Prev Sci (2014) 15:643–653 DOI 10.1007/s11121-013-0408-4
Negative Temperament as a Moderator of Intervention Effects in Infancy: Testing a Differential Susceptibility Model Stephanie Anzman-Frasca & Cynthia A. Stifter & Ian M. Paul & Leann L. Birch
Published online: 6 July 2013 # Society for Prevention Research 2013
Abstract A consideration of potential moderators can highlight intervention effects that are attenuated when investigating aggregate results. Differential susceptibility is one type of interaction, where susceptible individuals have poorer outcomes in negative environments and better outcomes in positive environments, compared to less susceptible individuals, who have moderate outcomes regardless of environment. In the current study, we provide rationale for investigating this type of interaction in the context of a behavioral childhood obesity preventive intervention and test whether infant negativity moderated intervention effects on infant self-regulation and weight gain and on two aspects of mothers’ parenting competence: parenting self-efficacy and parenting satisfaction. Results showed that infants’ negative temperament at 3 weeks moderated intervention effects on some, but not all, outcomes. The intervention led to greater parenting satisfaction in mothers with highly negative infants but did not affect parenting satisfaction in mothers with less negative infants, consistent with a model of differential susceptibility. There was also a trend toward less weight gain in highly negative Dr. Anzman-Frasca is now with the Friedman School of Nutrition Science and Policy, Tufts University. S. Anzman-Frasca : L. L. Birch The Center for Childhood Obesity Research, Pennsylvania State University, University Park, PA, USA C. A. Stifter : L. L. Birch Department of Human Development & Family Studies, Pennsylvania State University, University Park, PA, USA I. M. Paul Department of Pediatrics, Pennsylvania State University College of Medicine, Hershey, PA, USA S. Anzman-Frasca (*) Friedman School of Nutrition Science and Policy, Tufts University, 150 Harrison Avenue, Boston, MA 02111, USA e-mail:
[email protected]
intervention group infants. In contrast, there was a main effect of the intervention on infant self-regulation at 1 year, such that the intervention group had higher observed self-regulation, across levels of infant negativity. Results support the importance of incorporating tests of moderation into evaluations of obesity interventions and also illustrate that individuals may be differentially susceptible to environmental effects on some outcomes but not others. Keywords Temperament . Individual differences . Obesity prevention . Differential susceptibility . Infancy
A consideration of potential moderators can highlight intervention effects that would otherwise be masked when investigating aggregate results. There are numerous individual factors that may interact with intervention effects, such as temperament and genetics, and there are multiple types of interactions. Differential susceptibility is an interaction where susceptible individuals have poorer outcomes in negative environments and better outcomes in positive environments, compared to less susceptible individuals, who show moderate outcomes regardless of context. For example, children with higher negativity had more behavior problems when they experienced low-quality childcare, but they benefited more from high-quality childcare than children with low negativity (Pluess & Belsky 2010). In this example, negative temperament is the susceptibility factor that moderates environmental impact, and the environment is conceptualized in terms of childcare quality. In interpreting such interactions, it is posited that certain children (e.g., those with a negative temperament) have a heightened susceptibility to environments, which is why their outcomes are more likely to be predicted by environmental variability (Belsky et al. 2007). Differential susceptibility is framed as an evolutionary theory, where individual differences in susceptibility to environments are considered to be adaptive, as
644
Prev Sci (2014) 15:643–653
In the current study, we applied a differential susceptibility model to the assessment of an early obesity preventive intervention (Paul et al. 2011). Childhood obesity is an epidemic in the USA and in many other countries, affecting even the youngest children. Currently, one fourth of American preschoolers are overweight (Ogden et al. 2012). Some researchers have turned to obesity prevention during infancy to address these trends (Hesketh & Campbell 2010). Effects of one such intervention on infant weight outcomes have been published previously (Paul et al. 2011). The goal of the current study was to follow-up on these overall results, using a differential susceptibility model to examine whether highly negative infants were more susceptible to a Soothe/Sleep intervention, designed to promote responsive parenting around infant soothing and sleeping to attenuate excessive infant growth. We hypothesized that the intervention would positively affect highly negative infants, and that effects would not reach significance in less negative infants. This differential susceptibility model was proposed due to: (1) the likelihood that parents of highly negative infants would have a greater need for soothing and sleep strategies and (2) observed associations between negativity and obesity risk in the literature (Anzman-Frasca et al. 2012b).
