Sleep SAAF: a responsive parenting intervention to prevent excessive weight gain and obesity among African American infants

Sleep SAAF’s responsive parenting intervention is grounded in the developmental literature on parenting sensitivities, arguing that parenting that is developmentally appropriate, prompt, and contingent on the infant’s needs [20] can increase infant sleep and self-regulation, and reduce rapid infant weight gain. The intervention is focused on a critical developmental period when plasticity is high. How parents put their infant to bed (e.g., bedtime routines, falling asleep independently) and respond to nighttime awakenings represent important sources of variability in infants’ developing sleep patterns [21, 22]. Because infant sleep bouts are interrupted by waking to be fed, the routines parents establish around sleep can also affect feeding frequency and energy intake, with important consequences for infant weight gain. Similarly, how parents respond to infant crying represents an important source of variability in infants’ developing regulation skills; avoiding the use of feeding as the first response to infant distress by using alternative soothing strategies can increase parental responsiveness to infants’ needs and promote infants’ self-regulation.

Individual and contextual stressors are common among African American mothers and have the potential to affect intervention effectiveness. For example, depressive symptoms are prevalent among African American mothers of infants [23] and are likely to interfere with maternal responsiveness [24]. The broader social context is also likely to pose challenges [25]. The families participating in Sleep SAAF live in communities in Georgia where poverty rates are high and unemployment rates are above the national average [26]. Under conditions of economic distress, parents are vulnerable to anger, low frustration tolerance, and depressive symptoms [27‐30], all of which may be consequential for responsive parenting. African American families living in the South must also contend with race-based stressors such as racial discrimination, with negative consequences for depressive symptoms, health, and responsive parenting [31‐35]. Lastly, many African American mothers do not have a steady romantic partner at the time of their infant’s birth [36], potentially limiting available social support and increasing parenting stress [37]. We examine each of these stressors as moderators of program effectiveness. Our conceptual framework is shown in Fig. 1.

Sleep SAAF builds on our two previous RCTs providing evidence for the efficacy of a responsive parenting program. In a pilot study, Birch and colleagues developed and evaluated the SLIMTIME intervention for primiparous mothers who intended to breastfeed [19]. Participants were predominantly middle income and White. The first component of SLIMTIME, “Soothe/Sleep,” began when infants were 2–3 weeks old and focused on promoting aspects of responsive parenting to promote the development of infant self-soothing and longer sleep duration and to reduce feeding frequency. As predicted, there was evidence of changes in sleeping, feeding, and soothing as a result of the intervention [19, 38]. Infants whose mothers participated in the Soothe/Sleep intervention slept longer at night, had fewer total daily feeds, and had fewer nocturnal feeds from age 3 to 16 weeks compared to control group infants. Infants who received the intervention also gained weight more slowly over the first year, indicating that intervening on sleep and soothing behaviors can reduce the rate of weight gain during infancy.

In a subsequent larger scale randomized controlled trial (INSIGHT), the sample included primiparous mothers intending to formula feed as well as those intending to breastfeed and was predominantly middle income and White [39]. This comprehensive responsive parenting program addressed infant feeding, sleeping and soothing, active social play, emotion regulation, and growth record education, and was taught to mothers at four home visits at 2-, 16-, 28-, and 40 weeks. During the first 16 weeks, program content focused on soothing, sleeping, and breast or bottle feeding; after 16 weeks, the focus was expanded to include introducing solids, table foods, and active social play. Relative to a home safely control, infants in the responsive parenting condition gained weight more slowly than control group infants from birth to 6 months, had lower weight-for-length percentiles at 1 year, were less likely to be overweight at 1 year, and had lower BMI z-scores through age 3 [14, 17]. There were also positive sleep outcomes: infants in the responsive parenting condition had longer nighttime sleep duration, more consistent bedtime routines, and earlier bedtimes, and were more likely to self-soothe to sleep without being fed at 8-, 16- and 40 weeks of age [18].

