Reading Aloud, Play, and Social-Emotional Development

13 Pages·2017·1.19 MB·English

RAlan eL. Maenddelsoihnn, MgD, a CaArolyln oBrouckmdeyer, CaPtesl, PahD,y a Ad,r iaana nWeisdlede r , PhD, a,b Samantha Berkule Johnson, PhD,a
Anne M. Seery, PhD,a Caitlin F. Canfield, PhD,a Harris S. Huberman, MD, MPH,c Benard P. Dreyer, MDa
Social-Emotional Development
abstract
OBJECTIVES:
To determine impacts on social-emotional development at school entry of a
pediatric primary care intervention (Video Interaction Project [VIP]) promoting positive
– –
parenting through reading aloud and play, delivered in 2 phases: infant through toddler
METHODS:
(VIP birth to 3 years [VIP 0 3]) and preschool-age (VIP 3 to 5 years [VIP 3 5]).
Factorial randomized controlled trial with postpartum enrollment and random
assignment to VIP 0-3, control 0 to 3 years, and a third group without school entry
follow-up (Building Blocks) and 3-year second random assignment of VIP 0-3 and control
0 to 3 years to VIP 3-5 or control 3 to 5 years. In the VIP, a bilingual facilitator video
recorded the parent and child reading and/or playing using provided learning materials
and reviewed videos to reinforce positive interactions. Social-emotional development
at 4.5 years was assessed by parent-report Behavior Assessment System for Children,
Second Edition (Social Skills, Attention Problems, Hyperactivity, Aggression, Externalizing
RESULTS:
Problems).
VIP 0-3 and VIP 3-5 were independently associated with improved 4.5-year
’ ∼− −
Behavior Assessment System for Children, Second Edition T-scores, with effect sizes
− P “
(Cohens d) 0.25 to 0.30. Receipt of combined VIP 0-3 and VIP 3-5 was associated with
” P
d = 0.63 reduction in Hyperactivity ( = .001). VIP 0-3 resulted in reduced Clinically
P
Significant Hyperactivity (relative risk reduction for overall sample: 69.2%; = .03;
×
relative risk reduction for increased psychosocial risk: 100%; = .006). Multilevel models
CONCLUSIONS:
revealed significant VIP 0-3 linear effects and age VIP 3-5 interactions.
Phase VIP 0-3 resulted in sustained impacts on behavior problems 1.5 years
after program completion. VIP 3-5 had additional, independent impacts. With our findings,
we support the use of pediatric primary care to promote reading aloud and play from birth
NIH
to 5 years, and the potential for such programs to enhance social-emotional development.
WHAT’S KNOWN ON THIS SUBJECT: Pediatric primary care
represents a population-scalable platform for providing
aDepartment of Pediatrics, School of Medicine, New York University and Bellevue Hospital Center, New York, New effective, low-cost interventions to prevent poverty-related
York; bDepartment of Communication Sciences and Disorders, Northwestern University, Chicago, Illinois; and disparities by promoting positive parenting. A recent study
cDepartment of Pediatrics, State University of New York Downstate Medical Center, Brooklyn, New York revealed that the Video Interaction Project enhanced infant
through toddler social-emotional competencies critical for
Dr Mendelsohn conceptualized and designed the study, conducted statistical analyses, drafted the
education and health.
initial manuscript, and reviewed and revised the manuscript; Drs Cates, Weisleder, and Berkule
Johnson conceptualized and designed the study, designed the data collection instruments, WHAT THIS STUDY ADDS: In a randomized controlled trial
coordinated and supervised data collection, and critically reviewed the manuscript; Drs Seery of the Video Interaction Project, we demonstrated that
and Canfield conducted statistical analyses and critically reviewed the manuscript; Drs Dreyer promoting parent and child reading aloud and play reduced
and Huberman conceptualized and designed the study and critically reviewed the manuscript; and hyperactivity at school entry, with sustained impacts 1.5
all authors approved the final manuscript as submitted. years after completion and increased impacts with continued
intervention. With our findings, we support pediatric
This trial has been registered at www. clinicaltrials. gov (identifier NCT00212576).
preventive intervention from birth to 5 years.
DOI: https:// doi. org/ 10. 1542/ peds. 2017- 3393
Accepted for publication Feb 16, 2018 To cite: Mendelsohn AL, Cates CB, Weisleder A, et al. Reading
Aloud, Play, and Social-Emotional Development. Pediatrics.
2018;141(5):e20173393
Downloaded from www.aappublications.org/news by guest on December 31, 2018
PEDIATRICS Volume 141, number 5, May 2018:e20173393 ARTICLE
–
19 22
Poverty-related disparities in daily routines. The VIP has postpartum from November 2005
social-emotional development 2 components: (1) infant through through October 2008. Inclusion
emerge during early childhood toddler (VIP birth to 3 years [VIP criteria were as follows: planned
–
and represent a barrier to learning 0-3]) and (2) preschool-age (VIP pediatric care at BHC, full-term, no
1 3
after school entry. The pediatric 3 to 5 years [VIP 3-5]). We have significant medical complications or
primary care platform can deliver previously shown that VIP 0-3 early intervention eligibility at birth,
’
low-cost, low-intensity interventions enhanced 3-year social-emotional and mother primary caregiver was 18
– 23
for population-level prevention of development (at program s end). years or older, English- or Spanish-
4 6
disparities. Positive parenting speaking, with a working telephone
We had 3 aims in which we sought ∼
–
strategies are linked with improved or pager. Consent for the second
to determine the following: (1) if the
’
social-emotional development random assignment took place at 3
– infant-toddler component (VIP 0 3)
and are a potential focus of such years by research assistants masked
7 9 had sustained impacts on children s
interventions. social-emotional development at tRoa gnrdooump Aassssiiggnnmmeenntt.
