Primary Care Interventions for Early Childhood Development: A Systematic Review
- aInstitute for Healthcare Delivery and Population Science, Baystate Health, Springfield, Massachusetts;
- bDepartment of Pediatrics, University of Massachusetts Medical School–Baystate, Springfield, Massachusetts;
- cSchool of Medicine, Boston University, Boston, Massachusetts; and
- dDivision of General Pediatrics, Boston Medical Center, Boston, Massachusetts
Dr Peacock-Chambers conceptualized and designed the study, participated in study selection, conducted the analyses, drafted the initial manuscript, and revised the manuscript; Ms Ivy participated in the study selection, data collection, and analyses, drafted sections of the initial manuscript, and reviewed and revised the manuscript; Dr Bair-Merritt conceptualized and designed the study and reviewed and revised the manuscript; and all authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.
CONTEXT: The pediatric primary care setting offers a platform to promote positive parenting behaviors and the optimal development of young children. Many new interventions have been developed and tested in this setting over the past 2 decades.
OBJECTIVE: To summarize the recent published evidence regarding the impact of primary care–based interventions on parenting behaviors and child development outcomes; to provide recommendations for incorporation of effective interventions into pediatric clinics.
DATA SOURCES: A literature search of PubMed and PsycINFO was conducted from January 1, 1999, to February 14, 2017.
STUDY SELECTION: Publications in which primary care–based interventions and reported outcomes regarding the child’s development or parenting behaviors associated with the promotion of optimal child development are described.
DATA EXTRACTION: Forty-eight studies in which 24 interventions were described were included. Levels of evidence and specific outcome measures are reported.
RESULTS: Included interventions were categorized as general developmental support, general behavioral development, or topic-specific interventions. Two interventions resulted in reductions in developmental delay, 4 improved cognitive development scores, and 6 resulted in improved behavioral intensity or reduction in behavioral problems. Interventions used a variety of theory-based behavior change strategies such as modeling, group discussion, role play, homework assignment, coaching, and video-recorded interactions. Three interventions report the cost of the intervention.
LIMITATIONS: Community or home-based interventions were excluded.
CONCLUSIONS: Although several interventions resulted in improved child development outcomes for children aged 0 to 3 years, comparison across studies and interventions is limited by use of different outcome measures, time of evaluation, and variability of results.
- BB —
- Building Blocks
- CBCL —
- Child Behavior Checklist
- CFD —
- Care for Development
- ECBI —
- Eyberg Child Behavior Inventory
- HS —
- Healthy Steps
- IY —
- Incredible Years
- PCIT —
- Parent-Child Interaction Therapy
- POC —
- Play with Our Children
- PP —
- RCT —
- randomized controlled trial
- ROR —
- Reach Out and Read
- ROR + M —
- Reach Out and Read plus Mathematics
- SDP —
- Sit Down and Play
- Triple P —
- Positive Parenting Program
- VIP —
- Video Interaction Project
The first 3 years of life are a critical period for child brain growth and development, with the potential to impact later social, economic, and health-related quality of life.1 Pediatric primary care, which includes health supervision and anticipatory guidance, offers an important entry point for the promotion of optimal child development in the United States because it is a universal service with frequent encounters in the first 3 years of a child’s life.2,3 In a systematic review by Regalado and Halfon4 of publications between 1979 and 1999, the authors summarized developmental assessments, anticipatory guidance, and specific interventions used to promote development of children aged 0 to 3 years in primary care settings. They comment on the efficacy of anticipatory guidance and problem-focused counseling interventions as well as the challenges and barriers to implementing such interventions in practice. Importantly, the authors found that targeted skill-building interventions that were focused on improving parent-child interaction,5 or specific parenting behaviors (such as book sharing,6 coping with infant colic,7 sleep training,8 or non-harsh discipline9) proved to be the most effective.