The Moderator: Early Temperamental Negativity as the Susceptibility Factor Although many individuals develop obesity in the context of the current obesity epidemic, there are individual differences
Low
High
+ Outcome -
-
-
Outcome
Outcome
+
+
environmental conditions over time are uncertain (Boyce & Ellis 2005). Susceptible individuals are equipped to thrive in positive environments, while less susceptible individuals are equipped to withstand negative environments. The differential susceptibility literature typically examines interactions between susceptibility factors at the genetic and behavioral levels and natural variability in micro-level environments like parenting or childcare quality (Ellis et al. 2011). Alternatively, intervention and control conditions may be used to represent environmental variability. Applying a differential susceptibility model to an assessment of intervention effects can advance both developmental and prevention research, by allowing for causal investigations of differential susceptibility (Ellis et al. 2011) and elucidating subgroup differences in treatment effects (Wang & Ware 2011). Moderation of intervention effects by individual differences in initial risk, including child factors like aggression (August et al. 2001) and behavior problems (Dawson-McClure et al. 2004; Reid et al. 2004) and demographic factors like parent education level (Gardner et al. 2009) and socioeconomic status (Olds et al. 1997), has been demonstrated, showing that children at higher risk can benefit more from interventions, including parenting interventions. Differential susceptibility represents a special case as a particular pattern of interaction effects is hypothesized (Fig. 1a), where only the high-risk individuals are susceptible to intervention effects, showing better outcomes than low-risk individuals with the intervention and poorer outcomes than low-risk individuals without it. To our knowledge, such patterns have not been investigated in the context of obesity prevention.
Low
Negativity
High Negativity
A Differential susceptibility
B Contrastive effects
= Intervention group
Fig. 1 Patterns of hypothetical interactions between negativity and intervention group. The x-axis represents negativity, an aspect of temperament. The two lines represent intervention group assignment, which is a special case of environmental variability. The y-axis represents intervention outcomes, with higher standings indicating better outcomes. (For some outcomes, like excess weight gain, the "better" direction would be the reverse.) Differential susceptibility (a), contrastive effects (b), and dual risk (c) patterns of interactions are pictured. In the differential susceptibility example (a), more negative individuals (the susceptible individuals) have better outcomes in the intervention group and poorer outcomes in the control group. The less negative individuals (the less susceptible individuals) have moderate outcomes in either context. The
Low
High Negativity
C Dual risk = Control group
next two examples (b, c) demonstrate other ways in which individual differences like negativity may interact with environmental variability to predict outcomes. These patterns of interactions have previously been depicted by Belsky et al. (2007), in the context of general individual factors and environmental variability, as opposed to negativity and intervention groups specifically. Another difference is that environmental variability is depicted by the lines here, rather than being depicted on the x-axis. This was done for ease of interpretation given that environmental variability in this case is a dichotomous variable (intervention or control group) while negativity is continuous. The substantive conclusions drawn from the graphs remain the same
Prev Sci (2014) 15:643–653
in obesity risk, and early temperament is one factor that is associated with weight outcomes. Temperament refers to predispositions to behave a certain way in response to an eliciting context or situation. Rothbart and colleagues define temperament as “individual differences in reactivity and selfregulation” (Rothbart & Bates 2006), and both reactive and regulatory aspects of temperament have been linked to obesity risk (Anzman-Frasca et al. 2012b). Higher levels of infant negative reactivity, particularly distress to limitations, predict a greater weight status or weight gain (Darlington & Wright 2006; Anzman & Birch 2009), suggesting that a highly negative temperament may be a risk factor for obesity. Better selfregulation abilities, including the ability to delay gratification, predict a lower weight status (Francis & Susman 2009; Graziano et al. 2010), suggesting that self-regulation may be a protective factor. Early negativity may lead individuals to become obese if highly negative infants are fed more often as an attempt to soothe their distress (Anzman-Frasca et al. 2012a; Carey 1985; Darlington & Wright 2006; Stifter et al. 2011). This process could affect the development of behavioral patterns over time. Temperament may predispose individuals to certain behavioral patterns, but outcomes also depend on environmental contexts, such as the microlevel parenting environment. An emerging literature demonstrates links between parenting and obesity (Anderson et al. 2012; Brotman et al. 2012; Rhee et al. 2006; Wake et al. 2007), with evidence that temperament and parenting interact to predict child weight outcomes (Sleddens et al. 2011; Wu et al. 2011). For example, Wu et al. (2011) found that infants who were high on negativity and who also experienced insensitive parenting had an increased risk of obesity during childhood. In addition to being highlighted as a risk factor for subsequent obesity, early negativity is also a susceptibility factor, with evidence that negative infants are more susceptible to positive and negative rearing environments, compared to less negative infants (Feldman et al. 1999; Pluess & Belsky 2010; Stupica et al. 2011). This evidence comes from studies investigating behavioral (as opposed to weight) outcomes. For example, 6-month-olds who were higher on difficult temperament, a construct that overlaps with negativity, were more susceptible to variability in the quality of parenting and childcare: for difficult infants, positive relationships between these environmental variables and subsequent cognitive and social outcomes were stronger than they were for less difficult infants, and difficult infants who had experienced high-quality childcare had the lowest rates of behavior problems and teacher–child conflicts in sixth grade (Pluess & Belsky 2010). Negative infants’ differential susceptibility has been demonstrated via child outcomes like these, as well as outcomes of their parents (Klein Velderman et al. 2006).
