The Intervention Nurses Start Infants Growing on Healthy Trajectories (INSIGHT) study

Human biology evolved to promote survival in the context of food scarcity, biasing us to eat opportunistically and lay down fat stores essential for survival in times of scarcity. Within an ecological framework, parents protect their children from these perceived environmental threats to protect their health and development [55], and traditional feeding practices evolved during times when food scarcity and disease were threats to child health. Because loss of appetite is often a symptom of child illness and because adequate food and fluid intake is critical to child survival, traditional feeding practices are not contingent on child hunger, and feeding is the “default” response to infant distress [56‐58]. Accordingly, traditional feeding practices include strategies intended to soothe infant distress and promote intake, growth, and health such as use of feeding as the first response to crying and other distress, offering preferred foods, promoting intake in the absence of hunger, offering large portions, coercing children to eat beyond satiety, and providing palatable, preferred foods [52‐54].

Today in the US and many other countries worldwide, the obesogenic environment promotes an excessively positive energy balance, overweight, and obesity, and too much food has become the major environmental threat to child health, yet traditional feeding practices remain in place. Further, the use of feeding to soothe, especially with infants who are high in negativity and cry frequently, can exacerbate the effects of obesogenic environments, promoting overfeeding and overweight and fostering maladaptive eating behaviors affecting intake and obesity risk throughout the lifespan [54]. These eating behaviors include problems in controlling food intake: increased responsiveness to food cues [59‐61], reduced responsiveness to hunger and satiety, disinhibited eating and eating in the absence of hunger, and excessive weight gain and obesity [49, 62].

Currently, clinicians pay limited attention to obesity prevention during infancy [63]. While infant growth is monitored and most clinicians promote breastfeeding, suggest avoidance of infant cereal in bottles, and advise against early introduction of complementary foods and fruit juice, guidance aimed at early life obesity prevention often stops there. While initially appropriate, the focus for newborns and infants is typically on promoting adequate weight gain rather than preventing excessive weight gain. For example, immediately after birth mothers are typically instructed to wake their infants to feed at least every 3–4 hours for several weeks [64]. While this is necessary initial advice to help newborns regain weight lost shortly after birth, clinicians are far less consistent in instructing parents when to stop this practice and how to identify and rely on infant hunger and satiety cues to guide feeding. Typical anticipatory guidance in pediatric healthcare does not discourage the use of feeding to soothe for non-hunger related infant distress nor does it advise against feeding as a reward for positive behaviors later in infancy and childhood unless obvious overweight or obesity develops.

Current clinical care does focus on timing of introduction of foods – solids at 4 to 6 months, finger foods at 6 months, no cow’s milk until 1 year, etc. While guidance about what and when to feed infants is provided, little evidence-based direction is given on how to feed infants to promote subsequent healthy eating habits, including the aforementioned sensitive feeding styles, as well as feeding practices that promote acceptance of healthy foods and flavors in our obesogenic environment. However, interventions during infancy can produce healthier eating behaviors. For example, it is accepted that infants have innate preferences for sweet and salty tastes and are “neophobic”, rejecting new foods that are not sweet or salty [65]. Given this set of predispositions, typical infants will readily accept sweet and salty foods such as sweetened drinks and French fries. In contrast, healthy foods such as pureed vegetables, meats, infant cereals, and dairy products, which are not high in sugar or salt, are likely to be initially rejected. Birch and colleagues have conducted research demonstrating that infants typically need several opportunities to sample new foods before intake increases [66, 67]. The liking for complex flavors that are not dominated by sweet or salty tastes must be learned [65]. Interventions emphasizing repeated opportunities to try healthy, developmentally appropriate foods can have lasting positive effects on acceptance of healthy foods.

Current advice to parents to make healthy food choices for older infants and toddlers is failing to produce the desired outcomes and highlights the significant need for changes to clinical practice. Data from the Feeding Infants and Toddlers Study (FITS) revealed that unhealthy habits start early; energy intakes among infants and toddlers exceeded requirements by 20-30% [68]. In addition to consuming too much energy, children 4 to 24 months old ate significant amounts of developmentally inappropriate foods, high in energy density and sometimes deficient in key nutrients, and consumed too few of the foods that should form the basis of a healthy weaning diet [69, 70]. For example, in children aged 7 and 24 months, 18% and 33% consumed no servings of vegetables, respectively, during a given 24-hour period. Twenty-three percent of 7-month old and 33% of 24-month old children did not consume any fruits. By 15 to 18 months, French fries were the most common vegetable consumed. Clearly, parents of infants and toddlers need better direction on what to feed, how much to feed, and how to promote acceptance of a variety of healthy foods during the transition to the modified adult diet. Evidence that these outcomes can be promoted in a real-world setting is critical to transforming how parents view “picky eaters” while promoting acceptance of healthy diets.