There are several important 4.5 years, 1.5 year–s after program
research gaps. First, there has been completion; (2) if the preschool-age
a focus on families with additional component (VIP 3 5) had early, There were 2 phases of random
psychosocial risks and already independent impacts at 4.5-year assignment. Dyads were randomly
“ ” 8,1 0
emergent problems and less follow-up; and (3) if there were assigned postpartum to VIP 0-3,
study of primary prevention before additive impacts and“/or e”vidence control 0 to 3 years, or1 a9 third group
problem onset. Second, although of synergy related to receiving a (Building Blocks [BB]) by using a
developmental considerations larger intervention dose (infant random number generated by the
support intervening from infancy through toddler and preschool project director using Microsoft Excel
11
through preschool, there has together). We hypothesized that VIP (Microsoft Corp, Redmond, WA).
been limited study of additive 0-3 and VIP 3-5 would have additive, The second random assignment took
and/or differential impacts of infant independent effects on social- place by using the same methodology
through toddler and preschool emotional development at 4.5 years. at the time of the 3-year assessment.
12
interventions. Third, although METHODS All families assigned at postpartum
research has revealed sustained random assignment to VIP 0-3 or
Study Design
impacts of infant through toddler control 0 to 3 years were eligible at 3
primary prevention beyond program years for second random assignment
completion for home visiting to VIP 3-5 or control 3 to 5 years.
13 We performed a factorial, single-
(eg, Nurse Family Partnership ), Families assigned to BB received
blind RCT at an urban public
there has been more limited study of no follow-up beyond 24 months,
hospital serving low-income
sustained impacts for low-cost, low- did not participate in the 3-year
14 families (Bellevue Hospital Center
intensity primary care programs. random assignment or 3- and 4.5-
[BHC]), with enrollment and first
Fourth, although programs focused year assessments, and did not have
random assignment during the
on child behavior (eg, The Incredible data for the present analyses. Group
8 postpartum period, second random
Years ) demonstrate impacts on assignments were concealed from
assignment at age 3 years, and
social-emotional development, there staff and study participants until
5 assessments through 4.5 years.
has been limited study of social- consent for each random assignment
The trial was approved by the New
emotional impacts of primary care VwIaPs completed.
York University School of Medicine
programs promoting reading aloud
Institutional Review Board and BHC
and play; this is important given wide
– Research Review Committee and
dissemination of such programs
15 18 was registered with clinicaltrials.gov 1. VIP 0-3: As previously described,
(eg, Reach Out and Read [ROR] ).
(NCT00212576). Separate informed VIP 0-3 was provided by a
We sought to address these gaps consent was obtained for enrollment bilingual facilitator over 15
through the Bellevue Project for and first random assignment and for possible 1-on-1 30-minute
Early Language and Literacy Success, sPeacrotnicdip raanntdsom assignment. sessions from 2 weeks to 3
19 ∼
a factorial randomized controlled years. Parent and child dyads
trial (RCT) of the Video Interaction were video recorded during 5
Project (VIP), a pediatric primary Consecutive mother and infant minutes of play or shared reading
care intervention that builds on ROR dyads meeting inclusion criteria and by using a developmentally
and promotes positive parenting providing informed consent were appropriate learning material
through reading aloud, play, and enrolled and randomly assigned (toy or book). Videos were
Downloaded from www.aappublications.org/news by guest on December 31, 2018
2 MENDELSOHN et al
Sociodemographic Characteristics
immediately reviewed with the assigned postpartum to VIP 0-3 or
parent to identify and reinforce 19,2 3 control 0 to 3 years who had the
As previously described, we –
responsive interactions and 4.5-year assessment. The analytic
assessed baseline sociodemographic
promote self-reflection. Age- sample for aim 2 (early VIP 3 5
characteristics and socioeconomic
specific pamphlets were used to impacts) and aim 3 (additive and/or
status (SES; Hollingshead 4 Factor
provide suggestions for positive 27 ’ synergistic impacts) consisted of
Index ). Maternal literacy was
parenting and opportunities 252 families randomly assigned a
assessed at 6 months in the mother s
for parents to describe their second time at age 3 years who had
preferred language by using the
observations and develop goals ñ the 4.5-year assessment. Twenty-
Woodcock-Johnson III and/or the
for interacting and were taken í 28 three families who were randomly
Woodcock and Mu oz-Sandoval
home together with videos. After a assigned postpartum were not
Batera III Tests of Achievement,
3-year implementation, VIP 0-3's randomly assigned a second time
Letter-Word Identification Test; for n
cost is estimated at $175 to $200 because of either missing the 3-year
families who had not received the
per child per year, including staff, assessment ( = 22) or not consenting
literacy assessment, educational level n
equipment, supplies, rent, and 23 for second random assignment
∼ was used as a proxy. As in previous
utilities, with 1 interventionist ( = 1) but nonetheless received the
studies, mothers were categorized
providing the VIP for 400 to 4.5-year assessment because the
as having increased social risk if
500 families. original enrollment consent included
they had 1 or more of the following:
21 follow-up beyond age 3 years. These
2. VIP 3-5: VIP 3-5 took place homelessness, being a victim of
22 families were within the random
during 9 30- to 45-minute sessions violence, involvement with child
assignment protocol for aim 1 and
from 3 to 5 years with the same protection, financial difficulties, food
were included in the aim 1 analytic
structure as VIP 0-3 along with insecurity, smoking or alcohol use
sample but were outside the random
enhancements designed to promote during pregnancy, or previous mental
assignment protocol for aims 2 and 3
interactions in the context of iDllenpeesnsd, einnct lVuadriinagb ldeespression.
and therefore were not included in the
rapidly emerging developmental
aim 2 and aim 3 analytic samples.
capacities during the preschool Statistical Analysis
’
period. VIP 3-5 was developed At 3 and 4.5 years, we assessed
in consultation with preschool children s social-emotional
education experts who suggested development on the basis of parent
– Power analyses, previously described
integration of strategies from reports using 4 subscales from
24 26 for postpartum enrollment, were also
research-based curricula, the Parent Rating Scales of the
performed for the current sample
including the following: Behavior Assessment System for
29 and revealed that 275 dyads would
(1) building sessions around Children, Second Edition (BASC-2) α
provide 80% power to detect 0.35
themes (eg, grocery store, birthday (Social Skills, Attention Problems,
SD differences with 2-tailed = .05.
