Early parenting support on child development through age 6: The Smart Beginnings model

Abstract

Background and Objectives:

Despite longstanding efforts to design, implement, and study parenting interventions early in life to address disparities in school readiness, gaps remain related to understanding their long-term effects and pathways of influence on child development. Here we describe sustained impacts at child age 6 of the innovative, tiered birth-3-year Smart Beginnings (SB) model.

Methods:

We performed a single-blind, two-site randomized clinical trial (RCT) of the SB model. SB integrates PlayReadVIP, a universal, pediatric primary care-based program, and Family Check-Up (FCU), a targeted secondary home-based parenting intervention. Mother-infant dyads (N=403) were randomized at birth to standard pediatric care or the SB model. In line with SB’s theory of change that supporting parents will promote their children’s development, single and serial mediation pathways evaluated intervention effects of SB on age 6 child academic skills through parental cognitive stimulation at age 2 and child academic functioning at age 4.

Results:

We found significant single and serially mediated indirect effects of SB on academic outcomes through parental cognitive stimulation in toddlerhood and preacademic skills in preschool. The total indirect pathways were positive and statistically significant for all academic outcomes at age 6, including receptive vocabulary (Effect Size [ES]=.04, p=.04), oral comprehension (ES=.05, p=.04), letter-word recognition (ES=.04, p=.04), phonemic decoding (ES=.04, p=.04), and applied problems (ES=.05, p=.04).

Conclusions:

Findings build on the demonstrated scalability of the SB model, support the cumulative process of academic functioning in childhood, and offer a promising model to address disparities early in life.

Introduction

Poverty-related disparities in early brain development, school readiness, and school performance are well documented,¹ with cascading consequences across the lifespan.² Family stress³ and investment⁴ models of poverty posit that proximal family assets and vulnerabilities from inequality of resources and opportunity disrupt early relational health (ERH; parent-child relationship quality, structuring the home environment, cognitive stimulation^5-6). Disruptions in ERH, in turn, can adversely affect children’s development (Figure 1).⁷

Figure 1.

Family stress and investment models of poverty and their pathways of influence on children through early relational health.

Ecological perspectives posit that human development is shaped by interactions between a child and their proximal environments.⁸ As such, promoting cognitively stimulating parenting (e.g., reading, playing with one’s child) can mitigate poverty-related disparities in children’s development, decreasing them up to 25-50%⁹ and representing a key modifiable target for early preventive parenting interventions.¹ Moreover, prior research has demonstrated that parental cognitive stimulation is uniquely related to child academic functioning, the focus of this paper,^10-12 and that academic functioning is a cumulative process, in which later achievement is related to prior skills, and higher achievement promotes subsequent academic success.^13-15 Accordingly, this study tests pathways of child development from intervention-induced changes in early parenting to both initial and later academic functioning.

Smart Beginnings Parenting Model

Smart Beginnings (SB) is an innovative, tiered birth to 3-year parenting model designed to prevent poverty-related disparities in child development and school readiness. It integrates a universal healthcare-based intervention, PlayReadVIP,¹⁶ with a targeted home-based intervention, the Family Check-Up (FCU).¹⁷

PlayReadVIP

PlayReadVIP is an evidence-based preventive intervention that seeks to reduce income-related disparities in child development and school readiness through promotion of ERH.^16,18-20 The program offers up to 14 sessions delivered by bachelor’s-level coaches at well-child visits. PlayReadVIP’s core component is a 3-minute video of the parent and child interacting using a developmentally appropriate toy or book, which is reviewed in real time to identify and reinforce responsive parenting, including ERH. Building on evidence of impact and feasibility, implementation is supported by the PlayReadVIP National Center.²¹

The Family Check-Up

The FCU is an evidence-based home-visiting model to prevent the development of early disruptive behavior.¹⁷ The FCU is a brief (3-4 sessions/year), strengths-based, individualized intervention delivered by a Masters’ level clinician focused on improving parenting and parental well-being.²² Motivational Interviewing²³ and family management are key aspects of the FCU delivery, which has three core components: 1) a Get-to-Know-You session, designed to build rapport and develop a collaborative framework for change; 2) an ecological assessment of the family’s strengths and challenges; 3) a feedback session during which the clinician and parents engage in a collaborative discussion of family strengths, challenges, and goals. Parents can choose to address goals independently or via follow-up treatment with the clinician using the evidence-based Everyday Parenting curriculum,²⁴ and/or by referral. The FCU’s evidence base has led to wide implementation in the U.S. through the federal Maternal, Infant, and Early Childhood Home Visiting program.²⁵