New interventions to promote the optimal development of young children continue to be developed and tested. Additionally, in 2012, the American Academy of Pediatrics called for the pediatric community to “catalyze fundamental change in early childhood policy and services.”3 This declaration underscored the need for new, more effective strategies to increase the impact of primary care on promoting optimal development for young children in the United States. Taking this statement along with the rapidly changing landscape of US pediatric primary care systems, including the rise of the medical home model10 and new models of well-child care delivery,11 a comprehensive and current review of these interventions is needed. The authors of such a review must summarize the impact these interventions have on child development outcomes, building on a previous review focused on positive parenting outcomes.12 Furthermore, scientists and clinicians could benefit from a synthesis of available evidence-based interventions to inform areas still in need of future study. Therefore, our objective in this systematic review was to examine the interventions implemented in connection with primary care to promote optimal child development for children aged 0 to 3 years from 1999 to 2017 and to summarize the efficacy of these interventions on child and parent level outcomes.
This systematic review was conducted by following the PRISMA guidelines13 to identify interventions delivered in connection with primary care setting for the promotion of optimal early development of children aged 0 to 3 years. For the purpose of this review, pediatric primary care or pediatric clinics were defined as family medicine, pediatric, or public health clinics (outside of the United States) providing preventive care to children. Interventions were defined as education, counseling, or other provider-parent engagement beyond screening for developmental delay. Parents could include nonbiological parents, grandparents, or caregivers, but henceforth we refer to these caregivers as parents. Our inclusion criteria required that English, peer-reviewed publications: (1) examined interventions for the promotion of optimal child development for children aged 0 to 3 years, (2) described delivery of the intervention within or associated with a primary care setting, and (3) reported outcomes regarding the child’s development (cognitive, behavior, or social-emotional) or parenting behaviors associated with the promotion of optimal child development (eg, book sharing, positive discipline, or parental sensitivity using validated measures). We excluded review articles and articles that targeted children with developmental diagnoses (eg, autism, attention-deficit/hyperactivity disorder, oppositional defiant disorder) or extreme prematurity (<28 weeks’ gestation). We also excluded studies solely of parents with known substance abuse or depression because these mental health illnesses frequently require additional targeted treatment. We excluded studies that only reported outcomes regarding child sleep. Publications were considered from any country if they met the above listed criteria.
PubMed and PsycINFO were searched by E.P-C. for a selection of articles published from January 1, 1999, through February 14, 2017, using these key words: (“child development” OR “parenting”) AND (“pediatrics” OR “primary care”). In addition, we searched references of included articles and related review articles. The final list of included articles was reviewed with 2 expert child development consultants not affiliated with the study: 1 developmental behavioral pediatrician with more than 30 years of clinical and intervention development experience and a second developmental behavioral pediatrician that has written extensively about child developmental assessments and promotion in clinical settings.
Initial search terms were intentionally broad; therefore, 1 investigator (E.P-C.) screened the initial list of titles to exclude clearly irrelevant studies. Two investigators independently screened abstracts of all potentially relevant titles (E.P-C. and K.I.) using inclusion and exclusion criteria listed above to identify studies for full text review. The investigators resolved disagreements by consensus, reviewing full-text articles and involving a third investigator as needed.
Data Abstraction and Levels of Evidence
Data were abstracted from full-text articles by K.I. by using a structured form that included information on intervention delivery (content, frequency, duration, person implementing, location, specific developmental focus, and cost), study design, child age, sample size, study population, outcome measures (child or parent outcomes), and findings. Abstracted data were verified and checked for consistency by a separate abstractor (E.P-C.). Two reviewers (E.P-C. and K.I.) independently assessed the levels of evidence of the studies using a clinically relevant quality rating scheme modified from the Oxford Centre for Evidence-Based Medicine Levels of Evidence.14 Levels of evidence are classified as follows: level 1, properly powered and conducted randomized clinical trials; level 2, well-designed controlled trial without randomization, prospective comparative cohort trial; level 3, case-control studies, retrospective cohort studies; level 4, case series with or without intervention, cross-sectional study. Disagreements regarding level of evidence scores were resolved through discussion and consensus.
Description of Study Selection
Initial searches yielded 7605 titles from all search engines. Of these, 528 abstracts (411 from PubMed and 117 from PsychInfo) were retained, with an additional 40 abstracts identified from reference lists. After the removal of 55 duplicate articles, 513 abstracts were reviewed for inclusion and/or exclusion criteria (Fig 1). Four-hundred and twenty-five articles were excluded during abstract review, resulting in the full-text review of 88 articles. Forty articles were excluded on full-text review, yielding 48 articles in which 24 interventions were described that met all inclusion and exclusion criteria.