645
The Predictor: A Parenting Intervention as the Environmental Factor Just as individuals with different temperaments may be more or less susceptible to observed variability in parenting practices or childcare quality, they may also respond differentially to an intervention. Greater effects of preventive interventions may be revealed when accounting for differential susceptibility to the environmental variability that exists across intervention and control groups. There is some evidence consistent with this idea, where experimental manipulations of parenting showed greater effects in highly negative infants. In an intervention in which mothers were provided with feedback after interacting with their infants, dyads with highly reactive infants were more susceptible to intervention effects on parenting sensitivity and infant attachment security (Klein Velderman et al. 2006). Similarly, negative emotionality moderated effects of an intervention for preterm, low birth weight infants: for infants higher on negative emotionality, the intervention had greater positive effects on behavior problems and cognitive abilities at age 3 (Blair 2002). These findings add to the evidence that highly negative individuals are more susceptible to many types of environmental variability. As mentioned, negativity has also been linked to obesity risk, so it follows that negative infants may benefit most from a behavioral obesity preventive intervention aiming to promote healthier growth through responsive parenting around behaviors that have been linked to obesity risk, including soothing and sleep (Paul et al. 2011). Our Soothe/Sleep intervention allows an empirical test of whether highly negative infants are susceptible to obesity risk in the context of less responsive parenting but experience healthier weight outcomes in the context of more responsive parenting, compared to less negative infants. We explored the target outcome of weight gain but also hypothesized that dyads with highly negative infants would show greater intervention effects on infant self-regulation and maternal perceived parenting competence.
The Outcomes: Investigating Intervention Effects on Behavior, Weight, and Parenting We investigated multiple outcomes to address the specificity of the hypothesized differential susceptibility effects (Belsky et al. 2007). Significant interactions may be consistent with differential susceptibility in the case of some outcomes but not others; alternate types of interactions include dual risk, where negative infants are at risk in negative environments but do not attain additional benefits from positive environments, or contrastive effects, where the relationship between the intervention and outcome is statistically significant for both highly
646
negative and less negative infants, in opposing directions (Fig. 1; Belsky et al. 2007). Stupica et al. (2011) found that highly negative infants were differentially susceptible to the influence of attachment security on sociability, but when investigating a different outcome, exploration, the interaction was instead consistent with a dual risk model. In the current study, we were particularly interested in outcomes that could be (1) plausibly linked to the soothing and sleeping messages delivered in the intervention, and (2) implicated as potential antecedents of positive child and family outcomes in multiple domains. We identified infant self-regulation and perceived parenting competence and as outcome variables that fit these criteria. During infancy, self-regulation abilities are limited, and infants regulate distress with help from their caregivers. Effective parental soothing can scaffold children’s developing self-regulation abilities (Fox & Calkins 2003; Jahromi & Stifter 2007). Higher self-regulation is predictive of positive outcomes in many domains, including weight outcomes, as well as the development of conscience, empathy, social competence, compliance, and academic achievement (e.g., Anzman & Birch 2009; Moffitt et al. 2011). It follows that an intervention targeting responsive parenting could have a positive impact on a myriad of outcomes in childhood and beyond, by scaffolding and promoting the development of early self-regulation. Accordingly, a preschool adaptation of a widely implemented, evidence-based preventive intervention designed to promote self-regulation and emotional competence improved social and emotional competence in low-income treatment groups (Domitrovich et al. 2007), while an adaptation intended to prevent childhood obesity showed evidence of improving children’s food and activity attitudes (Riggs et al. 2007). Similarly, an intervention intended to impact the behavioral domain through improved parenting was later found to improve weight status as well (Brotman et al. 2012). These examples support the idea that interventions that strengthen parenting and/or children’s regulation abilities could in turn benefit both physical and psychological health. Responsive parenting and child