With these two interventions, additive and synergistic impacts are hypothesized. PlayReadVIP supports the initiation of ERH, through promotion of cognitive stimulation, with the FCU also supporting parenting in the context of barriers to ERH (e.g., parent depression, low social support). Underlying SB is the theory that tailoring of intervention to parents’ strengths and challenges will optimally support parents and promote children’s school readiness (Figure 2).⁷

Figure 2.

Tiered SB model and mechanisms of action.

Smart Beginnings Randomized Clinical Trial

A randomized clinical trial (RCT) of the SB model was conducted in New York, NY (NYC) and Pittsburgh, PA, with documented efficacy consistent with SB’s theory of change. First, we have shown direct effects of the SB model on parental cognitive stimulation beginning at child age 6 months and continuing through 24 months,²⁶ which led to subsequent indirect impacts on children’s language and literacy at age 4.²⁷ Additionally, we have shown direct effects of the SB model on harsh maternal discipline for families with pre-existing/emergent psychosocial stressors, leading to subsequent reductions in child problem behavior at age 6.²⁸ These results for the SB RCT are particularly impressive as existing large-scale preventive parenting interventions such as Early Head Start and the Nurse-Family Partnership program demonstrate limited impacts on children’s development by formal school entry.^29-30

The present study adds to the literature in substantive ways. First, it tested for indirect effects of SB through increases in parental cognitive stimulation on academic outcomes at the start of elementary school at child age 6 given that such parental cognitive stimulation is related to child academic functioning.^10-12 Second, building on findings demonstrating that academic functioning is a cumulative process in childhood,^13-15 the current study tested a specific mechanism (i.e., age 4 preacademic skills) by which supporting parents promotes children’s later academic skills. We therefore modeled serial pathways from SB to parental cognitive stimulation at age 2 to child preacademic skills at age 4 to child academic skills at age 6. Accordingly, the findings will have implications for the development of parenting interventions that seek to reduce poverty-related disparities in children’s development.^1-2

Methods

Study Design

This study was part of the single-blind, two-site RCT of SB with 403 families taking place in NYC (large urban public hospital) and in Pittsburgh (university-affiliated medical system). Informed consent was obtained from all study participants, IRB approval was obtained, and the study is registered in clinicaltrials.gov (NCT02459327).

Enrollment and Randomization

The study utilized a two-phased enrollment process with consecutive sampling (see Figure 3). Mothers and infants were enrolled in postpartum hospital units in NYC and Pittsburgh from June 2015-October 2017. In phase one, Medicaid-eligible families were offered enrollment and informed consent was obtained if they met inclusion criteria: 1) child: full-term, singleton, normal birthweight without significant medical complications, ineligible for Early Intervention at birth, plans to receive pediatric care at the institution; and 2) parent: primary caregiver/legal guardian, plans to stay in the birth city for three years, primary language English or Spanish, no known significant impairment (e.g., intellectual disability) or medical complication, no plans to stay in a shelter, baby discharged to mother, no prior participation in PlayReadVIP or FCU. In phase two, occurring through 6 weeks of age in the outpatient setting, families who continued to meet inclusion criteria were randomized by study staff to treatment or control groups within study site using study envelopes generated from a random number sequence designed to result in equal-sized groups.

Figure 3.

Participant enrollment and assessment in SB.