Until recently, studies aimed at primary prevention of obesity through infancy-based interventions had not been conducted [44, 45]. However, over the past several years several trials have been completed with additional studies underway (Table 1). The first to demonstrate an effect on infant weight status was our pilot study, The SLeeping and Intake Methods Taught to Infants and Mothers Early in life study (SLIMTIME) [71]. In this trial, which informed the design of INSIGHT, we selected two promising interventions for obesity prevention based on Birch’s research, recruited first-time mothers who intended to breastfeed, and followed their infants delivered at a single center from birth to 1 year. Research nurses delivered interventions in the home using a 2 × 2 randomized, experimental design. The first intervention, “Soothe/Sleep,” began when infants were 2–3 weeks old. This intervention focused on promoting responsive feeding especially at night via techniques that trained parents to use feeding and other soothing approaches appropriately and selectively to calm a fussy infant [72]. The goal was to help mothers learn to discriminate their infant’s hunger from other distress cues, to use appropriate soothing responses for infant distress, and to prolong infant sleep duration. Our hypothesis was that by helping parents respond appropriately and contingently, this would reduce the use of feeding as the traditional “default” response to infant fussing and crying, and that reducing the use of “feeding to soothe” would reduce the risk of overfeeding and overweight. The second intervention, “Introduction of Solids,” focused on the transition to solids, providing information on which foods to offer or limit, information about portion sizes, and strategies for promoting liking and acceptance of healthy complementary foods. Specifically, it was based on Birch’s research showing the effectiveness of repeated exposure at reducing food neophobia and promoting infants’ acceptance of new vegetables [66, 67] and toddlers’ acceptance of new table foods. 110 participants completed the year-long study.

In comparing “Soothe/Sleep” intervention infants vs. control, results showed that breastfed intervention infants slept significantly longer at night 16 weeks after birth (p = .04). Compared with controls, the intervention also reduced total number of daily feedings (p = .008) and nocturnal feedings (p = .003) for breastfed infants. Consistent with our framework, the results suggest that our intervention helped parents learn to be responsive while calming infants without feeding in the absence of hunger, thus promoting parenting competence. It also appeared to promote development of infant self-regulation by allowing infants opportunities to self-soothe [90, 91]. Third, given the consistently demonstrated relationship between short sleep duration and overweight, obesity and higher body fat for children of all ages including infants [9, 20, 43, 92‐98], an intervention that effectively lengthens sleep duration during infancy is potentially preventive, and may have both short- and long-term effects, given that infant sleep difficulty predicts later sleep problems [99, 100].

The “Introduction of Solids” intervention focused on the timing of and effective approaches for introducing solid foods. The intervention delayed introduction of solids, increased acceptance of vegetables upon introduction, and increased acceptance of healthy, unfamiliar foods (hummus, cottage cheese, or yogurt) at the age 1 year laboratory visit. Based on intervention-blinded coding of videos, only 10% of intervention infants rejected an unfamiliar food at age 1 year compared with 25% of those in the control group [71].

In addition to improving secondary outcomes related to sleeping and feeding, participants receiving both study interventions had significantly lower weight-for-length percentiles at 1 year (43rd percentile) than those in other study groups. The findings suggest both interventions, the first affecting early sleep, soothing, and feeding frequency and the second affecting the infants’ reactions to the introduction of solid foods, were required to have a significant effect on weight status, and therefore a multi-component intervention was selected for the larger INSIGHT trial.

INSIGHT is a two-arm randomized, controlled trial that involves nurses delivering interventions to first-time parents and their infants at four home visits in the first year after birth followed by clinical research center visits at ages 1, 2, and 3 years (Figure 1). Following completion of informed consent, research staff collected data from the newborn medical record, and participating mothers completed baseline demographic questionnaires. Because we are interested in the effect of the intervention program on formula fed and breastfed infants, mothers were contacted via telephone 10 to 14 days following childbirth, and randomized to a study group with stratification performed based on mothers’ intended feeding mode (breastfeeding or formula) and sex-specific birth weight for gestational age (<50th percentile or ≥50th percentile) [101]. This study was approved by the Penn State College of Medicine’s Human Subjects Protection Office and registered at http://​www.​clinicaltrials.​gov (NCT01167270) prior to enrollment of the first participant.

All newborns delivered at the Penn State Milton S. Hershey Medical Center in Hershey, Pennsylvania were screened for participation (Figure 2). Eligible mother-infant dyads for this trial included full-term (≥37 weeks gestation), singleton newborns delivered to English-speaking, primiparous mothers ≥20 years of age residing within 50 miles of the center. Mother-newborn dyads were excluded if there was a plan for the newborn to be adopted or move from Central Pennsylvania within 3 years, if a prenatal ultrasound demonstrated evidence of intrauterine growth retardation, if the newborn’s birth weight was <2500 grams, or if either the mother or newborn had significant health issues that would affect study participation.

The primary outcome for INSIGHT is BMI Z-score at 3 years. A 0.67 difference in BMI Z-scores represents the distance between major centile lines displayed on infant growth charts. This difference is clinically meaningful, and upward percentile crossing of this magnitude has been the operational definition frequently used for rapid weight gain [102]. For this reason we have powered the study to detect a 0.67 difference in BMI Z-scores between the safety control and the parenting intervention groups within each of the feeding modes (intent-to-breast feed or formula feed). To detect this difference with 90% power and a 5% Type 1 error rate, 276 participants are required. This includes an anticipated 30% attrition rate. INSIGHT’s initial sample size slightly exceeded this 276 with a total of 316 mother-infant dyads enrolled, 291 dyads randomized, and 279 participants completing the 3–4 week study visit where the interventions begin. This cohort of 279 will be considered the study cohort for outcomes and analyses.

