Maternal Mental Health Symptoms and Clusters Predict Toddler Sleep in Low-Income Homes

Abstract

This study examined the association of maternal mental health symptoms and symptom clusters with the sleeping behaviors of toddlers living in low-income homes. The sample is comprised of 174 racially diverse mother and toddler (aged 15 to 19 months) dyads who were participating in a longitudinal birth cohort study. Dyads were initially recruited from Women, Infant, and Children clinics in a Midwestern county of the U.S. Maternal mental health symptoms were measured with self-reports of depression (Center for Epidemiological Studies – Depression), parenting stress (Parenting Stress Index Very Short Form), and parenting self-efficacy. Toddler sleeping behaviors were measured with an adaptation of the Tayside Children’s Sleep Questionnaire that assessed toddler difficulties in initiating or maintaining sleep (DIMS). Maternal mental health symptom clusters were identified with latent profile analysis and toddler DIMS were regressed on maternal mental health symptoms. We found that 1 SD worse maternal depression, parenting stress, or parenting self-efficacy predicted 0.18-0.24 SD worse toddler DIMS (p < .05). Moreover, we found that the cluster of adverse maternal mental health symptoms predicted nearly 0.75 SD worse toddler DIMS as compared to the cluster with good maternal mental health symptoms (p < .05). These findings provide insight into how maternal mental health symptom clusters may be important in understanding sleep behaviors in early toddlerhood, a sensitive period of sleep development, and underscores the need to support mothers with co-occurring adverse mental health symptoms, as their children may be at higher risk for DIMS.

Poor sleep in childhood accounts for numerous disparities later in life, including adverse health outcomes (Chaput et al., 2017), premature aging (James et al., 2017), poor language abilities (Dionne et al., 2011), and limited academic success (Combs et al., 2016; Sadeh et al., 2015). Notably, the risks from adverse sleep on later health and developmental outcomes are more pronounced for children living in low socioeconomic status (SES) homes (El-Sheikh et al., 2010). Children living in low SES homes may also be at higher risk for sleeping problems (Acebo et al., 2005; Gellis, 2011). Even though the development of early childhood sleep and associated sleeping behaviors is a dynamic process and generally improves over time, toddlerhood is a sensitive period of development when sleep is mainly influenced by environmental factors (Brescianini et al., 2011; Dionne et al., 2011; Fisher et al., 2012). One notable environmental factor within low SES environments that may affect toddler sleep and sleeping behaviors is maternal mental health. The competing environmental demands associated with income constraints in low-SES homes may worsen individual symptoms or overall maternal mental health and, thus, a mother’s ability to manage her toddler’s sleeping behaviors (e.g., Jensen et al., 2017). However, to our knowledge, the relations between maternal mental health symptoms and clusters of symptoms on toddler sleeping behaviors have not been examined for mother-toddler dyads living in low-income homes. In this study, we examine how individual symptoms and symptom clusters of maternal mental health predict toddler sleeping behaviors for mother-toddler dyads living in low-income homes. With this study, we hope to provide a more nuanced understanding of potential malleable targets to support mothers and children living in low-income homes, improve sleep in early childhood, and inform the reduction of sleep-associated health disparities later in life.

Sleep and Sleeping Behaviors in Toddlerhood

While toddler-age sleeping problems comprise numerous concerns on parenting blogs and forums, parents and clinicians rarely discuss sleeping problems during well-visits (Meltzer et al., 2010). This may lead to underestimation of the scope of the problem and lack of professional guidance for parents. Generally, research suggests that sleeping problems are common in early childhood, and while they may improve over time, sleeping problems can persist for about one-third of children (Meltzer et al., 2014). The most prevalent child sleeping problems surround sleeping behaviors – including the child’s resistance to go to bed, delay to sleep initiation, and frequent or prolonged night wakings that require parental assistance (Owens, 2020). These sleeping behaviors are collectively termed disorders or difficulties initiating or maintaining sleep (hereafter DIMS; McGreavey et al., 2005). Toddlers may be at high-risk for developing DIMS, as toddlerhood is a sensitive period of development when children begin to mature into more consistent sleep-wake cycle patterns than infancy (Acebo et al., 2005) and when environmental influences have a more profound effect on their sleep than genetic factors (Brescianini et al., 2011; Dionne et al., 2011; Fisher et al., 2012).