Note: participants who were not eligible for the study may have met more than one exclusion criteria and therefore the individual criteria numbers do not sum to the total number not eligible

All families randomly assigned to the treatment condition (n=201) were offered the PlayReadVIP component of SB, with the opportunity to additionally receive the FCU based on eligibility. The number of completed PlayReadVIP sessions ranged from 0-14 (mean 9.3), with over 50% of eligible families completing 10 or more sessions. Whereas PlayReadVIP began soon after birth, qualifying families did not begin receiving the FCU until at least infant age 6 months, when treatment families were assessed for eligibility for the FCU through annual screening as part of the study’s assessments at 6 and 18 months. Treatment families could be eligible for the FCU based on either primary or secondary criteria known to compromise parenting quality.¹⁹ Families met primary criteria if they had one of the following: clinical threshold for maternal depressive symptoms, reports of family violence, reports of child welfare agency involvement with the study target child, or very low levels of maternal literacy (i.e., below 6th grade reading level). Families met secondary criteria if they had elevated scores in screening instruments for two of the following domains: child behavior (e.g., externalizing behaviors), family well-being and support (e.g., maternal stress), caregiving (e.g., low supportive parenting), and family capital (e.g., food insecurity). Around 50% of SB treatment families qualified for the FCU, and of those, 65% completed all three FCU components. Definitive reasons for not completing all three sessions are unknown, but potentially included sufficient parenting support from PlayReadVIP at such young ages.

Families randomly assigned to the control condition (n=202) received routine pediatric primary care, including immunizations, anticipatory guidance, and developmental surveillance.

Measures

Assessments were conducted at child ages 2, 4, and 6 years by research assistants trained in child development, supervised by project investigators, and masked to randomization group. Assessments were conducted with the child’s primary caregiver, 99% of which were the mother. Descriptive statistics of study outcomes and mediators are reported in Table 1.

Table 1.

Descriptive Statistics of Primary Child Outcomes and Mediators

	SB		Control
	M	SD	M	SD
Age 6 Primary Child Outcomes
Language Skills
Receptive Vocabulary (ROWPVT)	101.02	15.33	105.97	14.48
Oral Comprehension (WJ)	83.08	20.67	83.44	18.96
Literacy Skills
Letter-Word Recognition (WJ)	86.59	20.59	89.59	20.36
Phonemic Decoding (TOWRE)	88.91	15.26	89.06	16.21
Math Skills
Applied Problems (WJ)	82.12	16.28	86.13	16.58
Age 4 Mediator - Child Preacademic Skills
Receptive Vocabulary (ROWPVT)	99.78	16.88	97.30	14.22
Letter-Word Recognition (WJ)	85.53	13.90	85.60	14.86
Age 2 Mediator - Parental Support of Cognitive Stimulation
StimQ2 Total	20.57	5.62	18.37	6.50
Observed Cognitive Development Support	3.57	1.05	3.32	0.88
Observed Parent Language Quality	3.75	1.11	3.48	0.96
Observed Parent Language Quantity	4.48	1.27	4.27	1.27

Note. ROWPVT=Receptive One-Word Picture Vocabulary Test. WJ=Woodcock-Johnson. TOWRE=Test of Word Reading Efficiency.

Age 6 Primary Outcomes – Child Academic Skills

Three academic domains were assessed at child age 6 – language, literacy, and math. Standard scores (M=100, SD=15) were used in analyses for all measures.

Language Skills

Age 6 language was assessed with the Receptive One-Word Picture Vocabulary Test, 4th Edition (ROWPVT-4), ^31-32 which measures children’s receptive vocabulary by requiring them to match a word that is spoken to them with an image of said word (α=0.97). Language was additionally assessed with the Woodcock-Johnson (WJ) IV Oral Comprehension test,³³ which measures children’s ability to understand short oral passages (α=0.83).

Literacy Skills

Age 6 literacy was assessed with the WJ Letter-Word Identification test,³⁴ which measures children’s ability to distinguish letter sounds and words (α=0.98), and with the Phonemic Decoding Efficiency subtest of the Test of Word Reading Efficiency, Second Edition (TOWRE-2)³⁵ to measure phonemic (sound) accuracy and fluency (α=0.92) by having children sound out as many nonsense words as they could in 45 seconds.

Math Skills

Age 6 math was assessed with the WJ Applied Problems test,³⁴ which measures children’s ability to analyze and solve math problems (α=0.92).

Age 2 Mediator – Parental Support of Cognitive Stimulation

The age 2 mediator was parental support of cognitive stimulation, which was assessed in two ways: survey measures and coded video-recorded observations of parent-child interactions.

