A multimethod evaluation of bed provision and sleep education for young children and their families living in poverty

Abstract

Study Objectives:

We evaluated the impact of bed provision and sleep education through the Beds for Kids (BfK) program on early childhood sleep and behavior and maternal mood and sleep.

Methods:

Twenty-seven mother–child dyads (child_age= 2–5 years, 85.2% Black) living in poverty and without an individual child bed were randomly assigned (multimethod randomized waitlist control trial design) to BfK intervention ∼1 week postbaseline (initial intervention) or ∼2 weeks postbaseline (waitlist control), with follow-up at 1 month. BfK intervention (home bed delivery and written sleep health education) was provided to all families. Children wore actigraphs and mothers completed daily diaries to assess child and maternal sleep, child behavior, and maternal mood for the initial 1-week comparison period. Maternal-reported child sleep and behavior (internalizing and externalizing problems) were collected at 1 month after BfK participation for all families; 11 families completed a qualitative interview at 1-month follow-up.

Results:

At 1 week after BfK, mothers’ sleep duration increased by 1 hour compared to that of waitlist controls. No changes were found in child sleep, child behavior, or maternal mood. However, at 1 month after BfK intervention, improvements were found in mother-reported child night awakenings, sleep quality, sleep problems, and behavior. Mothers qualitatively reported significant BfK benefits for child sleep and family well-being, although they noted challenges to transitioning young children to sleeping independently.

Conclusions:

Bed provision and sleep education for families living in poverty has an immediate impact on maternal sleep and reported well-being. Child sleep and behavioral improvements are seen by 1 month, with children experiencing an initial adjustment period to sleeping independently.

Clinical Trial Registration: Registry: ClinicalTrials.gov; Name: Impact of Beds for Kids Program on Child Sleep; URL: https://www.clinicaltrials.gov/ct2/show/NCT03392844; Identifier: NCT03392844.

Citation:

Williamson AA, Min J, Fay K, Cicalese O, Meltzer LJ, Mindell JA. A multimethod evaluation of bed provision and sleep education for young children and their families living in poverty. J Clin Sleep Med. 2023;19(9):1583–1594.

BRIEF SUMMARY

Current Knowledge/Study Rationale: Socioeconomic sleep health disparities begin in early childhood and can adversely affect development, underscoring the need for efforts to promote sleep health. To date, no studies have evaluated the effect of sleep health education and/or bed provision on factors beyond caregiver sleep health knowledge and parent-reported child sleep.

Study Impact: Bed provision and sleep education for families living in poverty has a more immediate impact on maternal sleep duration, by about 1 hour, and perceived family well-being, whereas improvements in young children’s sleep and behavior emerge after 1 month, although some experience an adjustment period to sleeping independently. Collectively, results suggest broad improvements in family well-being following bed provision and sleep education.

INTRODUCTION

Early childhood is a critical period for sleep health promotion. Multiple aspects of poor sleep health¹ in early childhood, including short sleep duration,^2,3 later and more variable sleep timing,^4,5 frequent night awakenings,⁶ and an inconsistent bedtime routine,⁷ have been associated with impairments in the social-emotional and neurobehavioral skills that rapidly develop during this period. These linkages may be more pronounced in the context of socioeconomic disadvantage.⁸ For example, caregiver-reported poor sleep quality in preschoolers was associated with worse working memory, but only among children living in lower-socioeconomic-status homes.⁹ Another early childhood study found that greater exposure to cumulative sociodemographic risks, including lower family and neighborhood income, amplified associations between caregiver-reported child sleep problems and psychosocial concerns.¹⁰

Given this research and well-documented sleep health disparities,^11,12 effective sleep health promotion is needed for young children and families who are marginalized due to socioeconomic disadvantage. To date, the few available interventions have focused primarily on sleep health education. In a Head Start program for families of lower-socioeconomic-status backgrounds, a one-time sleep health education session for caregivers plus a teacher-delivered 2-week sleep health preschool curriculum increased caregiver sleep knowledge.¹³ This improvement was not maintained at 1-month follow-up, although there were durable improvements in caregiver-reported child weeknight sleep duration.¹³ A nonrandomized pilot study of family sleep health education for children aged 4–8 years and their caregivers living in supportive housing found evidence of feasibility, acceptability, and improvements in caregiver-reported child sleep.¹⁴

The benefits of sleep health education may be augmented by other modifiable determinants of socioeconomic sleep health disparities.¹² For instance, families experiencing socioeconomic disadvantage may lack tangible sleep resources, such as a dedicated sleep space for children, beds, and bedding. Beds for Kids (BfK) is a Philadelphia-area nonprofit program that provides beds, bedding, and other bedtime materials (age-appropriate books, stuffed animals, and dental hygiene tools) to children who are living in poverty and without an individual bed (eg, sleeping on the floor, on a sofa, or in a shared bed). We partnered with BfK to evaluate the effect of adding caregiver-based sleep health education to their bed provision program in 152 children (aged 2–12 years). Simply receiving a bed was associated with reduced electronics use and increased nighttime sleep duration (20 minutes) according to caregiver reports for all children; however, the addition of caregiver sleep health education further reduced bedroom electronics use and increased sleep duration (by 10 additional minutes).¹⁵ Sleep health education is now a permanent part of BfK.