The parenting intervention uses a responsive parenting framework with obesity prevention messages delivered at each visit that correspond to four infant/toddler behavior states: Drowsy, Sleeping, Fussy, and Alert and Calm (Figure 3). Within the Alert and Calm category are two sub-categories, Active, social play and Feeding.

INSIGHT’s control group receives a developmentally appropriate home safety intervention also delivered by visiting nurses. The home safety visits are designed to be equal in length and intensity to the parenting intervention visits and to avoid messages that could impact energy balance. The home safety intervention also is designed within the framework of the four behavioral states. Within the Drowsy and Sleep domains, prevention of Sudden Infant Death Syndrome is initially the focus with carbon monoxide poisoning prevention and bedroom childproofing topics covered later. Within the Fussy category, strategies to prevent Shaken Baby Syndrome and child abuse are discussed initially followed by avoidance of distracted driving, treatment of fever and other first aid remedies, and avoidance of physical force with discipline. Food safety is a repeated theme in the Alert and Calm – Feeding sub-domain beginning with breast milk and formula handling and storage, progressing to baby food handling, food allergens, choking hazards, and safe storage of food sent to childcare. High chair and stove safety issues are also discussed. For the Active, Social Play component, the first topics covered include fire, bath, and car seat safety. Prevention of falls, poison prevention, furniture safety, and toy safety are all covered within the first year. Gun safety and seasonal topics (insect repellant, sunscreen, trick-or-treating, holiday decorations) are then covered with pedestrian safety, booster seat transition, and hand washing.

To assess intervention impacts on both the primary outcome and intermediary behavioral processes, INSIGHT data collection includes measures divided into the following categories: Anthropometrics and Biological Specimens, Infant/Child Behavior, Parenting, Maternal Psychosocial Variables and Behavior, Family Context, and Background/Demographics/Covariates. As detailed in Table 2, many of the measurements are assessed repeatedly.

Beginning with Sir Francis Galton’s English Men of Science published in 1874, researchers have examined relationships between birth order and outcomes related to health, achievement, behavior, and intelligence [137‐139]. Yet despite observations over the past 140 years regarding birth order, few have prospectively evaluated differences in parenting between successive siblings as a source of the disparities between first- and second-born children. What is established is that while mothers spend more time interacting with their first-born during infancy [140, 141], they are more likely to use restrictive or coercive parenting strategies as first described by Lasko in 1954 [142]. Hilton later found that first-time mothers are significantly more interfering, extreme in response, and inconsistent with parenting response than mothers with their later born children [143]. These findings suggest that parents are less likely to use responsive parenting practices with their first-borns than with later-borns.

A recent review of studies examining the effect of birth order on parenting by Kaley et al. [144] reports that evidence is scant and that to date, no studies have prospectively examined differences in parenting of first- and second-born siblings within the same family beginning at birth. However, the review identified several potentially modifiable postnatal factors affecting infant obesity risk: sleep duration, feeding style, and parental regulation of distress.

Capitalizing on the infrastructure and extensive data collection occurring as part of INSIGHT, a second-related project, SIBSIGHT, adds two major pieces by enrolling second born siblings of INSIGHT participants and collecting genetic specimens from both siblings and their parents. Specifically, this translational research is a) prospectively evaluating obesity-related parenting similarities and differences as well as weight-related outcomes between first and second-born siblings, b) exploring how genetic differences among siblings that are associated with appetite, temperament, and obesity susceptibility affect parent–child interactions, degree of responsive parenting, and weight status, and c) determining whether INSIGHT study intervention carryover effects occur among families participating in the observation-only second-born child evaluation.

Data from diverse cultures have shown that first-born children have a higher risk for obesity [145‐152] despite the fact that pregnancy related risk factors for childhood obesity (high pre-pregnancy body mass index, high gestational weight gain, occurrence of gestational diabetes, high birth weight) are more common during pregnancies with second-born children. This suggests that postnatal factors related to parenting are the cause of the disparity between first and second born children’s obesity risk. Using the conceptual framework of responsive parenting, it can be hypothesized that mothers have improved responsiveness and more appropriate caretaking behaviors due to the experience gained with their first child. This hypothesis will be tested by comparing the >100 second-born siblings that can be expected to be born during the funding period with their older siblings with the additional goal to explore how differences in genetic susceptibility to obesity and observed differences in appetite and temperament moderate associations between responsive parenting and weight status at age one year. Data collection for second born children is similar to that demonstrated for the first born in the first year after birth with added aims assessing a) the relationship between infant oral and gut microbiome with weight outcomes and b) the role of epigenetics on obesity development as determined from placenta, cord blood, and peripheral blood obtained at 1 year.