selfregulation may exemplify "third variables" underlying associations between comorbid problems, such as obesity, low academic achievement, and behavior problems (e.g., Puder & Munsch 2010). Targeting self-regulation abilities early, as they are emerging, could be particularly powerful. In addition to intervention effects on infant selfregulation, we also investigated mothers’ perceived parenting competence. Parents’ reports of their own parenting competence have been identified as a predictor of parenting practices (Sanders & Woolley 2005) and as a mediator of intervention effects on parenting (Dekovic et al. 2010). Thus, interventions that impact parents’ cognitions and emotions about their parenting abilities may impact their actual parenting practices and may positively impact the
Prev Sci (2014) 15:643–653
child and the family through these effects. The current intervention was designed in part to give parents tools to soothe their infants; these tools may be particularly useful for parents of negative infants and may increase their perceived parenting competence via successful interactions around infant fussing and crying (Bandura 1982, 1986). We assessed two aspects of parenting competence: mothers’ parenting self-efficacy and parenting satisfaction. The former is an instrumental subscale, reflecting feelings of proficiency in the parenting role, and the latter is an affective subscale, reflecting frustrations and motivations around parenting (Johnston & Mash 1989). The overarching goal of our Soothe/Sleep intervention was to provide soothing and sleeping strategies that would lead to a lower infant weight status via improvements in responsive parenting and infant self-regulation (Paul et al. 2011). In the current study, we hypothesized that effects would differ based on initial infant negativity, and that interactions between negativity and the Soothe/Sleep intervention would be consistent with a differential susceptibility model (Fig. 1a): in dyads with highly negative infants, the intervention would lead to greater observed infant selfregulation, less infant weight gain, and greater maternal parenting competence, compared to controls, and the intervention would not affect these outcomes in dyads with less negative infants. We used the criteria of Belsky et al. (2007) to test whether interactions supported a differential susceptibility model, with the understanding that other types of interactions could also have important implications for preventive interventions.
Methods Participants Mothers were recruited from the maternity ward of an academic medical center in Pennsylvania and were eligible if they intended to breastfeed and to follow-up with a University-affiliated primary care provider and if they were first-time mothers and English speaking. Other inclusion criteria were singleton birth and at least 34 weeks’ gestation. Dyads were excluded if the mother or infant had a morbidity that would affect postpartum care or infant sleeping or feeding or if the mother stayed in the hospital for more than 7 days postpartum. At study entry, there were 160 mother– infant dyads, and 110 dyads completed the 1-year study. Among these dyads, 51 % of infants were female, 90 % were non-Hispanic White, and the mean birth weight for gestational age percentile was 45.0 (SD=28.7). Mothers’ mean age was 27.1 (SD=4.7), and their mean prepregnancy BMI was 24.8, with a range spanning from underweight to obese (17.8 to 49.5). Most mothers were
Prev Sci (2014) 15:643–653
647
college-educated, with annual family incomes greater than $50,000 (Table 1). Dyads were assessed at birth and at infant ages 3 and 16 weeks and 1 year. Research nurses conducted home visits at ages 3 and 16 weeks, and at 1 year, dyads visited a University-affiliated clinical research center. Weight data were obtained from medical charts at age 3 for a subsample of participants (n=75). Because the study was designed to end at age 1, it was not possible to locate all participants to obtain their consent to extract follow-up medical chart data. Following recruitment, mother–infant dyads were randomized to a control group or a Soothe/Sleep intervention group. The Soothe/Sleep intervention was administered by nurses during the home visit at infant age 3 weeks. There was also an Introduction to Solids intervention that addressed the introduction of solid foods, which was administered to some families after infants were 16 weeks old (Paul et al. 2011). This intervention was not the focus of the current study, but statistical models were adjusted for a dichotomous variable indicating participation in this second intervention. Study groups did not differ on demographic variables or maternal BMI, although there were trend-level differences between the Soothe/Sleep intervention and