Survey Measures

The StimQ₂ is a structured interview measure of caregiver cognitive stimulation developed and validated in English and Spanish for populations with low incomes.³⁶ Three core subscales of the Toddler StimQ₂ were administered: 1) Parent Verbal Responsivity (PVR), measuring verbal interactions; 2) Parental Involvement in Developmental Advance (PIDA), measuring teaching activities; and 3) Reading Activities (READ). A StimQ₂ Total score was calculated by summing the subscales (α=0.76).

Observational Measures

Parent-child dyads were observed and video-recorded participating in clean-up, free play, and teaching tasks in a laboratory setting. Videos were coded by undergraduate psychology students led by advanced graduate students in psychology using the Parent-Child Interaction Rating Scales – Infant Adaptation (PCIRS-IA)³⁷ for the domains of cognitive development, language quantity, and language quality. For each, parents were rated from very low (1), the parent did not demonstrate the behavior, to very high (7), the parent routinely demonstrated the behavior.

Latent Variable of Parental Cognitive Stimulation

A latent (not directly observed) variable of parental cognitive stimulation was generated at child age 2 from both the survey-reported StimQ₂ Total and the three observed domains of cognitive development, language quantity, and language quality. There were significant factor loadings for all indicators (p<.001) and this variable provided good model fit (χ²=.22, p=.90, CFI >.99, RMSEA <.01).

Age 4 Mediator – Latent Variable of Child Preacademic Skills

A latent variable of child preacademic skills was generated from the language (ROWPVT) and literacy (WJ Letter-Word Identification) measures administered as part of the age 4 direct child assessment.²⁷ Model fit statistics for this variable were unavailable due to the number of indicators; however, this analysis produced statistically significant factor loadings (language [β=.80, p<.001] and literacy [β=.60, p<.001]), suggesting that the latent variable of preacademic skills using these two indicators was appropriate.

Covariates

As randomization occurred within each site (NYC or Pittsburgh), a fixed effect for site was included in analyses. In addition, because data collection was on pause early in the COVID-19 pandemic (March-July 2020), also included was an indicator for child age at assessment to control for differences due to this delay (note there were no treatment-control differences in age or timing of assessments).

Analysis Plan

Study participant retention was quite high at child age 6 years (70%, n=282), and there were no differences in attrition for outcome data by intervention group (p=.31). There were also no significant differences on baseline characteristics between treatment and control groups within each site for the retained age 6 sample (Table 2). We further conducted a Little’s test and found that missing data were missing completely at random (MCAR; χ²=27.42, p=.12). Full Information Maximum Likelihood (FIML)³⁸ estimation was therefore used to account for missing data. FIML uses all available observed data to estimate model parameters, accounting for missingness patterns. As each participant in the SB RCT provided some data to the estimation, FIML amounted to a complete case analysis at age 6 (N=403).

Table 2.

Baseline Equivalence of Treatment and Control Groups by Site for the Retained Age 6 Sample

	NYC (n = 133)		Pittsburgh (n = 149)
	SB (n = 61)	Control (n = 72)	SB (n = 75)	Control (n =74)
	Mean (SD) / Percent of Sample
Child characteristics
Gender - female	44%	56%	53%	49%
Mother characteristics
Race/Ethnicity
Asian	0%	6%	0%	0%
African-American	7%	3%	80%	85%
White	2%	1%	12%	11%
Latinx	87%	89%	4%	1%
Marital status
Married	32%	27%	7%	3%
Cohabitating partner	47%	55%	35%	35%
Non-cohabitating partner	15%	8%	35%	36%
Bio. father current partner	96%	98%	93%	95%
Education
HS grad	53%	52%	85%	78%
Some college	36%	27%	43%	31%
Primiparous birth	38%	34%	33%	35%
Teenage mother	5%	3%	7%	11%
Family household characteristics
Income-to-needs ratioa	0.88 (0.68)	0.74 (0.45)	0.63 (0.54)	0.50 (0.43)
Crowding ratiob	1.41 (0.56)	1.45 (0.58)	0.83 (0.29)	0.88 (0.32)
Interview in Spanish	58%	70%	0%	0%

Note. N = 282. ^aIncome-to-needs ratio of 1.00 indicates that a family is right at the poverty threshold; 2.00 indicates that a family is 200% above that threshold.