party); (2) video recording both Hyperactivity, and Aggression) and
Statistical analyses were performed
story book reading and play, with an Externalizing Problems composite χ t
on the basis of intent to treat for each
play planned on the basis of the (Hyperactivity and Aggression). The 2
aim. and tests and/or analysis of
story; (3) integration of writing BASC-2 has been normed in English and
variance tests were used to compare
within play (eg, shopping list, Spanish. T-scores (mean = 50; SD =
’ families contributing data at 4.5
party invitations); and 10) were calculated. Recommended
“ ” years with those who did not and
(4) focusing on characters feelings. cut points were used to dichotomize
≥ also to compare across groups for
scores as Clinically At-Risk (T-score
Control 0 to 3 years and control 3 to “ ” ≥ sociodemographics and behavior
60, 1 SD above the mean) and
scores at age 3 years after phase VIP
5 years families received standard
Clinically Significant (T-score 70,
0-3 completion and before phase VIP
pediatric care during the birth- to –
2 SD above the mean) for Attention
3-5 initiation. For aim 1 (sustained
3-year and the 3- to 5-year periods,
Problems, Hyperactivity, Aggression,
VIP 0 3 impacts), we compared
respectively, including recommended
and Externalizing; Social Skills were
anticipatory guidance and ≤ VIP 0-3 and control 0 to 3 years
categorized as Clinically At-Risk for a
children for mean behavior scores
monitoring. All groups received ROR.
Measures ATn-sacloytriec S4a0m (p1le SD below the mean).
and proportions meeting clinical
thresholds using multiple linear and
logistic regressions, respectively,
–
Bilingual research assistants masked The analytic sample for aim 1 that adjusted for VIP 3-5 random
to group assignment performed (sustained VIP 0 3 impacts) assignment; comparisons of cells
assessments. consisted of 275 families randomly with 0% present were based on
Downloaded from www.aappublications.org/news by guest on December 31, 2018
PEDIATRICS Volume 141, number 5, May 2018 3
’ RESULTS
–
Fisher s exact test. For aim 2 (early Study Sample overall sample and an 80.2% RRR for
VIP 3 5 impacts), we similarly those with increased psychosocial
compared VIP 3-5 and control 3 to 5 risk (Table 3). In addition, VIP 0-3
years children ’with adjustment for Of 675 families enrolled postpartum, was associated with a reduction in
VIP 0-3 random assignment. For aims 450 were randomly assigned to VIP meeting the BASC Hyperactivity–
1 and 2, Cohen s d (mean difference 0-3 and control 0 to 3 years, and Clinically Significant cr–iterion for
and population SD), relative risk 275 (61.1%) completed the 4.5-year the overall sample (2.8% [VIP 0 3]
reduction (RRR), absolute risk assessment and were included in vPersus 9.1% [control 0 3 years];
reduction (ARR), number needed to aim 1 analyses (Fig 1). Two hundred RRR: 69.2%; 95% CI: 7.0 to 8–9.8;
treat (NNT; calculated as 1/ARR), and ninety-six of the 450 families = .03) and for those at –increased
and 95% confidence intervals (CIs) (65.8%) were randomly assigned a psychosocial risk (0% [VIP 0 3]
were calculated. Subgroup analyses second time at 3 years, of whom 252 versus 15.6% [control 0 3 yPears];
were performed for families at families (85.1% of second random RRR: 100’%; 95% CI not defined
increased psychosocial risk, given 23 assignment) completed the 4.5-year because of 0% proportion; = .006
greater impacts on 3-year behavior. assessment and were included in by Fisher s exact test). The ARR for
For aim 3 (additive and synergistic aims 2 and 3 analyses. Clinically Significant Hyperactivity
impacts), we performed cross- was 6.3% for the overall sample
sectional analyses as follows: (1) Participants in both the aim 1 and and 15.8% for those at increased
∼
compared mean 4.5-year behavior aim 2 and aim 3 samples were psychosocial risk, respectively,
comparable across study conditions
scores by using linear regressions corresponding to NNTs of 16
–
in which the predictor variable was for all baseline characteristics (95% CI: 8 to 156) and 6 (95% CI: 3
(Table 1), although VIP 0-3 families
the number of VIP doses (0 2), and P to 26), respectively. No significant
(2) estimated effect sizes by using had somewhat lowerP literacy reductions were found for any of the
multiple linear regressions dummy ( < .10), and VIP 3-5 families had other scales for the clinical criteria.