Sleep SAAF builds on the efficacy of these multi-component responsive parenting programs. Focusing on the initial soothe/sleep component of SLIMTIME and INSIGHT, we test whether a responsive parenting program can promote infants’ sleep and self-soothing and reduce rapid weight gain during the first 16 weeks of life among African American infants born in the South. Our specific aims are:

Sleep SAAF is a two-arm randomized, controlled clinical trial evaluating the effectiveness of a responsive parenting intervention relative to a child safety control. Following initial screening and consent in the hospital by the project’s Recruitment Coordinator, mothers and their infants were visited at home by CRAs at 1 week, 3 weeks, 8 weeks, and 16 weeks postpartum for intervention and data collection visits (Fig. 2). To minimize bias during the assessments, separate CRAs completed the intervention visits (at 3 weeks and 8 weeks) and data collection visits (at 1 week, 8 weeks, and 16 weeks). Further, CRAs were trained in (and subsequently delivered) only one of the two study interventions (responsive parenting or child safety).

Funding for this project comes from the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK; R01DK112874) and the study was registered on www.​clinicaltrials.​gov (NCT03505203) on April 3, 2018. This study was approved by the Augusta University Institutional Review Board and overseen by a Data Safety Monitor, based on NIDDK guidelines that a Data Safety Monitor is more appropriate for single-site clinical trials that are not masked and are minimal risk, as is the case in the current study. Additional details on the Data and Safety Monitoring Plan are provided in Additional file 1.

A total of three hundred African American mother-infant dyads will be recruited from the mother/baby nursery at Augusta University Medical Center (AUMC) in Augusta, GA. Recruitment began in the spring of 2018 and will continue through the summer of 2021. AUMC provides comprehensive medical care for the Augusta area and serves as a referral center for the state of Georgia and western South Carolina. All newborns delivered at AUMC were screened for participation. The project’s Recruitment Coordinator had access to electronic medical records systems in order to pre-screen mothers and infants based on the following criteria: (1) full-term infants (> 37 0/7 weeks gestational age), (2) singleton infant; (3) infant > 2500 g at birth; (4) primiparous mother > 17 years of age, (5) mother self-identifying as African American; (6) residence < 75 miles from Augusta; and (7) English speaking. In addition, mother-infant dyads were excluded if there was a congenital anomaly or neonatal physical or metabolic condition that would significantly affect the newborn’s feeding or growth (e.g. cleft palate, complex congenital heart disease); if there were major maternal morbidities or pre-existing conditions that would affect postpartum care or the mother’s ability to care for her newborn (e.g., narcotic drug use, uncontrolled depression causing social service contact); if there was a plan for the newborn to be adopted; or if the family planned to move from the area within four months of delivery.

Participants were randomized to either the responsive parenting intervention or child safety control after completion of the 1-week data collection visit, stratifying on sex-specific birth weight for gestational age (<50th percentile or > 50th percentile) [40] and intended feeding mode (breastfeeding or formula). Randomization was done using a secure Microsoft Excel application prepared by a statistician unaffiliated with the study and administered by the study’s Project Coordinator. Blinding was maintained after allocation only for the CRAs completing the data collection visits. Participants were made aware of their condition, CRAs serving as intervention facilitators were aware of condition because they were only trained in one of the two interventions (responsive parenting or child safety), and study investigators and other project staff (e.g., Project Coordinator, Intervention Coordinator) were aware of condition.

The responsive parenting intervention focused on responsive parenting in the context of (1) sleeping, (2) crying, (3) feeding, and (4) playing (Fig. 3). These topics were introduced and discussed in detail at the first intervention visit at 3 weeks postpartum and reviewed at a booster session at 8 weeks postpartum. All materials were tailored for African American families.

