Content uploaded by Nancy Rumaldo R.
Author content
All content in this area was uploaded by Nancy Rumaldo R. on Sep 24, 2021
Content may be subject to copyright.
Success at Scale: Outcomes of Community-Based Neurodevelopment
Intervention (CASITA) for Children Ages 6–20 months With Risk of Delay in
Lima, Peru
Ann C. Miller
Harvard Medical School
Nancy Rumaldo, Guadalupe Soplapuco, and
Alicia Condeso
Socios En Salud
Betsy Kammerer
Boston Children’s Hospital
Shannon Lundy
University of California San Francisco Benioff Children’s
Hospital
Fabiola Faiffer, Andy Monta~
nez, Karen Ramos,
Naysha Rojas, Carmen Contreras, Maribel
Mu~
noz, Hilda Valdivia, and Daojing Vilca
Socios En Salud
Nandy C
ordova and Patricia Hilario
Municipalidad de Carabayllo
Martha Vibbert
SPARK Center and Boston University School of Medicine
Leonid Lecca
Socios En Salud
Sonya Shin
Harvard Medical School and Brigham and Women’sHospital
This study is a randomized controlled trial of a 12-week community-based group parenting intervention
(“CASITA”) in Lima, Peru. CASITA improved neurodevelopment in a pilot study of 60 Peruvian children and
subsequently scaled to 3,000 households throughout the district. The objective of this study was to assess inter-
vention effectiveness when implemented at scale. A total of 347 children ages 6–20 months (52.7% male, 100%
identified as “mestizo”) at risk for developmental difficulties were randomized to immediate or delayed
CASITA. At 3 months after enrollment, the immediate arm showed significantly higher overall development,
based on the Extended Ages and Stages Questionnaire and Home Observation for Measurement of the Environ-
ment scores (Cohen’sds=.36 and .31, respectively). Programs demonstrably effective at scale could help
address children’s development risks worldwide.
Global Perspectives in Early Child Development
Addressing developmental difficulties among chil-
dren at the global level remains a public health con-
cern despite the fact that a simple and inexpensive
approach—early intervention combined with nutri-
tional support—has been recognized as evidence-
based for 40 years (First & Palfrey, 1994; Grantham-
McGregor, Fernald, Kagawa, & Walker, 2014;
Grassi & La Morto-Corse, 1979; Law, Garrett, &
Nye, 2004; Nelson, Nygren, Walker, & Panoscha,
2006). Estimates from a Lancet commission suggest
Ann C. Miller and Nancy Rumaldo contributed equally.
The authors acknowledge the data collectors, the CASITA group lead-
ers, and the participants of the project and their families, without whom
this work could not have been completed.
Funding for the study was provided by a grant from Grand Chal-
lenges Canada: Saving Brains: Scaling Impact Grant #0351-03 to Socios
En Salud (Leonid Lecca, PI). The CASITA scale-up intervention itself
was also supported by the Municipality of Caraballyo, Lima, Peru with
funding as part of their participatory budget, and with the identification
of community spaces for the intervention to take place. Support for
CASITA was also received in a grant to Socios En Salud from the Korea
International Cooperation Agency (KOICA) during the first year of the
intervention in the identification and training of community agents and
the construction of spaces in health facilities and conditioning of spaces
in the community to start the intervention. A grant to Partners In
Health from the Abbvie Foundation provided support with funding to
complete the goal of enrollment and delivery of sessions of the pro-
grammatic phase of the intervention. However, the Municipality,
KOICA, and Abbvie Foundation did not provide funds for the research,
or manuscript writing, and had no involvement in these latter activities.
Correspondence concerning this article should be addressed to Sonya
Shin, Division of Global Health Equity, Brigham and Women’s Hospital,
75 Francis Street, Boston, MA 02115. Electronic mail may be sent to
sshin@bwh.harvard.edu.
©2021 The Authors. Child Development ©2021 Society for Research in
Child Development.
All rights reserved. 0009-3920/2021/xxxx-xxxx
DOI: 10.1111/cdev.13602
Child Development, xxxx 2021, Volume 00, Number 0, Pages 1–15
that nearly 280 million children in low and middle-
income countries (LMICs) are still at risk of not
achieving their full developmental potential (Lu,
Black, & Richter, 2016). The kind of stressors in
early life that can have a detrimental impact on a
child’s development, such as exposure to familial or
societal violence, toxins such as heavy metals
(arsenic, lead), illnesses such as diarrhea, HIV, hel-
minth infections, anemia, and chronic malnutrition
are often overburdened in low resource settings
(Shonkoff et al., 2012). These stressors at early ages
especially strongly affect the development of bio-
logic, behavioral, and mental health systems, which
are intertwined and feed back into each other as
they develop (McEwen, Gray, & Nasca, 2015). Thus,
the policy advice from a recently published work-
ing paper from the National Council on the Devel-
oping Child makes the point forcefully: “Strategic
investments in young children and the adults who
care for them affect long-term physical and mental
health as much as they affect early learning”
(National Scientific Council on the Developing
Child, 2020, p. 2).