“ ” –
lower maternal age ( < .05). The
coded separately for receipt of VIP No differences were found for Social
– “ ”
double dose (combined VIP 0 3 275 families included in 1 or more SAkimill s2. Analyses: Early Impacts of VIP
– –
and VIP 3 5) or single dose (either analytic samples were similar to 3-5
–
nonparticipants for maternal age,
VIP 0 3 or VIP 3 5) and compared
–
marital status, social risk, and
with receipt of no VIP (control 0 3
firstborn child; however, in the
years and control 3 5 years) as the P VIP 3-5 was associated with early
−
analytic samples, there were more
reference group. Finally, we analyzed P P reductions (Table 4) for Aggression
−
who had lower education ( < .001),
trajectories from 3 to 4.5 years (d = 0.22) and Externalizing
−
literacy ( < .05), and SES ( < .01)
using multilevel models (MLMs) P Problems (d = 0.26) and marginally
and were more likely Hispanic and/
with Stata/SE 14 (StataCorp, College P for Hyperactivity (d = 0.26). There
or Latino ( < .001), immigrants
Station, TX). MLM analyses allowed P was a 60.7% RRR in meeting the
( < .001), and Spanish-speaking
us to determine sustained impacts P BASC Externalizing Clinically At-Risk
–
( < .001), with somewhat more
of VIP 0-3 (linear coefficient for criterion for the overall sample and
– ×
female children ( < .10). Among
VIP 0 3) and in separate models, a 72.3% reduction (not significant)
these families, a median of 10 out of
impacts of VIP 3-5 (VIP 3 5 age for those with increased psychosocial
15 VIP 0 -3 visits were attended, and a
interaction), while accounting for risk (Table 5). No differences were
median of 4 out of 6 possible VIP 3-5
repeated measures and random found for Social Skills or Attention
slopes; additional models allowed vAinsaitlsy sbeesfo orfe A4i.m5 y1e:a Srus swtaerinee adtt ended. PAirmob 3le Amnsa.lyses: Dose and Trajectory
us to test for potential synergy (ie, Impacts of VIP 0-3 Impacts
additional increase in impact due
to inclusion of both components)
– × – – –
through inclusion of a double VIP 0-3 was associated with Each dose (ie, receipt of either VIP
× −
interaction term (VIP 0 3 VIP 3 5 sustained reductions in Attention 0 3 or VIP 3 5) was associated with
−
age). Multiple comparisons were Problems (d = 0.25), Hyperactivity reductions in Attention Problems,
− ’
addressed through the Benjamini (d = 0.31), and Externalizing (d = Hyperactivity, Aggression, and
30 − −
and Hochberg procedure by using 0.24) at age 4.5 years, 1.5 years Externalizing Problems, with d s
a false discovery rate of 10% (ie, 1 after completion of the program ranging from 0.18 to 0.32 per
–
in 10 statistically significant tests (Table 2), with 59.6% RRR in meeting dose received (Table 6). Receipt of
31 –
expected to be false positive) ; the Behavior Assessment System double dose (ie, both VIP 0 3 and
application of this approach yielded for Children (BASC) Externalizing VIP 3 5) was associated with
equivalent findings. Clinically At-Risk criterion for the reductions in each of these domains,
Downloaded from www.aappublications.org/news by guest on December 31, 2018
4 MENDELSOHN et al
FIGURE 1
Participant flowchart. C, control.
−
−
with d = 0.63 for Hyperactivity at 3 years were sustained through in which VIP 0-3 and VIP 3-5 each
×
and d = 0.54 for Externalizing 4.5 years. In contrast, trajectory potentiated impacts of the other on
× ×
Problems. MLMs of trajectories for models revealed significant age VIP Attention Problems,P as indicated
Attention Problems, Hyperactivity, 3-5 interactions for Hyperactivity, by the age VIP 0-3 VIP 3-5
Aggression, and Externalizing (Table 7, Aggression, and Externalizing but interaction (Fig 2B; = .06).
Fig 2) revealed significant linear no significant linear coefficients DISCUSSION
×
coefficients for VIP 0-3 across the age for VIP 3-5, indicating increasing
range but no significant age VIP impacts for VIP 3-5 after second
0-3 interactions (latter not shown), random assignment. Furthermore, We found that promotion of positive
indicating that differences in VIP 0-3 a nonsignificant trend was seen parenting activities such as reading
Downloaded from www.aappublications.org/news by guest on December 31, 2018
PEDIATRICS Volume 141, number 5, May 2018 5
or
ncy,
a
aP gn aloud, play, and talking resulted
e
.99.30 .45 .07 .93 .20.62.10 .29 .43 .48.61 ng pr in enhanced social-emotional
uri development for children in low-
d
5 Status (Aim 3)– VIP 03 + VIP –35 (n = 62)– 92%37% 73% 31% 87% 13%97%97% 95% 57.9 (4.2) 39%57% oking or alcohol use itsB1nhu.e5crshoo tyaamuevignaeihore fs rdat oaa midfnmtiddelpil reeAa sprct.)tr tV esorn IegoPtsrni au0o mEln-t3xe Pct d(eori roinmnnbf apalenlliemtzt iisno gn .
P 03 and VIP 3– Control 03 –Years + VIP 35 (n = 61)– 93%23% 62% 25% 84% 13%97%85% 88% 57.9 (4.1) 39%48% od insecurity, sm ViwfnoIir Pta hA d3 tds-to5ietm ni(opteniro aeenlvs, ciPidhnreodonoebcpll eepe monefrds is eoiynndn t)we irrmhegispycua hlc ttesd,
By Combined VI VIP 03 + –Control 35 –Years (n = 70) 93%29% 63% 41% 84% 4%93%94% 87% 57.4 (3.9) 46%59% ancial difficulties, fo ipE0m.af3frpe2taci)ctc witspsiea zwreteisen rcpgeo ei msnre dpbeoaonrst aoehb n ocllfoey Vm tfoIopP rmo (cnuahenpinly dtt orse .n