Interventions were conducted in 12 different countries: United States, Australia, China, Holland, Canada, Norway, Turkey, Chile, Jamaica, Antigua, St. Lucia, and Iran. Interventions identified in this review could be grouped in 3 broad categories on the basis of the described intervention “focus” as summarized in Tables 1 through 3: (1) general developmental support, such as language, social, gross and fine motor development (9 interventions, 20 studies); (2) general behavioral development, such as managing negative child behaviors through developing positive disciplinary strategies (8 interventions, 19 studies); and (3) specific developmental topics (topic specific), including infant colic and reading aloud to children (7 interventions, 9 studies). Descriptions of interventions are found in Tables 1, 2, and 3, which include the intervention components, the “dose” of the intervention, the developmental focus of the intervention, who administered the intervention, the cost when available, and the connection to primary care.
General developmental interventions included the Video Interaction Project (VIP),17–23 Building Blocks (BB),19–21,23 Healthy Steps (HS),24–28 HS plus PrePare (PP),15,29 Care For Development (CFD) Intervention,16,30 Touchpoints,31 Play with Our Children (POC),32 Parenting Intervention,33 and Sit Down and Play (SDP).34 VIP and HS were the 2 most intensive interventions involving additional meetings with child development specialists at each well-child visit from birth to age 3 years. VIP sessions included videotaping the parent and child during a play interaction followed by review and coaching of the interaction with the specialist at a cost of $150 to $240 per child per year. BB provided a monthly newsletter and age-specific toy to families at a lower cost ($75) and was studied as part of a 3-arm randomized controlled trial (RCT) with VIP. HS clinic sessions offered parents an opportunity to discuss developmental concerns with a child development specialist as well as home visits, phone support, written material, and parenting groups. PP was an addition to HS providing 3 home visits during pregnancy. HS cost between $400 and $933 for 11 well-child visits and 2 home visits over 2.5 years. Touchpoints included similar services (clinic and home visits, phone support) delivered from birth to 18 months of age by a “parent coach” with college-level education in child development. CFD, developed by the World Health Organization and the United Nations Children’s Fund and officially called the Care for Child Development intervention, provided a picture card and guides for clinicians to share with parents to discuss the importance of developmentally appropriate play, home-made toys, communication, and reading. POC offered group sessions by “community monitors” as well as home visits and in-depth interviews from birth to age 4 years. The Parenting Intervention and SDP were 2 brief interventions delivered in waiting rooms for 10 to 15 minutes by community health workers or paraprofessional volunteers.
General behavioral interventions included Incredible Years (IY),35,36 Positive Parenting Program (Triple P),37–45 Parent-Child Interaction Therapy (PCIT),46 Universal Parenting Program (also called Toddlers without Tears),47,48 Webster-Stratton,49,50 Family Foundations,51 ezParent,52 and PriCARE.53 Triple P was the most widely disseminated and studied program, which included 9 publications from 4 different countries. Triple P had 5 levels of intervention intensity designed to match the severity of the child’s behavioral problems (Table 2). With increasing Triple P levels, the total number of sessions as well as length of sessions increased. The Webster-Stratton intervention, IY, and PriCARE were among the more intensive group-based parenting programs involving ∼10 parents in 6 to 10 sessions delivered by child psychologists or pediatric staff, including role play, homework, and group discussion. PriCARE adapted a trauma-informed program called Child-Adult Relationship Enhancement54 on the basis of principles and techniques derived from PCIT for the primary care setting. PCIT and the Universal Parenting Program required fewer sessions (3–4) ranging from 1.5 to 2.5 hours. Family Foundations included psycho-education for coparents to resolve conflicts between parents beginning prenatally. EzParent was the only behavioral intervention that was entirely delivered via an electronic device, adapting a parenting group curriculum into 6 multimedia interactive modules.