Table 1 Descriptive statistics Demographics Mean+SD or frequency Maternal education 65 % completed college Family income 72 % earned >$50,000 Maternal age 27.1±4.7 Maternal pre-pregnancy BMI 24.8±5.7 Infant sex 51 % female Infant race 90 % White Key study variables Daily min fussing/crying at 3 weeks Observed negativity at 1 year Observed regulation at 1 year Maternal self-efficacy at 1 year Parenting satisfaction at 1 year BMI-for-age z-scores at 1 year BMI-for-age z-scores at 3 yearsa
Mean+SD 87.4±59.9
Range 0–296.3
4.75+.64
3.25–6.33
.09+.12
0–.57
5.06+.59
3.43–6.00
4.77+.59
3.33–6.00
.30+.93
−3.97–3.57
.57+1.06
−1.56–3.66
Unstandardized variables are depicted in all tables and figures to facilitate interpretability. BMI-for-age z-scores indicate weight status relative to age- and sex-specific growth standards and were calculated using World Health Organization growth charts, as recommended by Grummer-Strawn et al. (2010) a
n=72, reflecting the number of participants with both height and weight data available at age 3. For all other variables, n=110
control groups on infant birth weight and sex, such that intervention group infants tended to be heavier at birth and were more likely to be male. There was also selective attrition, such that mothers who did not complete the study (n=50) were more likely to be single and non-White and tended to be younger and less educated than the mothers who participated for the entire year. Neither study group nor maternal BMI was related to participants’ likelihood of withdrawing from the study before completion. Unless otherwise indicated, the sample included the 110 dyads completing the 1-year study. The Human Subjects Protection Office of the Pennsylvania State University College of Medicine (Hershey, PA) approved all procedures, and mothers provided consent for the dyad’s participation. Dyads received monetary compensation and small gifts for their participation. Intervention The key messages of the Soothe/Sleep intervention were administered at a home visit when infants were 3 weeks old. Nurses explained and demonstrated alternate soothing techniques, besides feeding, to calm a non-hungry, fussing/crying infant. Dyads randomized to the Soothe/Sleep intervention also received a video (Karp 2006) that instructed parents on several ways to soothe a distressed infant, including swaddling, side or stomach position, shushing, swinging, and nonnutritive sucking. Additionally, parents were taught to emphasize day/night differences and were taught how to respond to night wakings to help the infant to self-soothe and return to sleep, as done by Pinilla and Birch (1993). In both the intervention and control groups, participants were given a standard infant parenting book (American Academy of Pediatrics 2010), and nurses answered questions about general infant care. Measures Infant Negativity Infant negativity was assessed using diary data and observation. Nurses provided and explained a set of 24-h infant behavior diaries at age 3 weeks, which were adapted from diaries used by Barr et al. (1988). Mothers completed the paper diaries, using colored markers to indicate their infant’s behavioral state (sleeping, awake/calm, fussing/crying, or feeding) every 15 min for 4 days. Average daily minutes of fussiness were calculated from these diary data at age 3 weeks and were used as an index of initial infant negativity. At the clinic visit at age 1 year, infants were observed during an adapted version of the LAB-TAB Toy Removal Task (Goldsmith & Rothbart 1996), as described by Stifter and Braungart (1995). During this procedure, the infant was allowed to play with an interesting toy for 2 min. Then, the toy was removed and placed out of the infant’s reach but
648
within sight. After 1 min, the toy was returned, and the infant was able to play for another minute. Trained research assistants coded infant reactivity and regulation during this procedure. The reactivity data contain the proportion of time spent showing mild, moderate, and high intensity negativity, as well as neutral and positive affect. Weighted intensity scores were calculated from the reactivity data and were used to adjust for observed negativity in analyses predicting self-regulation during the 1-min toy removal portion of the task: ((0*proportion of time spent exhibiting non-negative emotions)+(1*mild negativity proportion score)+(2*moderate negativity proportion score)+(3*high negativity proportion score)). Infant Regulation Infant regulation was coded during the Toy Removal Task described above. Trained research assistants coded regulatory behaviors in three passes. The regulatory behavior of interest was the proportion of time spent engaging in self-comforting behaviors, including hair twirling and finger sucking; these were coded during the second pass through the footage taken during third minute of the task (the part intended to elicit frustration). Infant Weight Status Infant weights and lengths were measured by research nurses at the clinic visit at age 1 year. Infant weights were measured using a calibrated Medela BabyChecker™ scale (McHenry, IL), and lengths were measured using the Seca 210 Mobile Measuring Mat for Infants and Toddlers (Hanover, MD). BMI-for-age