Control 03 –Years + Control 35 Years –(n = 59) 93%36% 59% 22% 83% 7%93%92% 85% 56.9 (3.5) 32%53% child protection, fin htwthhoeoemrsseeee e ifvmovi∼ersp indta iogcnturgs–eb pawletreeo drrgo.e3r s2afe,o 3 m 3u( sunN,p dow ttfhaooeb r0r lay.e6 ,a 3s)
with primary care based intervention
nt costing 1/5 to 1/25 of that of home
e
TABLE 1 Demographic Characteristics at Baseline (Postpartum) of Families in the Analytic Sample for Each Aim By VIP 03 Status (Aim 1)By VIP 35 Status (Aim 2)–– aaControl PControl PVIP 35 VIP 03 ––35 Years 03 Years (n = 123)(n = 143)––(n = 129)(n = 132) bFamily low SES91%93%.6593%92%.99Family increased 29%34%.4432%30%.79cpsychosocial riskMother not high 61%67%.3161%67%.36school graduateMother with low 25%36%.0733%28%.41literacy (less than ninth grade)Mother married or 83%83%.9984%85%.73partneredMother age <21 y9%8%.995%13%.048Mother Hispanic96%94%.7993%97%.26Mother born outside 88%94%.1093%91%.64United StatesMother primary 86%89%.4786%92%.16language SpanishMean child age in 57.8 (4.3)58.0 (4.8).7757.2 (3.7)57.9 (4.2).14mo (SD)Firstborn child36%41%.4740%39%.99Female child51%57%.2856%52%.61 a2 P value based on tests for comparisons of proportions and t tests and/or analysis of variance tests for comparisons of means.χb Low SES was defined as a Hollingshead 4 Factor Index of 4 or 5.c Increased psychosocial risk was defined as having 1 or more of the following: homelessness, being a victim of violence, having involvemprevious mental illness, including depression. vapdLw1HsrshhroolWpppdpadbwVlpoieaaucniunaunie5eeeorooraIyieswgsnmiPreobratmvrdtrplvlik pti hapoiylehgvegdt glshe e icud0cepctnpi rso ene netirloird cnhurl roo-maslotedclna o agttocee3irhaadg r suntyp smcr mtn sueal siryc6i';ttei aisea-tanmst vsst h ednmot hris sro fnuct srcifereearncoeeundea ioveoetu adlupeeel vlorsltr ttcsi yin w rnrsaasnc pke rfbeiaite n thf )Vp untewnyht olCiiltitspi eeic iaeanyonmeliapsg IamaismglnvrlfevitPdlc iungddru otoplledlheente eptyl pco o sgs ischrs vod0idpr ef3 inbashir wavtibgi cic,iaritsi-t roaei- tceeg ocnhitseaeae3htnttay e ttstnol ttsp uN sl sodhian.esenfgevrp,o rl, e tos as y oeadtaid Nw iaeeadrw aosrsosi rcpr dnvlriwne ino rSe kTk vwi ceotmcteo eiutddtetdniiai oioe asoshi gomiftnteaic epdae so ra tmnreotorden(neaxltd hrsharismnkp-nnrN rg innittaiy.oVae3 li vafy eiantte lcpilIN4oghi iiomevio Iet, ncncntcllnf,a ss P 3 fiaviniihTowr aata o d edntto5.lwnsi s3 r reluan a d,sa fetei=o 6ae ecmroaicigs,dthn e nnu eo W nat riyind d vddenc lo c iedh atten .l h
the program under study but also
implicit support for primary care
Downloaded from www.aappublications.org/news by guest on December 31, 2018
6 MENDELSOHN et al
TABLE 2 Aim 1 Analyses: Sustained Impacts of Completed VIP 0–3 on Mean (SD) BASC-2 T-scores at 4.5 Years (1.5 Years After VIP 0–3 Program Completion)
BASC-2 Composite-Scale Control 0–3 Years (n = 132), Mean VIP 0–3 (n = 143), Mean (SD) Effect Size (95% CI)a Pb
(SD)
Social Skillsc 51.7 (10.2) 52.3 (10.8) 0.07 (−0.19 to 0.32) .61
Attention Problems 49.2 (9.2) 46.7 (9.3) −0.25 (−0.47 to −0.03) .03
Hyperactivity 53.2 (11.4) 50.2 (10.1) −0.31 (−0.56 to −0.06) .02
Aggression 44.7 (6.9) 43.5 (7.2) −0.13 (−0.30 to 0.04) .12
Externalizing Problems 48.8 (8.9) 46.4 (8.4) −0.24 (−0.45 to −0.04) .02
a Difference between groups in SD units (Cohen’s d).
b P value was based on multiple linear regression adjusted for second random assignment group.
c Higher T-scores indicate better outcomes for Social Skills and worse outcomes for other subscales.
TABLE 3 Aim 1 Analyses: Sustained Impacts of Completed VIP 0–3 on Frequency of BASC-2 Scores in Clinically At-Risk Range at 4.5 Years (1.5 Years After
VIP 0–3 Program Completion)
BASC-2 Composite-Scale Entire Sample Increased Psychosocial Risk
Control 0–3 Years VIP 0–3 RRR % (95% CI) Pa Control 0–3 Years VIP 0–3 RRR % (95% CI) Pa
(n = 132), % (n = 143), % (n = 38), % (n = 48), %
Social Skills 13.6 11.2 N/A .53 10.5 8.3 N/A .76
Attention Problems 20.5 13.3 35.0 (−11.1 to 62.0) .11 15.8 12.5 N/A .77
Hyperactivity 27.3 18.2 33.3 (−4.1 to 57.3) .07 36.8 18.8 49.1 (−4.7 to 75.3) .06
Aggression 1.5 4.9 N/A .15 2.6 4.2 N/A .58
Externalizing Problems 12.1 4.9 59.6 (4.9 to 82.8) .03 21.2 4.2 80.2 (12.2 to 95.5) .04
N/A, not applicable.
a P value was based on multiple logistic regression adjusted for second random assignment group.
TABLE 4 Aim 2 Analyses: Early Impacts of VIP 3–5 on Mean (SD) BASC-2 T-scores at 4.5 Years
BASC-2 Composite-Scale Control 3–5 Years (n = 129), VIP 3–5 (n = 123), Mean (SD) Effect Size (95% CI)a Pb
Mean (SD)
Social Skillsc 51.6 (10.8) 52.2 (10.1) 0.06 (−0.20 to 0.32) .48
Attention Problems 48.3 (9.4) 47.4 (9.0) −0.10 (−0.32 to 0.13) .41
Hyperactivity 52.6 (10.9) 50.2 (10.2) −0.26 (−0.52 to 0.01) .05
Aggression 45.1 (8.0) 43.0 (5.9) −0.22 (−0.40 to −0.05) .01
Externalizing Problems 48.7 (9.2) 46.2 (7.7) −0.26 (−0.47 to −0.05) .01
a Difference between groups in SD units (Cohen’s d).
b P value was based on multiple linear regression adjusted for baseline random assignment group.
c Higher T-scores indicate better outcomes for Social Skills and worse outcomes for other subscales.