The topic-specific interventions included 4 interventions for colic (Family-Centered Treatment,55 Baby Business Program,56 The Happiest Baby,57 and Infant Massage58) and 3 interventions to promote reading aloud (Literacy Promotion Intervention,59 Reach Out and Read [ROR],6,60,61 and Reach Out and Read plus Mathematics [ROR + M]62). The 4 colic interventions varied greatly in intensity from development of individualized family treatment plans designed by a pediatrician and mental health specialist (Family-Centered Treatment) to multimedia videos with optional discussion (Baby Business Program and The Happiest Baby) as well as infant massage. ROR is a well-studied program to promote child literacy during well-child visits, also previously described in Regalado and Halfon’s4 review. The authors of 2 additional studies completed after 1999 evaluated the intervention through a quality improvement evaluation across 10 states in the United States.6,60 The authors of a third study modified the intervention to teach specific math topics.62 A literacy intervention similar to ROR evaluated the impact of reading aloud on receptive and expressive language development.59
The authors of the studies included in this review recruited participants from varying socioeconomic backgrounds, ethnicities, and nationalities. VIP involved primarily low-income and Spanish-speaking participants in the United States, a large proportion of which did not receive education beyond the seventh grade. HS studies included 15 primary care sites across the United States.25–27 Three interventions (CFD, POC, and the Parenting Intervention) were studied solely in low- or middle-income countries. In contrast to the general developmental interventions, behavioral interventions tended to target participants on the basis of behavioral concerns rather than demographic factors. The researchers for IY,35,36 Webster-Stratton,49,50 and PriCARE53 studies enrolled children that scored above a certain threshold on behavioral assessments or whose parents reported concerns regarding their child’s behavior. Triple P tailored the intervention intensity on the basis of the severity of the child’s behavioral problems.
Levels of Evidence
General developmental interventions were primarily conducted as quasi-experimental study designs, with the exception of the VIP and the Parenting Intervention studies in which RCTs were employed (Table 4). Sample sizes ranged from 50 at a single site31 to >3000 study participants in 10 sites.25,26 All of the general behavioral interventions were evaluated by at least 1 RCT (Table 5). For topic-specific studies, RCT study design was employed in colic interventions, whereas observational cohorts or cross-sectional comparisons with historical controls were used in the literary studies (Table 6). All Levels of Evidence scores are reported in Tables 4 through 6.
General Developmental Support
Among the general developmental interventions, cohorts enrolled in HS and VIP experienced fewer cases of developmental delay compared with control groups at 12 months28 and 3 years, respectively.17,18 The effects of VIP on child development were more pronounced among children of mothers with seventh to 11th grade educations in 1 study17,18 and for mothers with a literacy level of ninth grade or higher in a second study.19 VIP also resulted in decreased hyperactivity and externalizing behaviors.23 HS and PP resulted in a reduction in child aggressive behavior at 30 months of age measured by the Child Behavior Checklist (CBCL).15,29 No difference in child behavior was detected with HS alone. CFD resulted in improved adaptive, language, and social development 6 months postintervention.30 Touchpoints was associated with significant improvement in child language development (expressive and receptive) measured by the MacArthur Communicative Development Inventories at 16 months of age.31 The Parenting Intervention was associated with significantly improved cognitive development (Griffith Mental Development Scale), but not specifically language development.33
Multiple studies reported significant changes in both parenting behaviors and psychological well-being. Parents enrolled in the VIP,19 BB,19 HS,25,26 CFD,16 and SDP34 interventions engaged in educational activities (such as reading) at significantly higher levels compared with control groups. The overall quality of the home environment improved among families enrolled in VIP and BB at 6 months of age.19 HS,24 Touchpoints,31 and POC32 were associated with greater parental sensitivity and higher quality parent-child interactions. Parents enrolled in VIP reported significantly reduced stress and depressive symptoms compared with controls.22 One study in which HS and PP were evaluated revealed an increase in parental depressive symptoms and a decrease in parenting satisfaction among the intervention group at 30 months.15,29
General Behavioral Development
Four interventions resulted in decreased behavioral intensity between 2 and 24 months after the intervention: IY,35,36 Triple P levels 3 and 4,38,40–45 PCIT,46 and PriCARE.53 Children engaged in the Webster-Stratton intervention exhibited fewer behavioral problems at 6 months, but this difference did not persist at the 12-month follow-up.49,50 The Universal Parenting Program47,48 and ezParent52 did not result in any difference in child behavior between the intervention and control groups.
A number of the behavioral interventions also affected parenting behaviors and psychological outcomes. IY resulted in improved positive parenting behaviors 12 months postintervention,35,36 and multiple interventions (IY, Triple P level 3 and 4,38,40,42 PCIT,46 and PriCARE53) led to reductions in the use of harsh discipline. Parental satisfaction,35,40,41 self-efficacy,40,41,43,44 or confidence46 improved significantly with 3 different interventions, and marital health improved w,ith Triple P40–42 and Family Foundations.51 Although in most studies assessed parent mental health, only the Webster-Stratton49,50 intervention and Family Foundations51 were associated with a reduction in depressive symptoms beyond 10 months postintervention.