To date, no studies have evaluated the impact of sleep health education and/or bed provision on factors beyond caregiver sleep health knowledge and parent-reported child sleep. Thus, the current multimethod study evaluated the impact of bed provision plus sleep education on early childhood sleep and behavior, as well as caregiver mood and sleep, given research on linkages between poor child sleep and increased caregiver mood and sleep difficulties.^16,17 This was a 1-week, randomized, waitlist control pilot trial, with all families receiving BfK intervention following staggered baselines and assessed at 1-month follow-up. Study aims were to (1) examine initial 1-week changes in child sleep health, child behavior, caregiver sleep duration, and caregiver mood after BfK intervention relative to a waitlist condition; (2) examine changes in child sleep health and behavior from baseline to 1-month follow-up across all families; and (3) qualitatively identify caregiver perceptions about BfK participation and related family sleep outcomes.

METHODS

Setting and design

Study data were collected between 2017 and 2018 through a partnership between BfK and Children’s Hospital of Philadelphia (CHOP). BfK is part of the nonprofit organization One House at a Time and serves families referred from local social service agencies in the greater Philadelphia region. To participate in BfK, children must be between the ages of 2 and 20 years, live without an individual bed, and live in a home at or below 100% of the US federal poverty threshold. For each child in a participating family, BfK delivers a new twin mattress, a bedframe, and a “bedtime bag,” which contains a sheet set, blanket, pillow, age-appropriate books, a stuffed animal, a toothbrush, and toothpaste. Families also receive information about healthy sleep habits (importance of an early bedtime and guidance on avoiding caffeine and electronics in the bedroom; see Figure 1) via a magnet and bookmark.

Figure 1. Study design, measurement schedule, and Beds for Kids (BfK) program content.

BCSQ = Brief Child Sleep Questionnaire, CBCL = Child Behavioral Checklist, PANAS = Positive and Negative Affect Scale.

For the current study we recruited children aged 2–5 years with an English-speaking caregiver/legal guardian (100% mothers) referred to the BfK program by social workers located at 3 city-based primary care practices affiliated with CHOP. Families were excluded if the identified child or mother had a chronic medical, neurodevelopmental, or behavioral condition that would affect sleep; had a previously diagnosed sleep disorder; and/or used any sleep-promoting medications (eg, clonidine, Benadryl, melatonin). All children within the home (including all siblings) received BfK within ∼2 weeks of their program referral. Any family referred to BfK received all BfK items, regardless of study participation.

Following referral, eligibility screening, and baseline assessment, families were randomly assigned to receive the BfK delivery after 7 days (initial intervention; n = 14) or after 14 days (waitlist control; n = 13); see Figure 1 and Figure 2. All children wore an actigraph and mothers completed telephone-based daily diaries during this 2-week period. At the end of the second week of baseline, all waitlist control families were provided all aspects of the BfK intervention. One month after BfK participation, families completed a follow-up assessment (Figure 1). Twenty families were invited to complete the optional qualitative interview at the 1-month follow-up; 11 families completed interviews. Families who declined cited a lack of time for interview completion. Compensation was provided for measure completion. Mothers provided informed consent and this study was approved by the CHOP Institutional Review Board. This study was registered at ClinicalTrials.gov (NCT03392844).

Figure 2. CONSORT diagram with study procedures.

Measures

Mothers completed baseline measures, daily diaries, and 1-month follow-up measures. Measures collected at each time point are presented in Figure 1. Measurement decisions were made balancing the aims of the study against participant burden.

Trained research assistants blinded to study condition administered daily diaries via telephone between 4:00 and 8:00 pm. These diaries included questions about nightly child and maternal sleep, daily child behavior challenges, and daily maternal mood.

Mother-reported child sleep

The 33-item Brief Child Sleep Questionnaire (BCSQ) was used to assess child sleep outcomes across study Aims 1 and 2. The BCSQ is appropriate for 2- to 5-year-olds and has shown good reliability and moderate correspondence with actigraphic child sleep recordings.¹⁸

For Aim 1, selected items were used for the daily diary assessment, administered as part of standard procedures for actigraphy assessment of child sleep, and to capture aspects of child sleep that are not as reliably measured by actigraphy, such as night awakening frequency and duration (see below).^19,20 Aim 1 outcomes were assessed via daily diaries using BCSQ items, which included mother-reported child sleep onset latency, number and duration of night awakenings, and mother-perceived sleep quality. Sleep onset latency and duration of night awakenings were estimated in minutes. For sleep quality, mothers were asked how well their child slept on a 6-point Likert scale ranging from “very poorly” to “very well,” with higher values reflecting better sleep quality.

The full 33-item BCSQ was also used to assess multiple dimensions of child sleep at baseline and follow-up (see Figure 1) for study. These outcomes included mother-reported child nighttime sleep duration, bedtime, sleep onset latency, number and duration of night awakenings, sleep quality, and perceived sleep problems. BCSQ items were also used to describe the child’s sleep environment (eg, sleep location and space) and bedtime routine.