TABLE 5 Aim 2 Analyses: Early Impacts of VIP 3–5 on Frequency of BASC-2 Scores in Clinically At-Risk Range at 4.5 Years
BASC-2 Composite-Scale Entire Sample Increased Psychosocial Risk
Control 3–5 Years VIP 3–5 RRR %a (95% CI) Pb Control 0–3 Years VIP 0–3 RRR %a (95% CI) Pb
(n = 129), % (n = 123), % (n = 41), % (n = 37), %
Social Skills 14.0 11.4 N/A .52 9.8 8.1 N/A .99
Attention Problems 16.3 15.4 N/A .80 12.2 8.1 N/A .65
Hyperactivity 26.4 18.7 29.1 (−13.3 to 55.6) .13 26.8 27.0 N/A .78
Aggression 5.4 0.8 85.0 (−20.0 to 98.1) .08 7.3 0.0 100 (NDc) .24c
Externalizing Composite 12.4 4.9 60.7 (2.8 to 84.1) .03 19.5 5.4 72.3 (−22.2 to 93.7) .12
N/A, not applicable; ND, not defined.
a RRRs not calculated for P ≥ .15; 95% CI not defined for RRR 100%.
b P value was based on multiple logistic regression adjusted for baseline random assignment group except as indicated.
c Ninety-five percent CI for RRR and P value for logistic regression not estimated because of prediction of category with 0% present; P value was based on Fisher’s exact test.
Downloaded from www.aappublications.org/news by guest on December 31, 2018
PEDIATRICS Volume 141, number 5, May 2018 7
TABLE 6 Aim 3 Analyses: Impact of VIP Dose on Mean (SD) BASC-2 T-Scores at 4.5 Years
BASC-2 Both Control 0–3 VIP 0–3 Only VIP 3–5 Only Both VIP 0–3 Effect Size (95% CI) per Effect Size (95% CI) for Pc
Composite-Scale Years and Control (n = 70), (n = 61), and VIP 3–5 Dosea Double Doseb
3–5 Years (n = 59), Mean (SD) Mean (SD) (n = 62), Mean
Mean (SD) (SD)
Social Skillsd 51.2 (10.0) 52.0 (11.6) 51.6 (10.1) 52.7 (10.1) 0.08 (−0.11 to 0.26) 0.15 (−0.22 to 0.53) .42
Attention Problems 49.5 (8.7) 47.2 (9.8) 49.2 (9.6) 45.7 (8.1) −0.19 (−0.35 to −0.03) −0.38 (−0.71 to −0.05) .02
Hyperactivity 54.7 (11.4) 50.9 (10.2) 52.1 (11.4) 48.4 (8.6) −0.32 (−0.50 to −0.13) −0.63 (−1.0 to −0.26) .001
Aggression 46.0 (8.2) 44.4 (7.8) 43.5 (5.6) 42.4 (6.1) −0.18 (−0.31 to −0.06) −0.36 (−0.61 to −0.11) .005
Externalizing Problems 50.3 (9.6) 47.3 (8.8) 47.5 (8.4) 44.9 (6.9) −0.27 (−0.42 to −0.12) −0.54 (−0.85 to −0.24) <.001
a Effect size represents average additive value of a single dose increment (ie, receipt of either VIP 0–3 or VIP 3–5 compared with receipt of neither or receipt of both compared with receipt
of either) in SD units (Cohen’s d) on the basis of multiple linear regression, including linear term for number of possible doses (0, 1, or 2).
b Effect size for receipt of both VIP 0–3 and VIP 3–5 (double dose) compared with receipt of neither, in SD units (Cohen’s d), on the basis of multiple linear regression, with single and
double dose dummy coded.
c P values were based on multiple regression analyses and Eq for estimates of increment per dose and of estimates of double dose.
d Higher T-scores indicate better outcomes for Social Skills and worse outcomes for other subscales.
FIGURE 2
Trajectories of mean BASC-2 scores for (A) Social Skills, (B) Attention Problems, (C) Hyperactivity, (D) Aggression, and (E) Externalizing Problems from 3
to 4.5 years for children randomly assigned to each combination of VIP group and control group; the y-axis displays the predicted values based on MLMs.
C, control.
programs more broadly seeking seeking primary prevention before could have experienced differences in
to promote these activities (eg, emergence of child, parent and child, program impact. Second, assessments
ROR). In demonstrating additive and/or family challenges, in addition were conducted by parent report,
impacts of infant through toddler to secondary and tertiary prevention which can differ from report by
and preschool age intervention, our programs providing screening and other observers such as teachers
findings converge with a factorial referral for services (eg, Healthy and be susceptible to response bias;
42
RCT of the Play and Learning Steps Montefiore model, Help however, the primary focus of the VIP
10
Strategies home visiting program, Me Grow, Assuring Better Child on reading aloud and play may have
43
which found additive impacts Development ). reduced socially desirable responses.
resulting from extension of an Third, clinical-level Hyperactivity
infant through toddler parenting This study had a number of strengths, and/or Externalizing Behaviors on
intervention through the late toddler including the following: (1) a factorial the BASC do not necessarily indicate
12
to early preschool period. In our RCT, which is a strong design for a diagnosis of attention-deficit/
current findings, we support the estimating independent and additive hyperactivity disorder, and clinical
use of ongoing intervention from effects; (2) high follow-up of families assessments were not available for
birth to school entry in primary randomly assigned a second time at participating families.