^bThe crowding ratio indicates how many people live per room in the dwelling. A ratio above one indicates household crowding.

Joint test of all baseline sociodemographic characteristics by site: F(15, 275) = 1.01, p = .44.

To evaluate indirect effects of the SB intervention on age 6 child academic skills mediated parental cognitive stimulation, path analysis was first conducted using MPlus 8.10³⁹ to estimate single mediation models for each outcome. We estimated pathways from the SB model to age 2 parental cognitive stimulation (the α path), and then from age 2 parental cognitive stimulation to each age 6 child academic outcome (the β paths). Significant effects were established from bootstrapped 95% confidence intervals (5000 reps). As these confidence intervals are bias-corrected, they correct for skew in the population.⁴⁰

To evaluate continued indirect effects of the SB intervention mediated through initial child academic functioning, serial mediation pathways with bootstrapping from SB to parental cognitive stimulation to child pre-academic skills to age 6 child academic skills were then estimated for each outcome. We first tested paths from the SB model to parental cognitive stimulation (the α₁ path), then additional serially mediated paths from age 2 parental cognitive stimulation to age 4 pre-academic skills (the d path), and finally from age 4 preacademic skills to each age 6 academic outcome (the β paths).

In addition to indirect effects, which were hypothesized as the primary mechanism by which the SB model would have impacts on children, we also modeled pathways that did not include parental cognitive stimulation (Figures 4 and 5). These analyses estimated: 1) total effects of the SB model on each child outcome that represent the overall effect of the SB model in each case (τ pathways); and 2) direct effects that represent the residual relationships between the SB model and child outcomes after accounting for the hypothesized indirect pathways through changes in parenting (τ’ pathways).

Figure 4.

Standardized single mediation models from the SB model to age 2 parent cognitive stimulation to age 6 child academic outcomes.

(A) Single mediation model with child receptive vocabulary outcome. (B) Single mediation model with child oral comprehension outcome. (C) Single mediation model with child letter-word recognition outcome. (D) Single mediation model with child phonemic decoding outcome. (E) Single mediation model with child applied problems outcome.

Solid arrows represent statistically significant paths (p<.05), whereas dashed arrows represent nonsignificant paths.

Figure 5.

Standardized serial mediation models from the SB model to age 2 parent cognitive stimulation to age 4 child preacademic skills to age 6 child academic outcomes.

(A) Serial mediation model with child receptive vocabulary outcome. (B) Serial mediation model with child oral comprehension outcome. (C) Serial mediation model with child letter-word recognition outcome. (D) Serial mediation model with child phonemic decoding outcome. (E) Serial mediation model with child applied problems outcome.

Solid arrows represent statistically significant paths (p<.05), whereas dashed arrows represent nonsignificant paths.

Power

We used Monte Carlo simulations for all power analyses with 95% confidence intervals, 1,000 replications and 20,000 draws per replication.⁴¹ For the single mediation models, assuming small-to-medium associations between the mediator and the outcomes of interest, and a 70% retention rate, we had 82% power to detect significant indirect effects. For serial mediation models, assuming 70% retention and small-to-medium associations among mediators and outcomes, we had 81% power to detect significant indirect effects.

Results

Participants

The SB sample was comprised of mothers with low incomes at both sites. There were no significant differences on baseline characteristics between intervention and control groups within each site, the unit of randomization, in the retained age 6 sample (Table 2) or the full original accrued sample (Table 3).

Table 3.

Baseline Equivalence of Treatment and Control Groups by Site for the Original Accrued Sample