Actigraphy

Children were given an Actiwatch Spectrum (Philips Respironics, Inc., Bend, Oregon) to be worn continuously on the wrist (except when swimming or bathing). For this study, 7 days/nights of data (7 days after bed delivery for the initial intervention group and 7 days of baseline for the waitlist control group) were examined. Actigraphy was not used at follow-up to minimize participant burden and maximize study feasibility. Mothers were asked to press an event marker at lights-off and lights-on and report on sleep–wake patterns via the daily diaries, consistent with recommended procedures for actigraphy assessment.^19,20

Due to significant discrepancies between the actigraphy and diary data,²¹ the “nighttime sleep period” (scored as actigraphic sleep onset to sleep offset) was used instead of the traditional “sleep opportunity” (reported bedtime to wake time).^19,20 Sleep onset was defined as the first 5 consecutive minutes of sleep following caregiver-reported lights out time. Sleep offset was defined as the last 5 consecutive minutes of sleep prior to the diary-reported awakening time. Nighttime sleep duration was calculated as the average number of sleep minutes during the nighttime sleep period across the 7-day intervention period (ie, 7 days after bed delivery for the initial intervention group and 7 days of baseline for the waitlist control group; Figure 1). Sleep variability was calculated as the standard deviation of average nighttime sleep duration per individual.²² Actigraphy-based measurement of nighttime awakenings and sleep efficiency were not included because past studies indicate poor specificity for detecting wake after sleep onset in young children, leading to an overestimation of sleep fragmentation/awakenings and an underestimation of sleep efficiency.^20,23,24

Child behavior challenges

The 99-item mother-reported Child Behavior Checklist for children ages 1.5–5 years was used to assess internalizing and externalizing behaviors at baseline and 1-month follow-up. Items are rated on a 3-point Likert scale (“not true,” “somewhat true,” “very true”) with raw scores converted into T-scores for analysis. The Child Behavior Checklist has shown strong reliability and validity in large-scale studies and is a widely used measure of child behavior.²⁵ Aim 2 outcomes were the standardized internalizing behaviors and externalizing behaviors scales. Higher scores reflect more challenging child behaviors.

Five items were adapted from the Brief Problem Monitor (a validated, shortened version of the Child Behavior Checklist) to assess daily child behavioral challenges via the daily diaries.^25,26 Mothers used a 3-point Likert scale (“not true,” “somewhat true,” “very true”) to rate the following child behaviors that day: defiance, temper tantrums, sudden changes in mood/feelings, unhappiness/sadness, and worries. The average daily score was used, with higher scores reflecting more behavioral challenges.

Daily maternal mood

Three items from the validated Positive and Negative Affect Scale²⁷ asked mothers to rate their feelings of upset, stress, and happiness (reverse-coded) on a 5-point Likert scale ranging from “not at all/very slightly” to “extremely.” Questions were asked as part of the daily diaries, and an average daily score was generated to reflect mood.

Maternal nighttime sleep duration

Maternal nighttime sleep duration was assessed in hours through 1 self-reported daily diary item from the reliable and valid consensus sleep diary.²⁸

Family sociodemographic information

Child age, sex, race, and ethnicity were abstracted from CHOP’s electronic medical records. At baseline, mothers reported on their age, sex, race, ethnicity, education, income, employment, marital status, and number of adults and children in the home.

Covariates

In addition to child age and sex, we included baseline maternal clinically significant depressive symptoms, measured via the Center for Epidemiological Studies Depression Scale-Revised 10-item short form (≥ 10 = clinically significant)²⁹ given its potential impact on mothers’ own sleep and perceived child behavior and sleep.^16,17 We also included baseline family chaos as a covariate using the Confusion, Hubbub, and Order Scale,³⁰ because family organization and routines have been associated with early childhood sleep.³¹

Post-BfK qualitative interview

We aligned qualitative interview procedures and analysis with COREQ guidelines.³² Two study team members (A.A.W. and O.C.) administered in-person, 30-minute optional qualitative interviews that were audio-recorded. No field notes were recorded. Open-ended interview questions (supplemental material) solicited mothers’ perspectives about their participation in the BfK program, including their views about the written sleep health education and any program benefits on family sleep.

Analytic approach

All analyses were conducted in Stata (StataCorp LLC, College Station, Texas). Baseline sociodemographics were compared between children who received a bed at day 7 and those in the waitlist control condition (bed at day 14) using t tests and chi-squared tests. To test whether BfK participation improved actigraphy-derived child sleep and daily diary-derived mother-reported child sleep (using BCSQ items), behavior, and maternal mood and sleep duration (Aim 1), we used linear mixed-effects models that included random intercepts and adjusted for child age, sex, and baseline maternal depressive symptoms and family chaos. To determine whether BfK improved maternally reported child sleep health (full BCSQ measure) and behavior (Child Behavior Checklist) from baseline to 1-month post-BfK participation (Aim 2) we again used mixed-effects models adjusting for the same covariates in Aim 1, as well as study condition to account for variation in BfK participation timing. Logistic regression models were used for dichotomized outcomes; a McNemar test was used for the sleep problem outcome due to limited data coverage for covariates in the logistic regression model.

A total of 4 children did not have actigraphy data due to device errors or device loss. For the remaining 23 children, we included all possible nights of data (mean number of nights = 5.8, range = 1–7 nights), except in the case of outliers on specific variables. We excluded outliers in actigraphy-derived nighttime sleep duration (> 15 hours or < 5 hours), mother-rated child sleep onset latency in daily diaries (> mean ± 3 standard deviation), and pre- and post-BfK mother-reported child sleep onset latency (> 200 minutes), night awakening duration (> 120 minutes), and nighttime sleep duration (< 3 hours). In all quantitative analyses we assumed missingness was completely at random. Thus, missing data were handled using full information maximum likelihood estimation, which is appropriate for data missing completely at random.

Sample size for the qualitative interviews was based on data saturation and guidance from prior research.³³ Qualitative data analysis followed a grounded theory approach.³⁴ We created an operational definition for each code and decision rules for their application. One team member (O.C.) and a research assistant first separately coded 2 transcripts, compared their coding, and developed an initial codebook with oversight from another team member (A.A.W.). The codebook was then applied to 2 additional transcripts, compared across coders, and further refined. Coding disagreements were resolved through discussion. The finalized codebook was then applied to all transcripts; 100% of the transcripts were double-coded for reliability purposes (kappa = 0.80). Themes were determined by consensus among the research team.

