention study (ClinicalTrials.gov ID—NCT02939261). Inclusion criteria required that families have at least one child between the ages of 18 months to <6 years at the time of enrollment, reside in Guelph or the surrounding area and have at least one parent who could respond to surveys in English. Specific to the current analysis, mother–father–child triads were only included if they had a minimum of three valid nights of overlapping sleep. The study’s protocol was approved by the University of Guelph Research Ethics Board and conducted in accordance with the Declaration of Helsinki. Parent’s provided informed consent for themselves, at least one parent provided written consent for their child to participate, and children provided verbal assent.

Children and their parents were given ActiGraph wGT3X-BT (100 Hz sampling frequency) accelerometers and were instructed to wear them on the iliac crest of their right hip for a period of 7 days, 24 h a day. Additionally, parents were asked to complete activity log sheets that recorded bedtimes, rise times, nap times, and non-wear times for themselves and their child.

Details of the data analysis have been reported previously [6] but are summarized briefly below. Accelerometer data analyses were performed in ActiLife software (version 6.1.3). The raw acceleration data of both the children and their parents were exported to 60 s files with the low frequency extension feature enabled. Subsequently the Sadeh et al. [7] algorithm was applied to classify each 60 s epoch as either “sleep” or “wake.” Sleep periods were then detected and measured by applying the Tudor-Locke et al. [8] algorithm to calculate the sleep duration, quality, and timing metrics. This included total sleep time (TST), sleep efficiency (SE), wake after sleep onset (WASO), sleep onset time, and rise time.

Linear regressions with generalized estimating equations (GEEs) were conducted using children’s sleep metrics as the predictors (i.e. TST, SE, WASO, sleep onset, rise time) and the parent’s sleep metrics as the outcome variables (i.e. TST, SE, WASO, sleep onset, rise time). GEEs were applied to account for multiple children from the same family. The models were stratified by parent self-identified gender and adjusted for child sex, household income, and number of children in the family. Bonferroni adjustments were implemented and an alpha level of 0.05 was considered statistically significant.

Results

There were 53 mother–father–child triads included in this study, which was composed of 43 families. There were 43 cis-gendered mothers, 43 cis-gendered fathers, and 53 children, of which 30 were female and 23 were male. There were 8 sets of 2 siblings, and 1 set of 3 siblings who were from the same family. In this sample, 79% of the mothers, fathers, and children identified as white.

Descriptive information of the study sample, in addition to the linear regression results with GEEs are reported in Table 1. Child TST was a significant positive predictor of father TST (B = 0.187, CI 0.004, 0.370, p = .045), whereas child TST was not significantly associated with mother TST. Neither of the child sleep quality measures (WASO, SE) were found to be significant predictors of the parent (mother or father) sleep quality measures. Child sleep onset was a significant positive predictor of sleep onset time for both mothers (B = 0.299, CI 0.110, 0.487, p = .002) and fathers (B = 0.210, CI 0.017, 0.404, p = .033). Similarly, child rise time was significantly associated with mother rise time (B = 0.340, CI 0.143, 0.537, p = .001) and father rise time (B = 0.289, CI 0.095, 0.483, p = .004).

Table 1.

Descriptive information of the study participants (mean (SD)) as well as the results from the GEE statistical analysis, with child sleep variables predicting mother and father sleep variables. Estimates are presented with 95% confidence intervals (CI), and statistical significance was set at p < .05. The child sleep metric was used as the predictor and the parents sleep metric as the outcome variable, thus a positive beta would include that the parent sleep measure increases as the child’s sleep measure increases

	Children	Mothers	Fathers
Age, years (SD)	3.36 (1.18)	34.90 (4.42)	36.39 (4.67)
Ethnicity
White	42	34	34
Other	11	9	9
Household income (CAD)	<$59 999		10
	$60 000–$79 999		4
	$80 000–$99 999		4
	$100 000–$149 999		15
	>$150 000		9
	Not comfortable answering		1
TST (min) (SD)	565.85 (65.23)	491.16 (84.09)	463.99 (84.20)
SE, % (SD)	94.90 (2.85)	96.18 (2.40)	96.02 (2.74)
WASO (min) (SD)	30.36 (17.59)	19.89 (12.74)	19.60 (14.34)
Sleep onset (h) (SD)	20.85 (1.13)	22.70 (1.35)	22.82 (1.55)
Rise time (h) (SD)	30.96 (1.07)	31.04 (1.13)	30.89 (1.32)
Parameter	B	95% CI	P-value
Mother–child
TST (min)	0.027	−0.154, 0.209	.37
SE (%)	0.044	−0.545, 1.128	.495
WASO (min)	0.005	−0.115, 0.125	.936
Sleep Onset (h)	0.299	0.110, 0.487	.002
Rise Time (h)	0.34	0.143, 0.537	.001
Father–child
TST (min)	0.187	0.004, 0.370	.045
SE (%)	−0.036	−0.206, 1.134	.678
WASO (min)	−0.091	−0.217, 0.035	.158
Sleep onset (h)	0.21	0.017, 0.404	.033
Rise time (h)	0.289	0.095, 0.483	.004