	NYC (N = 200)		Pittsburgh (N = 203)
	SB (N = 101)	Control (N = 99)	SB (N = 100)	Control (N =103)
	Mean (SD) / Percent of Sample
Child characteristics
Gender - female	45%	54%	51%	50%
Mother characteristics
Race/Ethnicity
Asian	2%	4%	0%	0%
African-American	11%	4%	80%	83%
White	2%	2%	12%	13%
Latinx	81%	88%	5%	2%
Marital status
Married	33%	31%	6%	3%
Cohabitating partner	45%	53%	38%	35%
Non-cohabitating partner	13%	8%	32%	38%
Bio. father current partner	97%	99%	92%	96%
Education
HS grad	62%	51%	86%	82%
Some college	36%	27%	43%	31%
Primiparous birth	37%	34%	30%	35%
Teenage mother	4%	4%	5%	13%
Family household characteristics
Income-to-needs ratioa	0.87 (0.67)	0.77 (0.53)	0.68 (0.60)	0.60 (0.60)
Crowding ratiob	1.38 (0.54)	1.43 (0.61)	0.85 (0.30)	0.87 (0.32)
Interview in Spanish	58%	65%	0%	0%

Note. N = 403. ^a Income-to-needs ratio of 1.00 indicates that a family is right at the poverty threshold; 2.00 indicates that a family is 200% above that threshold.

^bThe crowding ratio indicates how many people live per room in the dwelling. A ratio above one indicates household crowding.

Joint test of all baseline sociodemographic characteristics by site: F(9, 350) = 1.11, p = .35.

Indirect Pathways from SB to Age 6 Child Academic Skills

Single Mediation Models through Age 2 Parental Cognitive Stimulation

As shown in Figure 4 A-E, for each of the five academic outcomes, the α path of the model was positively statistically significant (Effect Size [ES]=.14, p=.02) in that the SB model increased parental support of cognitive stimulation. This finding is consistent with previous research that revealed significant positive impacts of SB on parental cognitive stimulation from 6–24 months.²⁶

The β path in each of the models was also positively statistically significant. SB-induced increases in parental support of cognitive stimulation significantly increased age 6 child receptive vocabulary (ES=.29, p<.001), oral comprehension (ES=.34, p<.001), letter-word recognition (ES=.27, p<.001), phonemic decoding (ES=.22, p=.01), and applied problems (ES=.30, p<.001). The magnitude of the effect sizes of the mediating pathways was in the medium range.⁴²

Lastly, the indirect pathway from the SB model to age 6 outcomes through parental cognitive stimulation was positively statistically significant for child receptive vocabulary (ES=.04, p=.05), oral comprehension (ES=.05, p=.04), letter-word recognition (ES=.04, p=.05), and applied problems (ES=.04, p=.04). The indirect pathway for phonemic decoding was positive and marginally significant (ES=.03, p=.09). These findings are consistent with indirect impacts of the SB model on children’s early language and literacy at age 4²⁷ and extending them to elementary school.

Serial Mediation Models through Age 2 Parental Cognitive Stimulation and Age 4 Child Preacademic Skills

As shown in Figure 5 A-E, for each of the five academic outcomes, the α₁ paths mirrored those in the single mediation models and remained positively statistically significant (ES=.14, p=.02), consistent with previous research with the current sample.²⁶

The d path in each of the models was also positively statistically significant at p<.001, consistent with prior SB efficacy findings.²⁷ SB-induced increases in parental cognitive stimulation significantly increased children’s preacademic skills, with slight variations across models estimating age 6 child receptive vocabulary (ES=.40), oral comprehension (ES=.41), letter-word recognition (ES=.33), phonemic decoding (ES=.38), and applied problems (ES=.39). The magnitude of the effect sizes of the mediating pathways was in the medium to large ranges.⁴²

Lastly, the indirect pathway from the SB model to age 6 outcomes through parental cognitive stimulation and child preacademic skills was positively statistically significant at p=.04 for child receptive vocabulary (ES=.04), oral comprehension (ES=.05), letter-word recognition (ES=.04), phonemic decoding (ES=.04), and applied problems (ES=.05).

Multiple Comparisons Analysis

As path analysis is a rigorous analytic method for testing relations between variables that is informed by theory, we did not adjust for multiple comparisons in our primary analyses reported above.⁴³ Nonetheless, as a robustness check, we employed the Benjamini-Hochberg false discovery procedure.⁴⁴ We found the adjusted p-values for the indirect effects were .06 or below for the single mediation models and remained at p=.04 for the serial mediation models. Thus, we are confident in the findings presented above.