RESULTS

Preliminary analyses

Table 1 shows sociodemographic information for mother–child dyads across both study conditions (n = 27). No significant differences were found at baseline between dyads assigned to the intervention condition vs those assigned to the waitlist control condition for child age (mean difference [standard error] = −0.27 [0.46], P = .55), child sex (male: 46.2% vs 35.7%, P = .58), clinically significant maternal depressive symptoms (30.8% vs 28.6%, P = .91), and family chaos (mean difference [standard error] = −0.82 [2.27], P = .72).

Table 1.

Participants’ sociodemographic information (n = 27 mother–child dyads).

	Child	Mother/Family
Age, years	3.30 (1.17)	40.7% ages 18–29 years; 59.3% ages ≥ 30 years
Female sex	59.3%	100%
Race
Black/African American	85.2%	77.8%
Other or multiple races	7.4%	7.4%
White	3.7%	3.7%
Asian	—	7.4%
American Indian/Alaska Native	—	3.7%
Hispanic/Latinx ethnicity	22.2%	7.4%
Mother’s occupational status
Employed full-time		44.4%
Employed part-time		22.2%
Homemaker/at-home parent/on maternity leave		14.8%
Student		7.4%
Unemployed/between jobs		11.1%
Mother’s marital status
Single		77.8%
Married		7.4%
In a relationship, living with partner		3.7%
In a relationship, living separately		11.1%
Mother’s highest level of education
High school		48.1%
Diploma/preuniversity/junior college		25.9%
≥ College/university		45.9%
No. of children living in the home		2.96 (.90)
No. of adults (≥ age 18 years) living in the home		1.78 (1.12)
Clinically significant maternal depressive symptoms		29.6%
Average household chaos		10.89 (5.81)

Values are mean (standard deviation) or percent. Race and ethnicity are included as sociopolitical constructs, not as biological indicators.

Aim 1: outcomes 1 week after BfK

Table 2 shows the results of mixed-effects models with adjusted mean child and mother outcomes in the intervention condition compared to the waitlist control group. For the intervention group, there was no evidence of significant improvements in actigraphy-derived child nighttime sleep duration or variability compared to children in the waitlist control condition who had not yet participated in BfK. Similarly, no significant differences between the intervention and waitlist control groups emerged for improvements in mother-reported nightly child sleep onset latency, number and duration of night awakenings, or sleep quality.

Table 2.

Linear mixed-effects models for nightly/daily Beds for Kids intervention outcomes in the 1 week following bed provision.

Intervention Outcomes	Waitlist Control Predicted	Intervention Predicted	Adjusted Mean Difference [Intervention − Control] (SE)	P
	Means (SE)	Means (SE)
Child measures
Mother-reported child sleep onset latency (minutes)	33.70 (7.25)	39.98 (9.36)	6.28 (11.92)	.598
Mother-reported child night awakening frequency	0.66 (0.11)	0.40 (0.07)	−0.26 (0.14)	.066
Mother-reported child sleep quality	2.00 (0.17)	1.92 (0.14)	−0.08 (0.22)	.712
Mother-reported daily child behavior challenges	0.33 (0.09)	0.27 (0.07)	−0.06 (0.11)	.567
Actigraphy-derived nighttime sleep duration (hours)	9.03 (0.11)	8.53 (0.15)	1.67 (1.10)	.130
Actigraphy-derived nighttime sleep duration variability, mean (SD)a	1.35 (0.40)	1.56 (0.82)	0.12 (0.30)	.699
Maternal measures
Mother self-reported mood	0.99 (0.13)	0.96 (0.11)	−0.03 (0.17)	.863
Mother self-reported nighttime sleep duration (hours)	6.25 (0.26)	7.35 (0.28)	1.10 (0.37)	.003

Linear mixed-effects models accounting for random effects of children and after adjusting for child sex, age, and baseline mother clinically significant depression, and baseline family chaos score. ^aRegression model for nighttime sleep duration variability. SD = standard deviation, SE = standard error.

Mother-reported daily child behavior challenges and daily maternal mood also showed no significant differences between dyads in the intervention vs waitlist control condition (Table 2). However, in the 1 week following child bed provision, mothers reported significantly longer nighttime sleep duration compared to those who had not yet participated in BfK, with mothers in the intervention condition reporting an extra 1.1 hours of sleep.

Aim 2: outcomes 1 month after BfK

Table 3 summarizes child sleep arrangement, sleep patterns, and changes in outcomes from baseline to 1 month after BfK participation for all participants. At baseline, most (77.8%) children were sleeping in a shared bed with a caregiver or sibling(s); some slept on the floor (3.7%) or in another sleep space, such as in different family members’ beds or in a portable crib (7.4%). One month following BfK participation, nearly all children (87.9%) were sleeping in their own bed, in a room shared with a caregiver (22.2%) or sibling(s) (25.9%), or in their own room (40.7%). Three children (11.1%) continued to share a bed with a caregiver at follow-up.

Table 3.

Mother-reported child sleep patterns, ecology, and intervention outcomes before and after Beds for Kids (BfK) intervention (n = 27).