Specific Developmental Topic
All 3 interventions designed to address infant colic revealed improvement in colic symptoms in short-term follow-up, but they resulted in mixed results beyond 2 months postintervention (Table 6). The Baby Business Program was also associated with lower maternal depressive symptom scores at 6 months postintervention.56 In contrast, The Happiest Baby video was associated with higher parental stress scores at 12 weeks postintervention.57
Three interventions in which the promotion of early literacy was evaluated had relatively consistent results. Three ROR studies revealed a significant increase in the number of parents reporting reading aloud as their favorite activity with their child or increased frequency of reading on a weekly basis in cross-sectional samples or quasi-experimental studies.6,60,61 These effects were stronger among non-English speaking families60 and families with less than a completed high school education.6 The ROR + M intervention resulted in greater mathematics engagement by parents as well.62 The Literacy Promoting Intervention by High et al59 was associated with increased receptive and expressive vocabulary.
In the past 15 years, multiple new interventions have been developed and tested in connection with primary care settings with the goal of improving the developmental trajectories of young children. Twenty-four interventions are included in this review, with data from 48 studies, primarily evaluated by RCT or quasi-experimental study design. A number of these interventions were evaluated across multiple cities and countries in studies with >3000 participants.
The evolution of interventions to promote development of young children in primary care settings is evident when comparing our findings to those of the systematic review by Regalado and Halfon4 in 2001. As documented in the previous review, educational interventions were focused on positive parent-child interactions and improved parental knowledge about child development but did not change parenting behavior or child development outcomes.63–65 In contrast, 2 interventions explicitly designed to improve parental confidence and competence through skill building66 and coaching5 enhanced the quality of parent-child interactions in small trials (<50 participants).
The interventions summarized in this review continued this trajectory of incorporating new strategies to enhance parental skill building and parent-child interactions supported by behavior change theory.67,68 These strategies included vicarious learning from peers, role play, homework, feedback through coaching, videotaped observation, and identification of parenting strengths. In some studies, the parental factors that could explain mechanisms of action were measured and included parental resilience,31 satisfaction,15,35,38–40,45 self-efficacy,15,35,38–40,45,46,52 and stress.17,18,32,38,44,45,52 The authors of studies of behavioral interventions also frequently explored the impact on parental mental health as a potential mediating factor.42,47–51
The primary challenge in synthesizing and drawing firm conclusions about the efficacy or effectiveness of specific interventions, however, lies in the heterogeneity of outcome measures and reported findings. Among the general developmental studies, a different child outcome measure was used in almost every study, and 3 studies16,32,34 had no child outcome measures as part of the evaluation, making comparisons across these studies difficult. Even among studies of the same interventions, the results varied in terms of the time when evaluations occurred and the specific outcome measures that demonstrated improvement (eg, prevention of developmental delay versus behavior problems in VIP and HS studies).18,23,26,28 The behavioral development studies were more consistent in their use of the CBCL or Eyberg Child Behavior Inventory (ECBI) as outcome measures of child behavioral problems.
Greater consensus on which child outcomes are most important from a clinical perspective and which validated tools are optimal for assessment of these outcomes is needed, particularly for general developmental interventions. Other specialties, such as mental health, face similar challenges69 and may serve as examples for future standardization of quality and outcome measures.70,71 With respect to general developmental outcomes, measures of language development may be particularly important given the disparities that exist in this developmental domain72,73 and the association with later academic achievement.74,75 Studies with longer follow-up periods, such as studies of home visiting or preschool programs,76,77 may help identify measures in early childhood that predict long-term health and academic success. The behavioral development studies showed more consistent impact on the clinically relevant measures of behavioral intensity and behavior problems.