care. In demonstrating impacts in 3 years (85.2%); and (3) consistency CONCLUSIONS
a low-income sample, we build on across cross-sectional and trajectory
studies of ROR and Healthy Steps in analyses. There were also important
–
which impacts in similar populations limitations. First, attrition resulted in In this study, we provide strong
14,3 7 42
were demonstrated. With analytic samples comprising greater support for the use of pediatric
our current findings, we support proportions of Hispanic and/or Latino primary care to promote positive
a role for pediatric health care for immigrants with lower SES, and it is parenting activities such as reading
location of universal programs possible that less-engaged families aloud and play and the potential for
Downloaded from www.aappublications.org/news by guest on December 31, 2018
8 MENDELSOHN et al
TABLE 7 Aim 3 Analyses: MLMs of Trajectories From 3 to 4.5 Years (n = 252)
BASC-2 Linear Effect (95% CI) for Pb Linear Effect (95% Pb Interaction Effect (95% Pb Interaction Effect (95% CI) Pb
Composite-Scale VIP 0–3a CI) for VIP 3–5a CI) Age × VIP 3–5c Age × VIP 0–3 × VIP 3–5d
Social Skillse 0.06 (−0.15 to 0.28) .58 −0.01 (−0.23 to .92 0.01 (−0.003 to 0.02) .14 0.02 (−0.01 to 0.03) .31
0.20)
Attention Problems −0.23 (−0.43 to−0.03) .03 −0.07 (−0.27 to .49 0.002 (−0.01 to 0.01) .98 −0.03 (−0.06 to 0.001) .06
0.13)
Hyperactivity −0.38 (−0.62 to −0.13) .002 −0.09 (−0.33 to .46 −0.02 (−0.03 to .02 −0.02 (−0.04 to 0.01) .23
0.15) −0.003)
Aggression −0.17 (−0.34 to −0.01) .04 −0.13 (−0.30 to .12 −0.01 (−0.02 to −0.01) .04 −0.01 (−0.03 to 0.01) .34
0.03)
Externalizing −0.31 (−51 to −0.11) .003 −0.13 (−0.33 to .20 −0.02 (−0.03 to .01 −0.02 (−0.04 to 0.01) .20
Problems 0.07) −0.004)
a MLM coefficient predicting difference between intervention and control in BASC-2 mean T-scores in SD units (Cohen’s d) across the 3 to 4.5 y period; models included age and dummy
coded VIP 0–3 and VIP 3–5.
b P values were based on MLMs.
c MLM coefficient predicting additional difference in BASC-2 T-scores for families assigned to VIP 3–5 for each mo after second random assignment (regardless of enrollment random
assignment status); models additionally included age × VIP 3–5 interaction term.
d MLM coefficient predicting additional difference in BASC-2 T-scores for families assigned to both VIP 0–3 and VIP 3–5 for each mo since second random assignment; models additionally
included age × VIP 0–3 × VIP 3–5 interaction term.
e Higher T-scores indicate better outcomes for Social Skills and worse outcomes for other subscales.
ACKNOWLEDGMENTS
ABBREVIATIONS
such programs to promote social-
emotional development as reflected
We thank many colleagues for
through reductions in disruptive ARR:  absolute risk reduction
their guidance and support,
behaviors. With the effect sizes, it BASC:  Behavior Assessment
including J. Lawrence Aber, PhD,
is suggested that such programs System for Children
Clancy Blair, PhD, David Dickinson,
can result in clinically important BASC-2: B ehavior Assessment
EdD, Arthur Fierman, MD, Virginia
differences on long-term educational Flynn, MS, Gilbert Foley, EdD, System for Children,
outcomes, given the central2 role of Perri K’lass, MD, Lesley Morrow, PhD, Second Edition
behavior for child learning. There Erin O Connor, EdD, Cybele Raver, BB:  Building Blocks
BHC: B ellevue Hospital Center
is high potential for scalability of the PhD, Catherine Tamis-Lemonda,
CI: c onfidence interval
VIP together with relatively low-cost PhD, Wendy Tineo, PhD, and Linda
MLM:  multilevel model
support integration and alignment van Schaick, MSEd. We also thank
NNT: n umber needed to treat
with existing initiatives through those who contributed to this
RCT: r andomized controlled trial
home visiting and community-based project, including Angelica Alonso,
’ ROR:  Reach Out and Read
intervention (eg, Administration for Jenny Arevalo, Nina Burtchen,
RRR:  relative risk reduction
Children and Families Maternal, Diego Catalan Molina, Aida Custode,
SES:  socioeconomic status
Infant, and Early Child Home Yuliya Gurevich, Jennifer Ledesma,
44 ’ VIP: V ideo Interaction Project
Visiting, Bridging the Word Maya Matalon, Andrea Paloian,
45 46 VIP 0-3:  Video Interaction Project
Gap, Providence Talks, Citys Caroline Raak, Kristina Vlahovicova,
47 phase birth to 3 years
First Readers, and Thirty Million and Lisa White. We also thank
48 VIP 3-5:  Video Interaction Project
Words ) and dissemination within the parents and children who
phase 3 to 5 years
well-child care. participated.
Address correspondence to Alan L. Mendelsohn, MD, Department of Pediatrics, New York University School of Medicine, 550 First Ave, OBV A519, New York, NY
10016. E-mail: [email protected]
PEDIATRICS (ISSN Numbers: Print, 0031-4005; Online, 1098-4275).
Copyright © 2018 by the American Academy of Pediatrics
FINANCIAL DISCLOSURE: The authors have indicated they have no financial relationships relevant to this article to disclose.
FUNDING: Supported by grants R01 HD047740 (Mendelsohn) and R01 HD40388 (Huberman) from the National Institutes of Health and the National Institute of
Child Health and Human Development; the Tiger Foundation; the Marks Family Foundation; Children of Bellevue, Inc; KiDS of New York University Foundation,
Inc; and Rhodebeck Charitable Trust. Drs Weisleder and Canfield were supported in part by a National Research Service award from the Health Resources and
Services Administration (T32 HD047740), with training supported in part by the New York University Clinical and Translation Science Award grant (UL1TR000038)
from the National Institutes of Health National Center for the Advancement of Translational Science. Funded by the National Institutes of Health (NIH).
POTENTIAL CONFLICT OF INTEREST: The authors have indicated they have no potential conflicts of interest to disclose.