Sleep Arrangement and Patterns (BCSQ)	Pre-BfK	1-Month Post-BfK	Adjusted Post- vs Preintervention Difference	P
	Mean (SD) or %	Mean (SD) or %	Beta Coefficient, (SE)a
Sleeps in own room, own bed	0.0%	40.7%
Shared room with caregiver, own bedb	11.1%	22.2%
Shared room with sibling(s), own bed	0.0%	25.9%
Shared room with caregiver, shared bed	59.3%	11.1%
Shared room with sibling(s), shared bed	18.5%	0.0%
Floor	3.7%	0.0%
Other	7.4%	0.0%
Bedtime	8:44 PM (50.40 min)	8:59 PM (48.60 min)	.24 (0.13)	.053
Sleep onset latency (minutes)	42.61 (33.20)	37.71 (29.12)	−4.22 (5.96)	.479
Night awakening frequency	1 (0.78)	0.62 (0.64)	−0.39 (0.14)**	.007
Duration of night awakenings (minutes)	20.32 (23.52)	10.35 (13.65)	−10.31 (4.73)*	.029
Wake time	7:44 AM (74.40 min)	7:39 AM (73.80 min)
Nighttime sleep duration (hours)	8.91 (1.29)	9.13 (1.48)	0.20 (0.29)	.482
Takes naps	88.89%	81.48%
Nap duration (minutes)	104.44 (63.45 min)	76.67 (60.76 min)
Total (24-hour) sleep duration (hours)	10.72 (1.88)	10.46 (2.04)
Sleep quality	2.48 (0.98)	3.38 (.85)	.90 (.20)***	<.001
Child sleep problem	40.7%	11.5%	N/A	.03c
Bedtime routine ≥ 5 nights/week	48.2%	57.7%	0.66 (0.71)	.353
Consumes ≥ 1 caffeinated beverages/d (yes/no)	48.2%	57.7%	0.57 (0.73)	.432
No. of caffeinated beverages/day	0.81 (1.00)	0.65 (0.63)	−0.17 (0.19)	.354
Electronics in bedroom (yes/no)	88.9%	76.9%	−1.49 (1.13)	.185
No. of electronics	1.74 (0.98)	1.50 (1.21)	−0.26 (0.21)	.229
Internalizing problems T-score	49.96 (10.38)	46.23 (10.76)	−3.84 (1.56)*	.014
Externalizing problems T-score	52.15 (12.06)	48.62 (12.26)	−3.56 (1.65)*	.031

*P < .05, **P < .01, ***P < .001. ^aMixed-effects models accounting for random effects of children and after adjusting for child sex, age, and baseline maternal clinically significant depression, baseline family chaos score, and intervention, except for child sleep problem. ^bAt baseline assessment, “own bed” in caregiver’s room reflects independent child sleep space (eg, a rug or small crib mattress on the floor) and not an individual child bed. ^cMcNemar test. BCSQ = Brief Child Sleep Questionnaire, N/A = not applicable, SD = standard deviation, SE = standard error.

The number and duration of child night awakenings both significantly reduced from baseline to follow-up. Mother-reported child sleep quality and overall sleep problems were both significantly improved at 1 month after BfK participation compared to baseline. Specifically, the proportion of children with mother-reported sleep problems reduced from 40.7% at baseline to 11.5% at follow-up. However, there were no changes in mother-reported child bedtime, sleep onset latency, or nighttime sleep duration at 1-month follow-up. Both mother-rated child internalizing and externalizing concerns were significantly lower at the 1-month follow-up compared to baseline.

Post hoc and sensitivity analyses

We generated estimates of intervention effect sizes using Cohen’s d for continuous outcomes and Cohen’s h for categorical outcomes.³⁵ Moderate intervention effects were found for the improvement in maternal sleep duration (d = 0.56) in the 1 week following BfK. Across the sample at 1-month follow-up there were large improvements in child sleep quality (d = 0.99), night awakening frequency (d = 0.54) and duration (d = 0.52), and child sleep problems (h = 0.69). Small improvements were found for child internalizing (d = 0.35) and externalizing (d = 0.29) problems.

We also conducted a post hoc analysis to examine whether there were any subgroups of children whose mother-reported nighttime sleep duration improved by 20 minutes or more, stayed the same, or got worse by 20 minutes or more. Twenty minutes was selected as a cutoff to align with prior BfK research.¹⁵ Among 22 children having pre- and postnighttime sleep duration, 31.8% (n = 7) improved by 20 minutes or more, 22.7% (n = 5) got worse by 20 minutes or more, and 45.5% (n = 10) stayed the same from before to after BfK.

We additionally conducted a post hoc power analysis based on Aim 1 results, which showed that maternally reported child night awakenings in sleep diaries were marginally (P = .066) reduced in the intervention vs waitlist control groups. The post hoc power analysis using mixed models based on Monte Carlo simulations indicated that a sample of 54 children (n = 27 per condition) would yield 82% power to detect a significant reduction in this outcome. Thus, our study was likely underpowered. We then conducted sensitivity analyses by removing covariates from the models for Aims 1 and 2 to conserve power. Table S1 ^{(128KB, pdf)} and Table S2 ^{(128KB, pdf)} show that the pattern of significance is the same for all but one outcome (night awakening duration from before to after intervention), even with covariates removed.