SES and Toddler Sleep

While there are no known studies on the prevalence of DIMS in toddlers living in low-SES homes, other researchers have shown that children living in low-SES environments may be at high risk for poor sleep. For reference, about 25% of toddlers in SES-advantaged homes have a sleep problem (n = 5,006 infants and toddlers; Sadeh et al., 2009). In a national survey, the National Sleep Foundation (n = 1103; 2014) found that parents of children 10 years of age or younger and living with an annual income of $50,000 or less were more likely to report child night waking. Acebo et al. (2005; n = 169) found that toddlers and preschoolers in lower-SES homes (Hollingshead ratings) had more sleep variability, nocturnal wake minutes, and variability in bedtime compared to those in higher-SES homes. Additionally, while Singh and Kenney (2013) did not find significant differences in severely insufficient child sleep based on household income or parental education level, they did find that children (aged 6 to 17 years) living in the most disadvantaged neighborhoods had 37-43% higher odds of not getting “enough sleep” as compared to children living in more advantaged neighborhoods (n = 63,352, data from National Survey of Children’s Health). Singh and Kenney (2013) also acknowledged that their findings may be biased, as response rates from their sample were lower from individuals living in urban areas, with low-income, or of an ethnic minority. In a sample of primarily Black and White mother-preschooler dyads living in urban areas, with nearly half identified as single parents and living in distressed neighborhoods, Williamson and Mindell (2019) found that each increasing socioeconomic risk factor predicted a 10% increase in poor sleep habits. Thus, while there is no known research on SES and the prevalence of overall toddler DIMS, these reviewed studies suggest that children living in socioeconomically disadvantaged, urban homes may be at higher risk for sleep problems as compared to those living with more socioeconomic advantages. In this study, we focus on overall DIMS for toddlers living in low-income homes.

In a low-SES environment, there are few financial and material resources to raise a family. Fewer employment opportunities and low-wage work make it difficult to pay bills (Thompson & Dahling, 2019); access and purchase healthy food (Koury et al., 2020); obtain safe housing (Jensen et al., 2017); and adhere to a regular schedule, such as bedtime routines and sleep practices, due to nonstandard work schedules (Owens, 2020). Managing these stressors in a low-income home may also exacerbate adverse maternal mental health (Roubinov & Boyce, 2017), making it more challenging to learn about or engage in optimal practices that support optimal sleeping behaviors and regular routines in toddlers (Ystrom et al., 2017). Indeed, many studies have shown that women living in low-income homes experience higher rates of depression and stress, and lower feelings of self-competence than their more socioeconomically advantaged peers (Baskind et al., 2019; Bates et al., 2020; Chung et al., 2004; Gaynes et al., 2005). Therefore, it is crucial to examine how maternal mental health factors within low-income homes may be linked to toddler DIMS.

Maternal Mental Health and Toddler Sleep in Low-Income Homes

A large body of literature has linked maternal mental health to various early childhood behaviors, such as self-regulation, social engagement, negative emotionality, and stress reactivity (Bates et al., 2020; Feldman et al., 2009; Goodman & Gotlib, 1999; Granat et al., 2017). However, few studies have examined the relations between maternal mental health and toddler DIMS, particularly for racially diverse mother-toddler dyads living in low-income homes in the U.S. While poor toddler sleep may worsen maternal mental health, ultimately, it is the mother’s mental health that will enable her to manage her toddler’s sleeping behaviors. As research shows that 85% of mothers are the primary caregivers to toddler needs during the night (National Sleep Foundation, 2004), we focus on maternal mental health for this study.

In research and clinical screening tools, assessing maternal mental health often includes measuring her symptoms of depression, stress, and self-efficacy (i.e., feelings of competence in a specific situation – such as parenting or adopting a health-beneficial treatment). Symptoms of depression, stress, and decreased self-efficacy can be common for new mothers, but can be especially prevalent for mothers in low-income homes. For example, the National Institute of Mental Health (2017) shows that the prevalence of depression in adult females is around 8.7%, while Putnick et al. (2020) showed that about 20% of SES-diverse mothers experienced depression at some point within three years of their child’s birth. Yet, in cohorts with mothers living in low-income homes, estimates suggest that between 40-50% of these mothers experience significant depression symptoms (Bates et al., 2020; Chung et al., 2004). One of the most prevalent and sometimes persistent symptoms of depression, stress, and low self-efficacy is fatigue (Cooklin et al., 2012; Demyttenaere et al., 2005; Wiegner et al., 2015). Fatigue is linked to difficulty thinking, irritability, and reduced productivity (Boksem & Tops, 2008; Guo et al., 2018). These symptoms, which often manifest in adverse maternal affect and parenting behaviors, are often theorized as mechanisms through which maternal mental health shapes child outcomes (Goodman & Gotlib, 1999; Miller et al., 2011). In contrast, optimal mental health is often linked to higher energy, better problem solving, improved productivity, more positive affect, and better relationships (Fusar-Poli et al., 2020), each of which represent mental and supportive resources that are needed to manage the sleep behaviors of an active, locomotor-independent toddler. For example, toddlers have an increasing ability to leave their bed to seek maternal soothing (e.g., rocking and nursing) to fall asleep – this may be a challenging sleep behavior for mothers to manage, even more so if mothers are also struggling with fatigue from a symptom barrage of depression, stress, and low self-efficacy. The continuous limit setting needed to regulate this toddler sleep behavior may be difficult for overwhelmed mothers, particularly if they are living in low-income homes and have overall low-resources to understand or seek professional assistance to more effectively manage this toddler behavior.