Sleep SAAF’s control group received a developmentally-appropriate child safety intervention. The home visits were designed to be equal in length and intensity to the responsive parenting intervention visits and to avoid messages that could impact responsive parenting. At 3 weeks, mothers received an extended version of the safe sleep strategies presentation from NICHD [42], which provided more detail about the Back to Sleep Campaign, SIDS facts and myths, reducing baby’s risk, and the importance of good health care. In addition, information was provided about taking care of a crying baby, finding caretakers for baby, and food safety (e.g., formula and bottle handling, preparation, and storage). At the 8-week intervention visit, control group mothers received guidance on general child safety using materials from Safe Kids Worldwide [44]. This included car seat safety as well as home safety tips such as preventing falls, water safety, poison prevention, fire safety, and preventing burns.

To assess intervention impacts on the primary outcome, intermediary behavioral processes, and moderators of intervention effectiveness, data collection included measures divided into the following categories: Anthropometry, Sleep Outcomes, Infant Soothing and Feeding Outcomes, Mothers’ Outcomes, Child Safety Outcomes, Stress-Support Moderators, and Demographics and Health History. As shown in Table 1, many of these measures were assessed multiple times. CRAs also rated coparent involvement in the intervention and implementation quality following the intervention sessions.

Self-report data were collected using a secure, web-based survey interface, Qualtrics, on iPads. When literacy concerns arose, CRAs read the questionnaires to mothers, who subsequently entered their answers directly into Qualtrics. Anthropometry measurements were taken and recorded using Qualtrics by the CRAs.

The primary study outcome was a between-group comparison of infant conditional weight gain (CWG) scores from 3 weeks to 16 weeks, calculated following the method of Griffiths and colleagues [45] and described in our recent work [17]. There were several secondary weight-related outcomes that we compared between groups as well, including differences in change in infants’ weight for age over time and differences between groups in infants’ weight outcomes at age 16 weeks, including BMI z scores, weight-for-age z scores, and weight-for-length z scores.

Additionally, there were several other pre-specified outcomes assessing infant and maternal responses to intervention, including: 1) Maternal report of infant sleep (e.g., sleep duration, night awakenings); 2) Maternal sleep actigraphy (e.g., sleep duration, sleep efficiency, wakefulness); 3) Infant soothing (e.g., mothers use feeding for reasons other than in response to hunger); 4) Infant feeding frequency (e.g., feeding frequency, use of bottle feeding, introduction of solids); 5) Safe sleep practices; 6) Child safety practices; 7) Infant eating behavior; 8) Maternal self-efficacy; 9) Maternal feeding practices and beliefs; 10) Maternal depressive symptoms; and 11) Family routines.

The primary outcome for Sleep SAAF is between-group differences in CWG scores from 3 to 16 weeks. We used G Power 3.1.9.2 to estimate the smallest intervention effect size we could detect with 80% power, a 5% two-tailed Type 1 error rate, and two groups of 150 each (assuming 10% attrition over the course of the study) [46]. We achieved 80% power with an effect size d of 0.34. Our prior research showed an effect size of approximately 0.4 for infant conditional weight gain at 6 months [17], suggesting that we will have sufficient power in our sample to detect intervention effects.

All primary statistical analyses will invoke the intent-to-treat paradigm, analyzing data based on randomized assignment. Missing data will be handled using full-information maximum likelihood methods, which use all available information to estimate parameters, making this approach more efficient and less biased than other methods when data are missing at random [47]. All models considered allow for the inclusion of relevant covariates. Covariates to be considered in refining the analyses, in addition to those described in detail in the following paragraphs, include demographic factors (e.g., maternal employment and hours worked, gestational diabetes), weight-related factors (e.g., maternal pre-pregnancy weight/BMI, gestational weight gain, maternal smoking), and intervention-related factors (e.g., implementation quality, coparent involvement in intervention). We will also examine (1) infant sex as a covariate, given different growth charts for boys and girls [48] and (2) feeding mode as a covariate, given differences in weight gain between breastfed and formula-fed infants [49].