Aim 3: caregiver perceptions about BfK participation

Themes from qualitative data and representative participant quotations are shown in Table 4. Mothers reported a number of benefits to BfK participation, including benefits to child sleep as well as benefits to caregiver/family sleep and well-being. Mothers described their participating preschoolers as sleeping better and being in a better mood following BfK participation, and some mothers described how much their child “loved” the BfK bed and sleeping in it. Perceived family-wide benefits to BfK participation included similar sleep improvements in other children in the home who received a bed, with corresponding increases in positive mood (eg, feeling “happy” and “excited” to have an independent sleep space). One caregiver explicitly referenced reductions in child worries about family finances following BfK participation and described that “things are looking up for us [the family]” (ID 132, Table 4). Mothers also described BfK participation as helping them create routines for themselves and for their children.

Table 4.

Qualitative themes and representative quotations.

Theme	Mothers’ Quotes
Child sleep benefits	“Like her [child’s] bed was so comfortable to her, I could tell it was comfortable to her cause she slept. She slept throughout the night.” (ID 111)
	“You know, she’s sleeping more.… Now that she do have her own bed, she’s in her bed, you know, she loves it, she sleeps in it, she makes it up, she’s telling everybody ‘quick get off my bed.’” (ID 149)
	“He does actually sleep a lot better with his bed. Like you can totally see a difference between it. And in general, like his mood is better.… When he got the bed, it [sleep and behavior] like completely changed. Like he was sleeping better, his mood was better, so it was nice.” (ID 117)
	“So she [child] like, ‘Mommy, I got my teddy bears, I wanna sleep by myself.’ So she really loves sleeping on her bed. She doesn’t wanna sleep in my bed.” (ID 145)
Challenges with bed transition for young children	“He’s [child] never like really slept on his own, but now he’s learning to like, you know, this is your bed, this is mommy bed. So you have to sleep in your bed. He—he will sleep in the bed after like I have to put him to sleep in my bed then I have to transfer him to his bed. So he will sleep all night. That’s pretty good. So he’s getting used to that.” (ID 143)
	“They [the beds from BfK] seem good but he [child] one of them kids that like to cuddle so that’s the only hard part. Like him sleeping alone, ‘cause he’s so used to sleeping next to me so he just gotta get used to sleeping by his-self. Cause he used to be cuddled up like, since he was a baby he just like to be held. So that’s another fast way he falls asleep, to be held. So if he sleep in his bed, it just be little harder so I cuddle up with him first and then I put him in the bed. He a little scared. It’s new.” (ID 115)
	“When we put [the child] to bed between me and her dad, we lay down with her until she falls asleep. And then as soon as she fall asleep, I slide off the bed towards the door.… And she could be asleep for like a couple hours and the next thing you know, I hear a knock at the door and her crying…. I guess just because she didn’t feel us next to her. And she’ll come right back in my room and she’ll get on my bed and go right back to sleep.… I think it’s the comfort of us, her being next to us to be asleep.” (ID 116)
Impact of BfK participation on caregiver/family sleep and well-being	“It changed they [all children in the home] attitude. They’re happy now. They like, ‘Yeah, I got my own bed. I got my own room.’ … They were—all three of my kids were very excited. It’s comfortable.… Now everybody’s in they own space. I have a routine, I have a schedule, you know, clean up your room. It’s just—make your bed. It just teach you more things. Give you stuff to do with your child … more of a life.” (ID 110)
	“I think it’s good for parents to go into a service like this. Not just to help out for yourself and for your child, it’s to help out for, you know, families that struggle with they kids sleeping, you know?” (ID 111)
	“Before getting the beds, they [all children in the home] was always in the bed with me. So I thought once they got they beds, it was gonna be hard for them to like adapt to them but they actually was happy, I guess ‘cause they had their own room and stuff like that.… They stayed in they room more. They always on they beds. They just was happy about the whole thing. It came with little teddy bears and stuff like that. So they was really excited.” (ID 112)
	“It [having the beds from BfK] helps me to like figure out my own schedule too because I have to do things, I have to do for me, and I have to do things away from them. You know, so that kind of just helped me channel out a little bit about what I have to do with my space too.” (ID 149)
	“They [all children in the home] sleep better at night now. Like when I mean better, I mean actually get a good night’s sleep. Eight hours or more.… And they not—and it also puts in perspective like, they wasn’t worried about, ‘Oh, mommy don’t have enough money to do certain things now.’ Because now I feel like we got our beds, things are looking up for us. And that changed their sleep. Cause they not worried no more.” (ID 132)
Family beliefs and norms around electronics and sleep	“It [electronics] keeps ‘em up.… They [the children] will be—they will be tired and still be on it. And when you take it away from them, they instantly fall asleep. So, you know like they just be like up on it just because. Like even when they is sleepy. So that’s why I kind of take it from them.… They’ll be whining for a little bit but eventually it’ll stop.” (ID 112)
	“When I was a teenager, I didn’t have a TV in my room. I just wanted to be on my phone. But the thing is I had to learn to live without. So, when I didn’t have a phone, or TV, I just say, ‘Alright then sleep.’ You know? Got the best rest. I woke up feeling energized. So, I said to myself, huh, maybe TV in, you know, electronics do hold you back from the sleep. You know? And you gotta—sometimes I gotta have to make a bold choice. Cut the TV off. Even her [child’s] dad cuts the TV off.” (ID 111)
	“Electronics, I think they’re really addictive … but when it’s time to like go to bed … it’s like he [child] needs that [electronics] because that’s like more his—that’s like more of his downtime. That winds him down so. I try not to keep them [electronics] on at—that’s why I don’t give it to him all day. So, I just try to give it to him at nighttime so that way he can fall asleep.… Like if I take it away or if it dies, he will cry for probably about 10 minutes and then he’ll go and find something else to do.” (ID 143)
Family/cultural norms around caffeine consumption	“That’s hard because us as African Americans, we drink stuff like that [caffeine]. So, we—when you grew—when you grow up on that and it’s like a tradition, it’s like saying you can’t really change.” (ID 110)
	“I wouldn’t really give my kids soda like that. They—I don’t give my kids soda at all. So they barely get candy from me so I just think that caffeine’s too much for a kid, period.… I just grew up—my mom never gave me caffeine when I was little.… I just think it’s not healthy.” (ID 112)
Contextual contributors to an early child bedtime	“It took me a while to get like the same routine [earlier bedtime] because he’ll be up sometimes like late, 10:00. But once he started school, I had to really like be a little harder on him and stick to routine so he could have a good night’s sleep and then, like the next day go to school, he be well-rested. So that really made me like want to have a like, a set routine with him. ‘Cause he be up like 10:00 at night. He only 3, so like he need to have a set bedtime. And 7:30–8:00 and that work out good.” (ID 115)
	“I would say—well work changes is coming soon. So I’ll be starting a new job so that’s another reason why I’ve been trying to make it [child’s bedtime] earlier cause we do have to get up about 20 to 30 minutes earlier now. So that’s why I’ve been changing it.” (ID 117)