Thus far, there have been few studies on the relations of maternal mental health, including depression, stress, and self-efficacy, on toddler DIMS in low-income homes. From what is known, findings are mixed. This may be because studies have often examined individual symptoms of maternal mental health and individual DIMS behaviors (e.g., night waking, time in bed, sleep efficiency, or sleep duration) rather than a comprehensive assessment. For example, Pinheiro et al. (2011) found that Brazilian mothers with severe chronic depression symptoms had toddlers (12 months of age, n = 397) at higher risk for a maternal-reported sleep disorder. In a sample of 45 SES-diverse infants and toddlers (1 to 24 months of age) in the U.S, Teti and Crosby (2012) found more maternal depressive symptoms were significantly associated with more child night waking. However, in a Canadian sample of 388 mothers and toddlers (12 months of age), Pennestri et al. (2018) found that maternal depressive symptoms were not associated with maternal-reports of toddler sleep duration. Similarly, in a sample of 21 low-income and minority dyads in the U.S., Caldwell and Redeker (2015) did not find that maternal depression was related to preschooler sleep efficiency, sleep periods, or time in bed. Yet in the same study, higher maternal lifetime stress was associated with less preschooler sleep duration (Caldwell & Redeker, 2015). In another study of 85 toddlers in Canada, maternal parenting stress did not predict toddler sleep duration at 18 or 24 months (Benier et al., 2013). In regards to parenting self-efficacy, one study with 601 dyads found that higher parenting self-efficacy was associated with maternal report of longer preschooler sleep duration (Heerman et al., 2017). Similarly, Teti and Crosby (2012; n = 45) found that maternal self-efficacy (confidence in managing infant and toddler sleep) was associated with less observed preschooler night waking.

While no study has specifically examined relations between maternal mental health and toddler DIMS for dyads in low-income homes, there may be several reasons for the reviewed mixed findings. First, these studies have often measured maternal mental health symptoms individually, while it may be necessary also to examine symptom clusters of maternal mental health on toddler DIMS. Symptom clusters are at least two or more related, co-occurring symptoms (Barsevick, 2016). For many chronic conditions such as adverse mental health, experiencing symptom clusters is associated with lower quality of life than experiencing individual symptoms (Kim et al., 2005), as well as with increased barriers to treatment (Harris & Edlund, 2005). Thus, clusters of maternal mental health symptoms may have more of an effect on toddler DIMS than individual symptoms, as mothers with more overwhelming adverse mental health may have more difficulties managing toddler sleep behaviors, including seeking help for managing these sleep behaviors, than mothers with few symptoms. Additionally, symptom clusters are common in clinical practice, yet little research has examined the synergistic effect in maternal-child mental health research, particularly in regards to child sleep.

The Current Study

Building upon prior research, there were two aims of this study. The first aim was to understand whether and the extent to which individual symptoms of maternal mental health (depression, parenting stress, self-efficacy) were associated with DIMS in toddlers living in low-income homes. We hypothesized that higher depression, more parenting stress, and lower self-efficacy would each predict higher levels of toddler DIMS. The second aim was to determine whether and the extent to which clusters of maternal mental health symptoms were associated with DIMS in toddlers living in low-income homes. We hypothesized that a cluster of adverse maternal mental health symptoms would predict higher levels of toddler DIMS compared to clusters with better maternal mental health symptoms. With this study, we build upon the reviewed prior research in three ways: (1) we examine separate maternal mental health symptoms and symptom clusters in relation to (2) aggregated toddler sleeping behaviors related to DIMS with a validated questionnaire (3) in a sample of mothers and toddlers living in primarily urban, low-income homes.

Materials and Methods

Design, Participants, and Procedure

This study was a cross-sectional, observational design. A subsample of mothers with toddlers from the Kids in Columbus Study (KICS) longitudinal cohort (Salsberry et al., 2016) who completed questions on toddler DIMS and maternal mental health (n = 174) at time point four (TP4; 2016-2017) of data collection constitute the analytic sample of this study. KICS researchers recruited mothers from Women, Infant, and Children clinics (WIC) following a robust quota sampling method to represent those were living at or below the Federal Poverty Level in Franklin County, Ohio, by race and ethnicity. Mothers were eligible to participate in KICS if they met the following enrollment criteria: 1) were pregnant or recently gave birth; 2) were at least 18 years of age; 3) were eligible for WIC clinic services; 4) planned on living in the target county during the 5-year study; 5) their target child did not have a known serious medical disorder at enrollment; and 6) provided consent for themselves and their child. Study approval came from the Institutional Review Board of the authors’ university. After enrollment data collection in WIC clinics, data were primarily collected in mothers’ homes.

The current analytic sample was comprised of participants primarily living in low-income homes: nearly 78% of the dyads had an annual household income of $30,000 or less. For reference, 100% of the Federal Poverty Level in 2017 was $24,600 for a family of four (U.S. Department of Health and Human Services, 2017). About 11% of mothers did not graduate high school. About 56% of mothers had Black toddlers, 47% had White toddlers, 3% had Asian toddlers, and 8% had Latino toddlers (mothers could select more than one race that applied to their toddler). About 54% of toddlers were female.

Measures

This study used data collected when sleep behaviors were first collected at the 1.5-year visit (TP4: child age ~15 to 19 months) - the beginning of toddlerhood. Maternal mental health - including measures of self-efficacy, depression, and parenting stress - and demographics were also measured at TP4. Over 97% of surveys were completed by mothers.