Discussion

This study examined effects of the tiered, integrated SB model on children’s academic skills at age 6 as part of the SB RCT. In models testing for single mediation, we found statistically significant indirect effects of SB on children’s academic outcomes at age 6 through SB-induced increases in parental cognitive stimulation at age 2. These findings are consistent with earlier work demonstrating similar indirect effects at age 4 ²⁷ and extending them to elementary school. In models testing for serial mediation, we also found statistically significant indirect effects from the SB model to parental cognitive stimulation to children’s preacademic skills to academic outcomes at age 6. The robust and meaningful effect sizes for academic outcomes are notable and represent strong indices of observed achievement, providing explanatory and practical use in the short and long runs,⁴² particularly in comparison to other preventive parenting interventions for which effect sizes were typically much lower.^29-30

These findings provide evidence for the theory of change underlying SB, which posits that supporting parents in very early childhood can help them promote their children’s development. Importantly, the findings also extend previous work on parenting⁹ by linking increases in cognitive stimulation in infancy and toddlerhood^26-27 to improvements in an extended array of children’s academic outcomes at the start of elementary school that are foundational for later school success. The current findings also corroborate previous research demonstrating parental stimulation is related to child academic functioning,^10-12 and that later achievement is related to prior skills.^13-15

This study had many strengths including the use of a RCT design and multimethod assessment across a racially and ethnically diverse sample from two cities. There were also some limitations. First, as the SB model employed a tiered design in which all families randomized to the treatment condition received PlayReadVIP and only eligible families received the FCU, it was not possible to causally determine the added benefit of the FCU over and above PlayReadVIP alone. Second, there was some loss of participants to follow-up over the course of the study (70% retention at age 6). However, our missingness analyses demonstrated that the data were MCAR, for which FIML can be utilized as a gold standard technique for missing data.³⁸ Lastly, although the sample included families from two cities and was racially and ethnically diverse samples, we do not know how the results might generalize to other populations and locations. Future work can expand the SB model to include other contexts.

In line with ecological systems theory⁸ and family stress and investment models of poverty,^3-4 the results of this study add to the literature in three important, transformational ways. First, they indicate the enduring salience of parenting practices as proximal processes for children’s development and provide support for the SB model to impact several domains of academic functioning by increasing ERH. Second, the findings provide strong support for the utility of tiered approaches focused on positive parenting and tailored to specific family needs. In particular, we have demonstrated parallel pathways related to its two interventions, including a path primarily attributable to PlayReadVIP linking SB to academic outcomes through cognitive stimulation shown here and previously,²⁷ and a previously documented path primarily attributable to the FCU linking SB to child behavior outcomes through reduction of maternal harsh discipline.²⁸ Lastly, the magnitude of the effects was substantial and suggest that SB is a particularly effective parenting intervention with demonstrated efficacy to support parents and promote their children’s development.

Conclusion

The results of this study are especially important in light of recent recommendations by the American Academy of Pediatrics and the National Academies of Sciences, Engineering, and Medicine to layer services linked across platforms to promote school readiness by enhancing ERH,^5-6 and therefore, have implications for preventive parenting interventions to mitigate poverty-related disparities in children’s development.^1-2 In particular, our results are broadly applicable to recent population-level initiatives seeking to achieve these goals, such as The Pittsburgh Study, ⁴⁵ 3-2-1 IMPACT,⁴⁶ and Together Growing Strong.⁴⁷ The findings further shed light on the unique program features of SB and the pathways through which such models impact child development, including the importance of parenting⁹ and the stability of academic functioning across childhood.^10-12

Article Summary:

This study demonstrates sustained effects at child age 6 of the tiered birth-age 3 Smart Beginnings parenting model to address disparities early in life.

What’s Known on This Subject:

Despite longstanding efforts to design, implement, and study parenting interventions early in life to address disparities in school readiness, gaps remain related to understanding their long-term effects and their pathways of influence on child development.

What This Study Adds:

We describe sustained impacts at child age 6 of the innovative, tiered birth-3-year Smart Beginnings (SB) model. We find significant single and serially mediated indirect effects of SB on academic outcomes through parental cognitive stimulation and child preacademic skills.

Abbreviations and Acronyms:

ERH

early relational health

SB

Smart Beginnings

FCU

Family Check-Up

RCT

randomized clinical trial

NYC

New York City

ES

effect size