At the same time, challenges with bed transition for young children were common, with many mothers describing children’s difficulties in adjusting to an independent sleep space. Some mothers attributed these challenges to young children’s being accustomed to sleeping with family members since infancy. In addition, feedback on BfK healthy sleep guidance (an early bedtime, no electronics in the bedroom, no caffeine) was mixed, because some mothers described barriers to these practices. Contextual contributors to an early child bedtime were common and included social and environmental factors that could help or hinder an early and consistent child bedtime, such as school and caregiver work schedules. Family beliefs and norms around electronics and sleep generally reflected the idea that electronics can negatively affect sleep; however, in some cases these items were described as helpful to “wind down” children at bedtime. Family/cultural norms around caffeine consumption emerged in relation to the caffeine avoidance recommendations. Whereas some mothers described avoiding caffeine already, often referencing the affect of their own upbringing in this regard, one mother (ID 110) described drinking caffeine as a “tradition” among Black/African American families.

DISCUSSION

This multimethod pilot randomized trial found that participation in a program involving child bed provision and sleep health education (BfK) resulted in improved self-reported maternal sleep by about 1 hour among families assigned to the intervention condition compared to those in the waitlist control group. However, during this 7-day intervention period no group differences were observed in actigraphy-derived child nighttime sleep duration and variability or in mother-reported nightly child sleep patterns and quality, child behavior, or maternal mood. Across conditions, at 1-month follow-up we found improvements in the frequency and duration of mother-reported child night awakenings, child sleep quality, overall child sleep problems, and child internalizing and externalizing problems, even after adjusting for child and family characteristics linked to child sleep, although child sleep duration and bedtimes did not improve. In addition, qualitative findings suggested that mothers perceived improved sleep and well-being for all family members following BfK participation.

The null findings for actigraphy-based and mother-reported child sleep duration outcomes 1 week following bed provision and at 1-month follow-up differ from those in our prior trial evaluating BfK bed provision and the impact of written sleep health education (vs dental hygiene education), which demonstrated an average improvement of about 20 minutes in child nighttime sleep duration at 4-week follow-up.¹⁵ These previously found improvements were observed across children ages 2–12 years and were increased among those who received sleep health education.¹⁵ These differences in study findings at 1 week after bed provision vs 1-month follow-up are interesting and may help inform expectations of the impact of such programs over time. It also should be noted that the age range in each of these studies differs; older children were included in the prior study, which may indicate differences in sleep outcomes based on the age of the child.

In the current study, BfK participation may have resulted in greater instability in child sleep and behavior due to the challenges of transitioning young children from bed-sharing to an independent sleep space. Even at 1 month after bed delivery, mothers qualitatively reported current challenges in making this transition, noting that their young children were still adjusting to sleeping alone. Sleep-space transitions are often challenging in early childhood (for instance, moving from a crib to a bed)³⁶ and this could be the case when transitioning from bed-sharing to an independent sleep space, which may also involve changing child sleep onset associations. It is also possible that, especially for young children, sharing a sleep space is preferable and comforting after living in foster care³⁷ and/or unstable, temporary, or transitional housing conditions,¹⁴ which is common among families referred to BfK, although we did not evaluate prior child foster care involvement or family housing arrangements in this study. Even for young children who have not faced family separation, bed-sharing may increase children’s feelings of security and/or safety, leading to reduced sleep onset latency, and caregiver presence overnight may result in fewer child awakenings that disrupt and/or signal the caregiver.

Longer-term findings, at 1-month follow-up, indicated improvements in child sleep, including reduced night wakings and better mother-perceived sleep quality. These findings support the need for inclusion of information on the challenges of transitioning young children to an independent sleep space in sleep education and providing realistic short-term and long-term expectations for families. In addition, although families are referred to BfK because they have expressed a need and desire for individual beds for their children, there is no assessment of family preferences around sleep arrangements as part of BfK. Examining family preferences for bed-sharing as part of BfK enrollment may be especially relevant for families with young children, who are more likely to share a bed compared to older youth. Understanding these preferences prior to program participation could help tailor BfK recommendations and bed provision for families that are bed-sharing with their young child.