Toddler DIMS

DIMS were measured with a maternal report using an adaptation of the Tayside Children’s Sleep Questionnaire (McGreavey et al., 2005). The Tayside has strong reliability and validity for children 1 to 5 years of age (McGreavey et al., 2005). Mothers rated 10 child sleep behavior items on a 4-point scale, with 0 = strongly agree, 1 = agree, 2 = disagree, and 3 = strongly disagree. Mothers were not asked to rate the items based on a particular time period. Example items included: “the child usually falls asleep within 5–10 minutes after going to bed,” “the child goes to bed reluctantly,” “the child has sleeping difficulties,” and “the child sleeps in the parent’s bed at some time during the night.” Following reverse scoring rules, items were summed. Scores ranged from 0-30, with higher scores indicating less difficulties initiating or maintaining sleep (i.e., higher scores indicate better sleep). Cronbach’s alpha was .86.

Maternal Mental Health Symptoms (Depression, Stress, and Self-Efficacy)

Maternal depression was measured with the Center for Epidemiological Studies-Depression (Radloff, 1977) scale. Twenty items were scored on a symptom frequency scale from 0-3 in the past week, where 0 = rarely or none of the time (less than 1 day) and 3 = most or all of the time (5-7 days). Following reverse scoring rules, items were summed with a possible scoring range of 0-60. Higher scores indicate more depression symptoms. A score of 16 or greater has a sensitivity of 95% and specificity of 70% for major depressive disorder in women living in low-income homes (Thomas et al., 2001). Cronbach’s alpha for the current sample was .84.

Maternal stress was measured with the total stress score from the Parenting Stress Index Short Form 4^th edition (Abidin, 2012). There are 36 items, scored on a scale from 0-4, with 0 = strongly disagree and 4 = strongly agree. Items were summed with a possible range from 0-144. Scores were converted to standard or percentile scores, with higher scores indicating more stress while in a parenting role. For reference, the standard score in the 50^th percentile was 48, and high scores are generally at the 81^st percentile and higher (standard score of 61 and greater). The Cronbach’s alpha for the current sample was .93.

Maternal self-efficacy was measured with the Teti and Gelfand (1991) parenting self-efficacy questionnaire at TP4. There were 10 items scored on a scale from 0 = not good at all to 3 = very good. Sample items included: “When your baby is upset, fussy, or crying, how good are you at soothing him or her?” and “In general, how good a parent do you feel you are with your baby?” Total scores ranged from 0-30, with higher scores indicating better self-efficacy while in a parenting role. In the current sample, the Cronbach’s alpha was .87.

Covariates

Demographics were used as covariates and included maternal education, and toddler sex, race, ethnicity, and age. Maternal education was scored as 0 = not a high school graduate or 1 = high school graduate, as child sleeping behaviors may differ by maternal education (Weimer et al., 2002). Toddler sex was scored as 1 = female and 0 = male to account for potential sex differences in early childhood behaviors, such as DIMS. We controlled for toddler race and ethnicity, as non-Hispanic White children may have better sleep and sleep-related behaviors than other racial and ethnic minorities (Smith et al., 2019). Toddler race was scored as 1 = Black/African American and 0 = not Black/African American, as the majority of those who were not Black in this sample were White. Toddler ethnicity was scored as 1 = Latino and 0 = not Latino.

Analytic Overview

Analyses were performed in SPSS 27 (descriptives, correlations, regressions) and Mplus7 (latent profile analysis [LPA]). Preliminary statistics were examined with descriptives, frequencies, and correlations with all study variables. Aim 1 was examined with regressions. Specifically, we standardized the three maternal mental health variables and toddler DIMS, then completed three sets of linear regressions: one for each maternal mental health symptom (depression, stress, and self-efficacy). Covariates in each regression were mother education and toddler sex, race, ethnicity, and age. Aim 2 was examined with LPA, a person-centered approach that groups characteristics within a particular sample of individuals (Denson & Ing, 2014; Muthen & Muthen, 2000), and regressions. Specifically, we used LPA in MPlus to identify profiles of maternal mental health symptoms. Then, we regressed toddler DIMS on the identified symptom profiles in SPSS, controlling for demographics (mother education and toddler sex, race, ethnicity, and age). Effect sizes for regressions were calculated with Cohen’s d_s (Lakens, 2013) and interpreted with guidelines from Cohen (1988).

For the LPA, we used different starting values (up to 1500) and numbers of profiles (1-7) to find the model that best fit the individuals’ data. Differences in the standardized means between the profiles were determined with one-way ANOVA and post-hoc Tukey tests. We examined 10 criteria to evaluate fit: loglikelihood, Bayesian Information Criteria (BIC), sample-adjusted BIC (ABIC), Akaike Information Criteria (AIC), Vuong-Lo-Mendell Rubin (VLMR) Likelihood Ratio Test, Lo-Mendell-Rubin (LMR) Likelihood Ratio Test, Bootstrapped Likelihood Ratio Test (BLRT), entropy, the number of mothers in each profile and theoretical interpretability. For reference, lower values of BIC, ABIC, and AIC indicate better fit (Kline, 2011). Significant p-values from VLMR, LMR, and BLRT indicate that the model with K profiles fits better than the model with K-1 profiles (Nylund et al., 2007). Higher entropy values (range 0-1) indicate better profilization (Clark & Muthén, 2008). The model should also have a sufficient sample size for each profile to provide power for analyses on associations between profile membership and outcomes (Jung & Wickrama, 2008; Nylund-Gibson & Choi, 2018).