Although the factors related to the mechanism of action of interventions were assessed in many studies, the variability of results limits our ability to understand precisely why some interventions work over others, who benefits from interventions, and what may be the common pathway by which different interventions impact the same outcomes. This basic understanding is necessary to inform future policy and services.3 The collection of VIP studies provides an example research strategy to thoroughly study an intervention. The studies identified families that benefited most (on the basis of education level),18 potential mechanism of action (mediators including responsive parenting and maternal depression),20 and explored clinically significant outcomes such as language, harsh discipline, and social emotional development.21,23
In addition to examining the individual mechanism of action of the interventions, researchers for future studies should attempt to identify key components that may be common across multiple interventions. Drawing again from the mental health literature, the examination of key components should include intervention content as well as factors that impact the process of intervention delivery (eg, participant and provider attitudes, specific types of interactions, and provider approach to influencing behavior change).78 Common elements of evidence-based psychologic treatments for adult anxiety and depression, for example, include altering cognitive appraisal, modifying emotion-driven behavior, preventing emotional avoidance,79 as well as demonstrations of warmth and empathy, clear explanations, focused discussion on practical concerns, and well-organized sessions.78 Identification of common key components of developmental interventions could have implications for the feasibility of dissemination of these interventions into community settings where the majority of children receive primary care.
Several notable gaps in the literature limit the feasibility of replicating individual interventions in community settings. First, the cost of many of the identified interventions is not well described. VIP and HS reported relatively low costs compared with programs such as home visiting or Early Head Start, although both required a child development specialist.20,26,27 The descriptions of trainings are similarly limited. Most of the interventions were delivered by child development specialists or health care professionals that likely bring additional skills and knowledge to interactions with parents. Access to child development specialists is limited in rural settings; thus, understanding the specific skills needed to conduct the intervention is critical. Chang et al33 provided an example of a detailed description of paraprofessionals training that could be used in resource-limited settings.
Although multiple studies showed positive change in a number of different child development and parental outcomes, the heterogeneity of results limit the clear distinction between effective and noneffective interventions. To identify the services and policies that will optimally promote the development of young children, we need a clear understanding of the end point we are collectively trying to achieve. We need greater consensus on the best measures of these outcomes, and we need to understand why interventions work and who benefits most. When considering implementation of these interventions in community settings, health professionals must consider the availability and expertise of their staff and the cost of training and implementation. Examples of highly targeted interventions delivered by paraprofessionals may provide some direction for future dissemination strategies.33,34
This review has several limitations. Only articles published in English were reviewed. In addition, the inclusion criteria requiring a connection to primary care limited the focus to US populations and countries with similar health systems. Initially, we considered limiting our review to studies conducted in the United States given the significant differences between health care systems around the world. Ultimately, we sought to include international studies because multiple interventions were studied in the United States and internationally, and interventions designed for settings with varying levels of resource allocation may also be relevant to lower-resource or rural settings in the United States. Our review may have missed eligible studies for various reasons. Although our search terms were intentionally broad, other terms used to describe similar interventions may have resulted in missed studies. Additional studies may have been missed or excluded because of ambiguity with respect to the age of children enrolled or the connection to primary care. Interventions conducted without a connection to primary care were excluded even if other primary care–based studies of the same intervention were included. We also excluded studies targeting specific populations such as extremely premature infants and parents with diagnosed depression or substance abuse disorders because these populations frequently require targeted therapies or treatment with medication. Finally, the Levels of Evidence system used to summarize and compare the study designs across interventions is not a definitive indicator of the quality of the study because it does not assess bias related to blinding, instrument validation, or observer versus self-report measures. The decision to use this system was based on feasibility given the variety of study designs used and the variability of reported factors that impact study quality.
Several interventions delivered in primary care settings can positively impact developmental and behavioral trajectories of children aged 0 to 3 years. These interventions involve the child’s parents and frequently a team of health professionals. Common measures of child development outcomes are needed to compare efficacy across studies. Further exploration of why certain interventions work and who benefits from interventions could inform future policy to make these services more widely available.
- Accepted September 11, 2017.
- Address correspondence to Elizabeth Peacock-Chambers, MD, MSc, Department of Pediatrics, Baystate Medical Center, 3601 Main St, Springfield, MA 01199. E-mail:
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Supported by a Health Resources and Services Administration T32 grant HP10028 (I09940000730) from June 2014 to June 2015.
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
COMPANION PAPER: A companion to this article is available online at www.pediatrics.org/cgi/doi/10.1542/peds.2017-3136.
- Copyright © 2017 by the American Academy of Pediatrics