Downloaded from www.aappublications.org/news by guest on December 31, 2018
PEDIATRICS Volume 141, number 5, May 2018 9
REFERENCES
1. M cClelland MM, Cameron CE, Connor 10. H elp Me Grow National CenterWhat 20. M endelsohn AL, Dreyer BP, Flynn V,
CM, Farris CL, Jewkes AM, Morrison FJ. is Help Me Grow. Available at: https:// et al. Use of videotaped interactions
Links between behavioral regulation helpmegrownationa l. org/ what- is- help- during pediatric well-child care
and preschoolers’ literacy, vocabulary, me- grow/ . Accessed March 27, 2018. to promote child development: a
and math skills. Dev Psychol. randomized, controlled trial. J Dev
11. Vygotsky LS. Mind in Society: The
2007;43(4):947–959 Behav Pediatr. 2005;26(1):34–41
Development of Higher Mental
2. R hoades BL, Warren HK, Domitrovich Processes. Cambridge, MA: Harvard 21. M endelsohn AL, Cates CB, Weisleder
CE, Greenberg MT. Examining the University Press; 1978 A, Berkule SB, Dreyer BP. Promotion
link between preschool social– of early school readiness using
12. L andry SH, Smith KE, Swank PR,
emotional competence and first grade pediatric primary care as an
Guttentag C. A responsive parenting
academic achievement: the role of innovative platform. Zero to Three.
intervention: the optimal timing across
attention skills. Early Child Res Q. 2013;34(1):29–40
early childhood for impacting maternal
2011;26(2):182–191 behaviors and child outcomes. Dev 22. V ideo Interaction Project. Video
3. G ershoff ET, Aber JL, Raver CC, Psychol. 2008;44(5):1335–1353 Interaction Project. Available at:
Lennon MC. Income is not enough: http:// www. videointeractionp roject.
13. O lds DL, Kitzman H, Hanks C, et al.
incorporating material hardship into org/ what- is- vip. html. Accessed
Effects of nurse home visiting on
models of income associations with September 30, 2017
maternal and child functioning:
parenting and child development.
age-9 follow-up of a randomized trial. 23. W eisleder A, Cates CB, Dreyer
Child Dev. 2007;78(1):70–95 Pediatrics. 2007;120(4). Available at: BP, et al. Promotion of positive
4. A merican Academy of Pediatrics www. pediatrics. org/ cgi/ content/ full/ parenting and prevention of
Council on Community Pediatrics. 120/ 4/ e832 socioemotional disparities. Pediatrics.
2016;137(2):e20153239
Poverty and child health in
14. M inkovitz CS, Strobino D, Mistry KB,
the United States. Pediatrics. et al. Healthy steps for young children: 24. S chickedanz JA, Dickinson DK;
2016;137(4):e20160339 sustained results at 5.5 years. Charlotte-Mecklenburg Schools.
5. C ates CB, Weisleder A, Mendelsohn AL. Pediatrics. 2007;120(3). Available at: Opening the World of Learning:
Mitigating the effects of family poverty www. pediatrics. org/ cgi/ content/ full/ A Comprehensive Early Literacy
on early child development through 120/ 3/ e658 Program. Parsippany, NJ: Pearson
Early Learning; 2005
parenting interventions in primary 15. N eedlman R, Fried LE, Morley DS,
care. Acad Pediatr. 2016;16(suppl Taylor S, Zuckerman B. Clinic-based 25. D omitrovich CE, Cortes RC, Greenberg
3):S112–S120 intervention to promote literacy. MT. Improving young children’s
A pilot study. Am J Dis Child. social and emotional competence:
6. C ouncil on Early Childhood, High
PC, Klass P. Literacy promotion: an 1991;145(8):881–884 a randomized trial of the preschool
“PATHS” curriculum. J Prim Prev.
essential component of primary 16. Z uckerman B. Promoting early literacy
2007;28(2):67–91
care pediatric practice. Pediatrics. in pediatric practice: twenty years
2014;134(2):404–409 of reach out and read. Pediatrics. 26. B redekamp S, Morrow LM, Pikulski JJ;
7. U rsache A, Blair C, Raver CC. The 2009;124(6):1660–1665 Houghton Mifflin Company. Houghton
Mifflin Pre-K : Where Bright Futures
promotion of self-regulation as a 17. K lass P, Dreyer BP, Mendelsohn AL.
Begin! Boston, MA: Houghton Mifflin;
means of enhancing school readiness Reach out and read: literacy promotion
2006
and early achievement in children in pediatric primary care. Adv Pediatr.
at risk for school failure. Child Dev 2009;56:11–27 27. Hollingshead AB. Four Factor Index
Perspect. 2012;6(2):122–128 of Social Status. New Haven, CT: Yale
18. R each Out and Read. Reach Out
University Department of Psychology;
8. P errin EC, Sheldrick RC, McMenamy & Read. Available at: http://
1975
JM, Henson BS, Carter AS. Improving reachoutandread. org/ about- us/ .
parenting skills for families of young Accessed September 30, 2017 28. W oodcock RW, Muñoz-Sandoval AF.
children in pediatric settings: a Batería Woodcock-Munoz: Pruebas de
19. M endelsohn AL, Huberman HS, Berkule
randomized clinical trial. JAMA Pediatr. Habilidad Cognitiva–Revisada. Itasca,
SB, Brockmeyer CA, Morrow LM,
2014;168(1):16–24 IL: Riverside Publishing; 1996
Dreyer BP. Primary care strategies for
29. R eynolds CR, Kamphaus RW. BASC-
9. S choler SJ, Hudnut-Beumler J, promoting parent-child interactions
2: Behavior Assessment System
Mukherjee A, Dietrich MS. A brief and school readiness in at-risk
for Children. Minneapolis, MN: NCS
intervention facilitates discussions families: the Bellevue Project for Early
Pearson, Inc; 2004
about discipline in pediatric Language, Literacy, and Education
primary care. Clin Pediatr (Phila). Success. Arch Pediatr Adolesc Med. 30. B enjamini Y, Hochberg Y. Controlling
2015;54(8):732–737 2011;165(1):33–41 the false discovery rate: a practical
Downloaded from www.aappublications.org/news by guest on December 31, 2018
10 MENDELSOHN et al