Missing data were as follows: maternal depression (0%), maternal stress (2.3%), maternal self-efficacy (0%), toddler DIMS (1.1%), mother education (2.9%), toddler sex (3.4%), toddler ethnicity (0%), toddler race (0.6%), and toddler age at DIMS measure (0.6%). Little’s MCAR test indicated no sufficient evidence against the hypothesis that the data were missing completely at random (p = .47). Missing data for correlations and regressions were multiply imputed in 20 datasets using fully conditional specification in SPSS 27 using all study variables. For the LPA on maternal mental health symptom clusters, we employed the default maximum likelihood estimation with robust standard errors (MLR), which assumes data are missing at random (Nylund-Gibson & Choi, 2018).

Results

Descriptive statistics of the sample and analytic variables are in Table 1. Spearman correlations between analytic variables are in Table 2. For this sample, the mean standard score of maternal parenting stress was 43.26 (SD = 7.74), around the 60^th percentile. Mean maternal depression was 14.07 (SD = 7.87); about 36.2% of the sample had symptoms of major depressive disorder (score of 16 or higher; Thomas et al., 2001). Mean maternal self-efficacy was 26.31 (SD = 3.54). Mean toddler DIMS was 19.05 (SD = 6.16). All maternal mental health symptoms were significantly associated with toddler DIMS (p < .05): mothers with higher depression (r_s = −.19), higher stress (r_s = −.25), and lower self-efficacy (r_s = .19) had toddlers with higher DIMS.

Table 1.

Descriptive Statistics

	N	%	Min.	Max.	Mean	SD
Maternal Depression	174		1	43	14.07	7.87
Maternal Stress	170		32	72	43.26	7.74
Maternal Self-Efficacy	174		16	30	26.31	3.54
Toddler sleep behaviors	172		0	30	19.05	6.16
Toddler age at TP4 (months)	173		14.1	20.7	17.24	1.78
Mother age at TP4 (years)	174		20	44	28.47	5.32
Mother education	169
8th or less		1.8
Some HS no diploma		8.9
HS diploma or GED		34.9
Some college no degree		37.9
Associate’s degree		5.3
Bachelor’s degree		10.1
Master’s degree		1.2
Household income	172
$10K or less		32.6
$10,0001 to $20K		24.4
$20,0001 to $30K		20.9
$30,0001 to $40K		8.1
$40,0001 to $50K		5.8
$50,001 to $60,000		4.7
$60K or more		3.5
Toddler sex	168
female		53.6
male		46.4
Toddler race and ethnicity
Black	n/173	55.5
Latino	n/174	8.0
White	n/172	46.8
Asian	n/173	2.9

Note. Parents could select more than one race. TP4 = time point four.

Table 2.

Spearman Correlations with Multiply Imputed Data (n = 174)

	1	2	3	4	5	6	7	8	9
1. Maternal depression	--	.40**	−.16*	−.19*	.04	−.19*	.03	.00	.10
2. Maternal stress		--	−.50**	−.25**	−.14	−.15	.06	.00	.02
3. Maternal self-efficacy			--	.19*	.12	.01	.11	−.14	.01
4. Toddler DIMS				--	−.04	.03	−.11	.09	.00
5. Mother education					--	.07	.15	−.24**	−.01
6. Toddler sex (1= Female)						--	.03	.06	−.08
7. Toddler race (1= Black)							--	−.19*	.07
8. Toddler ethnicity (1= Latino)								--	.04
9. Toddler age									--

Note.

^*p < .05;

^**p < .01.

Aim 1: Whether and the Extent to Which Individual Symptoms of Maternal Mental Health were Associated with DIMS in Toddlers Living in Low Income Homes

We found support for our first hypothesis that adverse maternal mental health symptoms independently predicted worse toddler DIMS (Table 3). In the first model, maternal depression significantly predicted toddler DIMS (β = −0.18, SE = 0.08, p = .02, d_s = −0.35), such that 1.0 SD increase in maternal depression predicted nearly −0.2 SD worse toddler DIMS. In the second model, maternal stress significantly predicted toddler DIMS (β = −0.24, SE = 0.08, p < .01, d_s = −0.45), such that mothers who had 1.0 SD increase in stress had a toddler with nearly ¼ SD worse DIMS. In the third model, maternal self-efficacy also significantly predicted toddler DIMS (β = 0.18, SE = 0.08, p = .02, d_s = .35), such that mothers with 1.0 SD higher self-efficacy had a toddler with nearly 0.2 SD better DIMS. In each of the models, none of the covariates significantly predicted toddler DIMS. In sum, 1.0 SD increase in individual adverse maternal mental health symptoms tended to individually predict about 1/5-1/4 SD more toddler DIMS.