z-scores, indicating weight status relative to growth standards, were calculated using the World Health Organization (WHO) growth charts (Grummer-Strawn et al. 2010). Birth weight was obtained from infants’ medical charts. After the study’s conclusion, heights (n=72) and weights (n=75) were collected from participants’ age 3 medical charts when possible. There were no significant differences when comparing infants with follow-up weight data (n=75) to the study completers without follow-up weight data (n=35) on birth weight (p=.25), BMI-for-age z-scores at age 1 (p=.18), or weight gain between birth and age 1 (p=.36). Residualized weight gain scores were calculated from the residuals of regression models where weight at age 1 was regressed on birth weight and where weight at age 3 was regressed on birth weight. These calculated variables allow an investigation of whether certain individuals experienced more weight change relative to the rest of the sample over time periods of interest. We also tested BMI gain scores, and results were similar to the reported results for weight gain. Maternal Parenting Self-efficacy and Satisfaction At an earlier visit, mothers were provided with a packet of surveys that included the Parenting Sense of Competence Scale (Gibaud-Wallston & Wandersman 1978), which assesses
Prev Sci (2014) 15:643–653
two subscales of parenting competence: parenting selfefficacy and satisfaction. Mothers received a phone call prompting them to complete the survey and to bring it to the clinic visit at infant age 1 year. Statistical Analyses Analyses were conducted using SAS version 9.2. The overall aim was to test whether initial infant negativity moderated effects of the Soothe/Sleep intervention on outcomes in three domains: infant self-regulation, infant weight outcomes, and perceived parenting competence. Ordinary least squares regressions were used to test whether negativity moderated intervention effects on observed infant selfregulation at 1 year, residualized infant weight gain scores from birth to 1 year and from birth to 3 years, as well as maternal parenting self-efficacy and parenting satisfaction at 1 year. Birth weight, infant sex, and Introduction to Solids intervention group were tested as covariates in all models. Additionally, observed negativity was a covariate in the model predicting observed regulation at 1 year, and feeding mode (breastfeeding versus formula-feeding, where dyads were considered breastfeeders if >80 % of milk feeds were breast milk at age 16 weeks) was included as a covariate in the models predicting weight gain. For all analyses, negativity and intervention group variables were standardized. If an interaction was statistically significant, we tested whether it showed evidence of differential susceptibility by investigating correlations and plots as recommended by Belsky et al. (2007): we (1) tested for the independence of negativity (the susceptibility factor) and intervention group (the predictor); (2) tested for a lack of association between negativity and the outcome; and (3) examined a plot of the interaction to verify that highly negative infants had better outcomes with the intervention and poorer outcomes without it, and that less negative infants were less affected by group assignment. In addition, post hoc tests of simple slopes were conducted to probe intervention effects for infants higher (+ 1 SD) versus lower (−1 SD) on negativity (Aiken & West 1991).
Results The means and standard deviations of variables of interest appear in Table 1. On average, infants fussed and cried for an hour and a half per day at age 3 weeks, which is less than the ∼2 h reported in previously published studies using middle-class samples and similar methodology (St. JamesRoberts & Plewis 1996). There are some differences between this study and previous studies that may account for these differences: all dyads in our study included first-time mothers, and a greater number of days of diary data were
Prev Sci (2014) 15:643–653
649
recorded. In both our study and comparable studies, the standard deviations around average daily minutes spent fussing/crying were large, indicating that there are individual differences in infant fussing and crying durations at this age and justifying the use of this variable as an index of temperament. Mean BMI-for-age z-scores at ages 1 and 3 were positive, showing that children were heavier than the average standard for their age and sex. This is consistent with national data (8 % of White infants and toddlers had a weight status greater than the 97.7th WHO weight-forlength percentile for their age and sex; Ogden et al. 2012) and illustrates a need for early prevention. Effects of the Soothe/Sleep Intervention on Infant Outcomes Infant negativity at age 3 weeks did not moderate the effects of the Soothe/Sleep intervention on infant self-regulation, but there was a significant main effect of the intervention, such that intervention infants showed higher observed regulation at 1 year (β=.23, p