Discussion

Findings from this preliminary analysis indicate that, when objectively measured, parent–child sleep metrics are significantly associated, regardless of the parent’s self-identified gender, contradicting our original hypothesis. In contrast to our findings, several studies employing subjective data collection techniques have reported significant mother–child associations but not father–child associations. For instance, Ford et al. [9] assessed sleep in children with night settling problems, before and after an intervention and found significant improvements in mothers’ sleep duration, while fathers’ sleep parameters remained unchanged. Prior research suggests that within heterosexual couples, mothers are typically the primary overnight carer for their children, yet studies investigating general parenting involvement have suggested that over recent decades, fathers have become increasingly engaged in child rearing [10].

Our findings diverge from previous literature likely due to our use of objective methodology. To date, only two published studies have used accelerometers to assess sleep in children and both parents [2, 11] and these earlier studies focused on older aged children. Matricciani et al. (2019) assessed 11–12-year-old Australian children and their parents’ using accelerometers worn for several days. Aligned with our observations, they also report the strongest agreement between children and both mothers and fathers for sleep onset and rise times. In contrast, in their sample of 163 mother–father–child triads (9–12 years), Kouros and El-Sheikh (2017) reported that child sleep was a significant predictor for mother sleep and not a significant predictor for father sleep. A possible explanation for the distinct findings may be that the age range of our population is much younger and perhaps fathers are more engaged in their child’s sleep at this younger age as compared to when children are older. Moreover, as outlined in the Introduction, another reason our findings may diverge from much of the broader literature is that our device-based measurements eliminate the need for parental proxy reporting, thereby reducing bias.

To our knowledge, this is the first study to objectively assess sleeping relationships between mothers, fathers, and toddler/preschool-aged children. Given that sleep plays a critical role in child development [12], supporting healthy sleep in the early years is vital, particularly as health habits are thought to be established during this period [13]. Understanding family dynamics and how parents can support children’s sleep is therefore essential. Our findings provide preliminary evidence that supports the evaluation of sleep in young children within a familial context—the key take home message for researchers in this area is to emphasize the importance of using objective tools (e.g. accelerometers) to quantify human health behaviors when possible.

Despite the novel findings of this preliminary study there are some limitations. The studied sample was composed of predominantly white families with high household income; thus, findings may not be generalized to racially diverse and lower income families. Additionally, due to the cross-sectional nature of this study, neither causation nor direction of the relationship can be explicitly determined; it remains ambiguous as to the direction of influence between parent and child sleep. Future efforts should continue this line of investigation with a larger sample and assess whether the findings are generalizable to more diverse populations. In addition, longitudinal studies are needed to provide greater insight into causation.

Author contributions

Hannah J. Coyle-Asbil (Conceptualization [lead], Data curation [lead], Methodology [equal], Validation [lead], Writing—original draft [lead], Writing—review & editing [equal]), Bridget Coyle-Asbil (Data curation [equal], Methodology [supporting], Writing—review & editing [supporting]), Julia Gruson-Wood (Conceptualization [supporting], Methodology [supporting], Writing—review & editing [supporting]), David W.L. Ma (Conceptualization [supporting], Methodology [supporting], Project administration [supporting], Resources [lead], Writing—review & editing [equal]), Jess Haines (Conceptualization [equal], Methodology [equal], Resources [lead], Supervision [supporting], Writing—review & editing [equal]), Lori Ann Vallis (Conceptualization [equal], Methodology [equal], Supervision [lead], Writing—original draft [equal], Writing—review & editing [equal])

Disclosure statement

Financial disclosure: The authors would like to acknowledge funding support provided by the Canadian Institutes of Health Research (Project grant #376067) awarded to J.H.; a Canada Foundation for Innovation John R. Evans Leaders Fund grant awarded to D.W.L.M. (Grant #36619), a NSERC PGS-Doctoral grant awarded to H.C-A.

Non-financial disclosure: The authors declare that there are no competing interests.