Table 3.

Regressions of Individual Maternal Mental Health Symptoms Predicting Child DIMS (n = 174)

		β	SE	t	p	ds
Model 1	Intercept	−0.09	0.79	−0.11	0.91
	Maternal depression [ISPS-CHK-TAGS-ALL]*	−0.18	0.08	−2.30	0.02	−0.35
	Mother education	0.00	0.26	−0.02	0.99
	Child sex- Female	−0.03	0.16	−0.20	0.84
	Child race- Black	−0.10	0.16	−0.66	0.51
	Child ethnicity- Latino	0.12	0.29	0.42	0.67
	Child age	0.01	0.04	0.21	0.84
Model 2	Intercept	0.26	0.79	0.33	0.74
	Maternal stress *	−0.24	0.08	−3.00	0.00	−0.45
	Mother education	−0.17	0.26	−0.65	0.52
	Child sex- Female	−0.05	0.16	−0.31	0.75
	Child race- Black	−0.08	0.16	−0.50	0.62
	Child ethnicity- Latino	0.05	0.29	0.18	0.86
	Child age	0.00	0.04	−0.05	0.96
Model 3	Intercept	0.10	0.79	0.13	0.90
	Maternal self-efficacy *	0.18	0.08	2.31	0.02	0.35
	Mother education	−0.05	0.26	−0.19	0.85
	Child sex- Female	0.01	0.16	0.06	0.96
	Child race- Black	−0.15	0.16	−0.97	0.33
	Child ethnicity- Latino	0.19	0.30	0.65	0.52
	Child age	0.00	0.04	0.01	0.99

Note.

^*= statistically significant result. DIMS and maternal depression, stress, and self-efficacy standardized in SPSS.

Aim 2: Whether and the Extent to Which Profiles of Maternal Mental Health Symptoms were Associated with DIMS in Toddlers Living in Low-Income Homes

We also found support for our second hypothesis that a cluster of adverse maternal mental health symptoms predicted worse toddler DIMS as compared to clusters of better maternal mental health symptoms. Fit indices of the profile selection of maternal mental health symptom clusters are in Table 4, which indicated that the three-profile model was the best fit to the data, theory, and interpretability. Specifically, the three-profile model had adequate numbers of mothers per profile; lower BIC, ABIC, and AIC than the 2-profile model; a significant BLRT p-value, and a higher entropy than the 2-profile model. Shown in Figure 1, the three profiles were distinguished by (1) low self-efficacy (66% of mothers); (2) good mental health (25% of mothers); and (3) adverse mental health (10% of mothers). Shown in Table 5, stress and self-efficacy were significantly different in each profile, and depression differed between the adverse mental health and the other two profiles. Profile membership explained 43% of the variance in depression scores, 39% in stress scores, and 70% in self-efficacy scores.

Table 4.

LPA Fit Indices

# profiles	Profile and n in each	Log-likelihood H0	# free parameters	BIC	ABIC	AIC	VLMR	LMR	BLRT	Entropy
							p value	p value	p value
1	1 = 174	−735.010	6	1500.975	1481.975	1482.020	N/A	N/A	N/A	N/A
2	1 = 49 2 = 125	−695.913	10	1443.416	1411.750	1411.825	0.0099	0.0119	0.0000	0.813
3	1 = 43 2 = 114 3 = 170.838	−675.574	14	1423.375	1379.043	1379.149	0.2115	0.2240	0.0000	0.838
4	1 = 46 2 = 5 3 = 95 4 = 28	−662.587	18	1418.038	1361.038	1361.175	0.2827	0.2940	0.0000	0.834
5	1 = 1 2 = 97 3 = 5 4 = 44 5 = 27	−652.814	22	1419.127	1349.461	1349.628	0.0720	0.0775	0.0000	0.865
6	1 = 33 2 = 19 3 = 21 4 = 95 5 = 5 6 = 1	−642.455	26	1419.045	1336.713	1336.910	0.1953	0.2070	0.0000	0.871
7	1 = 28 2 = 95 3 = 19 4 = 23 5 = 5 6 = 3 7 = 1	−634.958	30	1424.688	1329.689	1329.916	0.0353	0.0393	0.5000	0.882

Note. BIC = Bayesian Information Criteria; ABIC = Adjusted BIC; AIC = Akaike Information Criteria; VLMR = Vuong-Lo-Mendell Rubin; BLRT = Bootstrapped Likelihood Ratio Test. Final fit indices performed with 1500 starts and 120 final stage optimizations.

Figure 1. Standardized Means of the Maternal Mental Health Symptoms for Each Identified Profile.

Note. Higher z-scores mean higher depression, higher stress, and higher self-efficacy symptoms.

Table 5.