The observed pre-BfK to 1-month follow-up improvements across conditions in mother-perceived child sleep quality, sleep problems, frequency and duration of night awakenings, and child internalizing and externalizing are very promising but should be confirmed in a larger randomized trial. Mothers’ qualitative reports of improved child sleep further support these quantitative findings, as does the large body of research documenting associations between healthy child sleep and better behavioral functioning.^2,6,38 It is especially notable that mother-perceived child sleep problems were reduced from before BfK to follow-up, indicating this may be another benefit of child bed provision and sleep health education. The reduction in sleep problems could be related to reductions in the frequency and duration of night awakenings, given that frequent night awakenings are robustly associated with the likelihood of an overall caregiver-perceived child sleep problem in early childhood.^39,40 Sleep outcomes beyond timing and duration were not assessed in our prior BfK study,¹⁵ underscoring the need for additional research on potential benefits of bed provision and sleep education to multiple domains of sleep health, sleep problems, and daytime functioning in both younger and older children.

The improvements in intervention mothers’ self-reported nightly sleep duration during the 7-day intervention period, as well as the qualitative findings across study conditions, suggest that the most salient immediate benefits of BfK participation may not be children’s sleep, at least not for young children, but mothers’ outcomes. A 1-hour improvement in sleep duration among adult caregivers is clinically meaningful and could have important downstream benefits on maternal mood,⁴¹ health,⁴² and parenting practices.^16,43 Qualitatively, mothers also described feeling as though BfK participation, and the bed provision component in particular, facilitated the implementation of family routines, allowed for mothers to focus on their own needs, and helped improve the well-being of all children in the home. These maternal and family benefits should be explored further in sleep health promotion interventions in future research.

In line with previous research,⁴⁴ our qualitative findings also suggest that increased tailoring of both BfK-specific and broader sleep health education in early childhood could be beneficial to better align content with families’ cultural practices and context. As noted above, including recommendations to help with the transition to an independent sleep space may be helpful in sleep health guidance, especially for families with young children. Cultural preferences and family norms/traditions around caffeine consumption and evening electronics use should also be integrated into this education. Additional qualitative research in collaboration with families is needed to identify how such information could be packaged in a culturally humble manner⁴⁵ that does not make assumptions about preferences based on family race and ethnicity.

Study findings are limited by the pilot nature of this evaluation and its small sample, which affected our ability to detect small improvements in child sleep and other outcomes. In addition, participants were recruited from one hospital system (CHOP), whereas the BfK program receives referrals from a variety of social service agencies in the greater Philadelphia area. Future research on early childhood BfK outcomes should utilize a larger sample while balancing the program’s need to serve a broad population of families, including children of older ages, referred by multiple social service agencies. Of note, despite the small sample in this study the significant intervention effects identified were in the small-to-large range.

The use of multiple methods to assess BfK outcomes is a study strength; however, there are still limitations to these approaches in accurately assessing early childhood sleep and behavior. For example, the impact of bed-sharing on the reliability and validity of actigraphy scoring in early childhood is unknown²⁰ and could artificially inflate movement and/or awakening readings on actigraphy data. In addition, mothers’ daily diary reports of child sleep were not always in agreement with actigraphy data.²¹ Furthermore, we only examined 1 week of actigraphy after bed provision to minimize family burden and conform with recommendations from BfK regarding study feasibility. These data collection methods should be examined further over a longer period of time, as well as in sociodemographically similar samples, given that most of the actigraphy validation research, with a few rare exceptions,^23,46 has been conducted with children of higher socioeconomic status and/or non-Latinx White backgrounds. Given that electronic deployment of daily diaries was not available for this study and participants had to complete these measures by telephone with research assistants, social desirability could have affected these ratings. Mothers’ awareness of their intervention group assignment may also have influenced their reporting on maternal and child sleep and behavior postintervention. In addition, it would be beneficial to assess longer-term outcomes within the context of an appropriate control group; however, ethical considerations of limiting bed provision to eligible families needed to be considered in this study.

CONCLUSIONS

Receiving an individual child bed and sleep health education in preschool-aged children has immediate benefits on maternal sleep duration, with evidence of improvements in mother-reported child night awakening frequency and duration, sleep quality, and behavioral concerns at 1-month follow-up. Our findings suggest that BfK benefits to young children in particular may not emerge until the children are accustomed to sleeping independently. More importantly, qualitative findings point to improvements in broad family and child well-being following program participation. Collectively, these findings have important implications for policies that seek to improve maternal and child health by providing families of lower socioeconomic backgrounds with tangible goods and resources, such as the Special Supplemental Nutrition Program for Women, Infants, and Children or the US Public Housing Program, which assists with rent but does not provide household furnishings. Expanding such programs to include beds, bedding, and sleep education is a crucial direction for researchers and policymakers to explore, because providing more comprehensive child health resources may help ensure long-term child and family well-being.

DISCLOSURE STATEMENT

All authors have reviewed and approved this manuscript. Work for this study was performed at Children’s Hospital Philadelphia, Philadelphia, Pennsylvania and One House at a Time, Huntingdon Valley, Pennsylvania. This work was supported by a Children’s Hospital of Philadelphia Center for Pediatric and Perinatal Health Disparities pilot grant awarded to Dr. Williamson. Dr. Williamson is also supported by the National Institute of Child Health and Human Development (K23HD094905). Dr. Mindell has served as a consultant for Johnson & Johnson Consumer Inc. Dr. Williamson is a board member for the Beds for Kids program (no compensation provided). The other authors report no conflicts of interest.

ABBREVIATIONS

BCSQ

Brief Child Sleep Questionnaire

BfK

Beds for Kids

CHOP

Children’s Hospital of Philadelphia