Standardized Mean Comparisons of Maternal Mental Health Symptoms between Profiles (n = 174)

	Unimputed means by profile			Standardized means by profile
	Low self-efficacy	Good mental health	Adverse mental health	Low self-efficacy (a)	Good mental health (b)	Adverse mental health (c)	Effect size (η2)
Depression	12.86	12.19	29.71	−0.15c	−0.24c	1.99a,b	0.43
Stress	47.81	39.91	53.47	0.59b,c	−0.43a,c	1.32a,b	0.39
Self-Efficacy	21.42	28.32	25.24	−1.39b,c	0.57a,c	−0.31a,b	0.70

Note. Unimputed means by profile provided for reference. Profiles with standardized means calculated from Mplus; ANOVA performed in SPSS only with standardized means. Next to the standardized mean, alphabetic subscripts indicate a significant difference between two noted profiles at p < .05 with Tukey Honest Significant Difference post hoc tests in SPSS.

The results of toddler DIMS regressed on maternal mental health symptom profiles are in Table 6. Toddler DIMS were nearly 3/4 SD worse in the adverse mental health profile as compared to the good mental health profile (β = −0.74, SE = 0.26, p < .01, d_s = −0.43). Compared to the good mental health profile, the low self-efficacy profile did not significantly predict toddler DIMS (β = −0.24, SE = 0.18, p = 0.18, d_s = −0.20). Finally, compared to the adverse mental health profile, the low self-efficacy profile did not significantly predict toddler DIMS, yet there was a noticeable trend. Despite having low self-efficacy and mild stress, these profiled mothers tended to have children with almost ½ SD better sleep behaviors than mothers with adverse mental health (β = 0.49, SE = 0.29, p = 0.08, d_s = 0.26). No covariates were significant in the models. In sum, mothers with overall adverse mental health had toddlers with worse DIMS as compared to good mental health and a trending comparison to mothers with low self-efficacy.

Table 6.

Regressions on if the Maternal Mental Health Symptom Profiles Predict Standardized Child DIMS (n = 174)

Model	β	SE	t	p	ds
Referent group: Good mental health
Intercept	0.13	0.79	0.17	0.87
Low self-efficacy	−0.24	0.18	−1.35	0.18	−0.20
Adverse mental health *	−0.74	0.26	−2.84	0.00	−0.43
Mother education	−0.03	0.26	−0.12	0.91
Toddler sex- Female	0.01	0.16	0.05	0.96
Toddler race- Black	−0.11	0.16	−0.70	0.48
Toddler ethnicity- Latino	0.17	0.29	0.57	0.57
Toddler age	0.00	0.04	0.10	0.92
Referent group: Adverse mental health
Intercept	−0.61	0.83	−0.73	0.47
Low self-efficacy	0.49	0.29	1.73	0.08	0.26
Good mental health *	0.74	0.26	2.84	0.00	0.43
Mother education	−0.03	0.26	−0.12	0.91
Toddler sex- Female	0.01	0.16	0.05	0.96
Toddler race- Black	−0.11	0.16	−0.70	0.48
Toddler ethnicity- Latino	0.17	0.29	0.57	0.57
Toddler age	0.00	0.04	0.10	0.92

Note.

^*p < .05.

DIMS = Difficulties initiating or maintaining sleep (higher scores mean better sleep).

Discussion

To our knowledge, this is the first study to examine how both individual symptoms and symptom clusters of maternal mental health predict toddler DIMS for dyads living in low-income homes. We found support for both of our hypotheses. First, our results suggested that individual symptoms of adverse maternal mental health were related to higher levels of toddler DIMS. Second, we found that as compared to mothers with good mental health, a symptom cluster of maternal adverse mental health predicted nearly ¾ SD higher toddler DIMS.

Our findings are consistent with prior studies on individual symptoms of adverse maternal mental health on DIMS-related behaviors in SES- and racially-diverse infants and toddlers. For example, our findings that increased depression symptoms are linked to worse toddler DIMS are in line with findings from Pinheiro et al. (2011) on Brazilian infants (worse maternal rated sleep disorders) and Teti and Crosby (2012) on an SES-diverse U.S. cohort of infants and toddlers (more observed infant night wakings). Our findings on associations between maternal stress and toddler DIMS are also in line with findings by Caldwell and Redeker (2015), who found higher maternal lifetime stress was associated with a worse DIMS-related behavior (less sleep duration) in minority preschoolers from low-income homes. Our findings regarding associations between better parenting self-efficacy and less DIMS are in line with findings by Heerman et al. (2017) who showed that higher parenting self-efficacy was associated with longer preschooler sleep duration. Therefore, this study builds upon and adds support to the existing literature on individual maternal mental health symptoms such as depression, parenting stress, and self-efficacy and more comprehensive toddler DIMS for a low-income and racially diverse sample that may be at high-risk for child sleeping problems.

In this study, we uniquely found that a cluster of adverse maternal mental health symptoms predicted worse toddler DIMS than the best cluster of maternal mental health symptoms. We also found a trending relationship that mothers with a mix of average and adverse maternal mental health symptoms (low self-efficacy profile) had toddlers with better DIMS as compared to the mothers with overall adverse mental health. Symptom clusters have not been well examined in maternal mental health, particularly in relation to early childhood sleep. This finding aligns with previous research showing that symptom clusters with particular conditions are associated with lower quality of life than experiencing individual symptoms (Kim et al., 2005). Indeed, good sleeping quality can influence overall quality of life, and poor sleep behaviors interrupt the sleep of both toddlers and mothers. Based on our findings, it seems that being depressed, stressed, and having low self-efficacy is related to high toddler DIMS. This particular cluster of mental health symptoms may be related to child DIMS for several reasons. First, when mothers are depressed and stressed, they may feel overwhelmed and exhausted by the demands of parenthood. Coupled with low self-efficacy, or feelings of being inadequate or unable to parent well, these mothers may have difficulty providing the optimal environment for healthy toddler sleep even if they may have the capability. Instead, mothers with several adverse mental health symptoms may have different priorities surrounding child sleeping behaviors than mothers with optimal mental health, and one priority may simply be getting through another night. It may be difficult for mothers with adverse mental health to adhere to routines, particularly with toddlers who are often discovering and testing limits of their locomotor independence. For instance, healthy sleep habits for toddlers include consistent routines and limit setting to ensure coordination between potential daytime naps and nighttime sleep. Managing these consistent routines may also be particularly difficult in low-income homes with competing priorities and nonstandard work schedules. Toddler sleep and sleeping behaviors may also be complicated by varying childcare schedules, painful teething, or difficulties with self-soothing. Thus, for mothers with several symptoms of adverse mental health who are also managing the demands of toddlerhood and challenges with living in low-income homes, their goal may be to not optimize toddler sleep, rather help the toddler at least get some sleep and survive the day.

These findings strengthen the argument for providing mothers with adequate mental health resources. In our sample, 36.2% of mothers may have had a major depressive disorder compared to about 8.7% of adult women nationally (National Institute of Mental Health, 2017). It may be difficult for mothers living in low-income homes in the U.S. to get adequate health care coverage, as health care insurance in the U.S. is most often employer-based and with high deductibles. Thus, at minimum, mothers living in low-income homes may not be getting the professional mental health care they need, and opportunities are also lost to help these mothers manage their child’s sleep behaviors. Without this kind of support, a cascading cycle can occur of poor child sleep exacerbating adverse maternal mental health, in turn influencing worse child sleep. Thus to care for young children, comprehensive care must be provided to the mothers.

Limitations and Future Research

As an observational, correlational study of cross-sectional data, there are limitations to this work. First, no causality can be inferred from this work: the associations we found may still be explained by unmeasured variables. Second, while we used a reliable and valid measure of toddler DIMS, this study was based on maternal report, and we did not confirm the findings with observational measures of toddler sleep (e.g., actigraphy, videography, or home-based polysomnography). There is ongoing debate on the accuracy versus bias of maternal report measures, particularly when mothers may be depressed (Gartstein et al., 2009; Ordway, 2011; Richters, 1992) or have anxiety (Briggs-Gowan et al., 1996). For example, significant symptoms of depression may bias how mothers recall their child’s sleep behaviors, or it may reflect the sensitivity or true difficulty mothers perceive they are having with their child’s sleep while managing their own depression. Nevertheless, while some studies show that mothers may slightly overestimate child sleep quantity (Combs et al., 2019), other studies show that maternal report of child sleep is valid (Sadeh, 2008). Further, maternal perception of child behavior can help researchers understand a more nuanced view of the child’s environment that strict laboratory measures may not be able to capture. Maternal perception has also been shown to influence child behavior (Bates & Bayles, 1984; Mangelsdorf et al., 2000). Thus, while a limitation of this study was that we did not have observational measures of child sleep, we provide insight into understanding that mothers in low-income homes with adverse mental health may be struggling more with child sleep behaviors as compared to mothers with better mental health. This is important because clinical assessments of child sleep often begin with maternal report of the child’s behavior. This first screening opportunity with an expert clinician opens up an important conversation to help mothers: a) feel their concerns are respected and heard; b) hear and reflect upon typical child sleep behaviors; and c) learn to manage safe and optimal child sleep behaviors. One final limitation was that this study was focused on the relations between mothers specifically and their children in low-income homes in the U.S. While over 85% of caregivers who respond to child sleep demands during the night are primarily mothers, 15% of caregivers are allomaternal. Given the changing diversity of family structures and responsibility distributions in the U.S., it will be important for future work to investigate the effects of allomaternal care and allomaternal mental health on child DIMS.

In conclusion, we found that both individual and clusters of adverse maternal mental health symptoms predicted worse toddler sleep for dyads living in low-income homes. These findings provide a nuanced insight into how maternal mental health symptom clusters may be necessary for understanding sleep behaviors in early toddlerhood, a sensitive period of sleep development. These findings also underscore the need to support mothers with co-occurring adverse mental health symptoms.

Biographies

Author Bio

Dr. Randi Bates is an Assistant Professor of Nursing at the University of Cincinnati.

Dr. Britt Singletary is a Postdoctoral Researcher at the Crane Center for Early Childhood Research and Policy at The Ohio State University.

Dr. Jaclyn Dynia is a Senior Research Specialist at the Crane Center for Early Childhood Research and Policy at The Ohio State University.

Dr. Laura Justice is a Distinguished Professor at the College of Education and Human Ecology and Executive Director of the Crane Center for Early Childhood Research and Policy at The Ohio State University.