According to Bronfenbrenner’s Ecological Sys-
tems theory (Bronfenbrenner, 1979), the interac-
tions between children and their environments
determine growth and development. The Socio-
Ecological Model of Development posits that child
development is influenced by embedded, dynamic,
and interconnected social spheres; from individual
factors to the primary caregiver and household
characteristics to larger fields of culture, commu-
nity, and health policy (Bronfenbrenner, 1979;
Steele, Nelson, & Cole, 2007). In this model, the
most central and influential sphere for a very
young child is that of the family, particularly the
caregivers. It’s now accepted that caregiver–child
interactions stimulate the developing brain, and
shape the brain architecture (Katz & Shatz, 1996;
Knudsen, 2004). While sustained experiences of
deprivation and lack of caregiver response can
hamper growth and development (Levin, Zeanah,
Fox, & Nelson, 2014; Nelson, Bos, Gunnar, &
Sonuga-Barke, 2011; Pollak et al., 2010), positive,
responsive, enriching interactions are demonstrated
to promote language, cognitive and social develop-
ment in very young children (Nelson, de Haan, &
Thomas, 2006).
Contingent reciprocal interaction (popularly
known as “serve and return”; National Scientific
Council on the Developing Child, 2004) is a type of
responsive and enriching interaction. In this
dynamic, the infant or child makes an effort to
interact with the caregiver, through vocalizations,
eye contact, or gestures (the “serve”) and the care-
giver notices and responds in kind (the “return”).
These interactions serve to stimulate and encourage
language and cognitive development and positive
social relationships (Committee on Integrating the
Science of Early Childhood Development, 2000;
Golinkoff, Can, Soderstrom, & Hirsh-Pasek, 2015;
Pianta, Nimetz, & Bennett, 1997). However, the cir-
cumstances surrounding a child in a family are not
always conducive to these enriching interactions.
As highlighted by Harvard’s Center on the Devel-
oping Child, for overwhelmed caregivers with mul-
tiple young children to attend to, demands of
poverty, societal disruptions, physical health con-
cerns, maternal depression, or other mental health
issues can hinder a caregiver’s ability to recognize a
child’s serve or to be able to respond with a posi-
tive return (Center on the Developing Child at Har-
vard University, 2021; National Scientific Council
on the Developing Child, 2004). Caregivers also
may not know or understand that the baby is com-
municating with them—a child’s babbling or facial
expressions may not register as something to be
responded to.
The World Health Organization’s recent adop-
tion and promotion of the Nurturing Care Frame-
work globally is a testament to the strength of the
evidence in this regard. Although many of the
same issues impact children’s development in
High-Income Countries, children in LMICs often
have multiple risk factors (Walker et al., 2011), and
both developmentally supportive programming
and research in these areas has historically been
very limited (UNESCO, 2012). Jeong et al. pub-
lished a series of meta-analyses of interventions tar-
geting responsive caregiving and early learning for
the WHO’s Recommendations on Caregiving Inter-
ventions to support child development in the first
3 years of life (Jeong, Franchette, & Yousafzai,
2018; Jeong, Pitchik, & Yousafzai, 2018). In these
analyses, programs targeting responsive care and
early learning stimulation showed significant bene-
fits to both children’s development (motor, lan-
guage, cognition, and socioemotional), and
caregiver practices and responsive interactions.
These results were particularly strong in LMICs
(Cooper et al., 2009; Jeong, Pitchik, et al., 2018;
Kalinauskiene et al., 2009), although studies per-
formed in LMICs were limited in number in the
meta-analyses.
2 Miller et al.
Evidence-Based Interventions and the Challenge of Scale
As with many diseases of poverty, having an
effective intervention is not enough. Given the vast
number of children at risk for developmental diffi-
culties globally, small, targeted interventions, no
matter how effective, will be unable to address the
global need. Effective programs need to be brought
to scale. Based on literature in scaling interventions
to address complex diseases in resource-poor set-
tings (Farmer et al., 2001; Shin et al., 2004), we
believe that effective, long-term implementation
begins with interventions that are designed for
scale—including the ability to exponentially expand
without diluting intervention fidelity, participant
retention, and ongoing program evaluation.
Some early child development programs have
been brought to scale over time. Programs that
focus on parenting, such as Jamaica’s Reach Up
Early Childhood Parenting program (Walker,
Chang, Smith, & Barker-Henningham, 2018), or the
WHO’s Care for Child Development (CCD; UNI-
CEF and World Health Organization, 2012) have
had successes at scale. The Reach Up program was
based on the Jamaica Home Visit (JHV) program,
targeting families of children with malnutrition and
stunting. The model of Reach Up has similar core
principles to that of CASITA: working through par-
ents to impact children’s development, building
caregivers’skills, enjoyment, and self-efficacy in
their interactions with children. In trials of these
interventions, marked success has been demon-
strated in improving children’s cognition and
behavior (Walker, Chang, Younger, & Grantham-
McGregor, 2010), making this part of the evidence-
base, and as such, the JHV program has been
adapted and expanded to other countries, including
Peru’s large-scale Cuna M
as program. Cuna M
as
supports day-care programs with a special focus on
nutrition. A component targeting development
through enhanced parenting knowledge and higher
quality adult–child interactions is available in
impoverished rural areas of Peru (Araujo, Dormal,
& Rubio-Codina, 2018). CCD is WHO-UNICEF’s
long-utilized package for addressing the support of
at-risk children’s development, addressing parent-
ing skills, among other things. In an appendix to a
2016 Lancet article addressing scale-up of these pro-
grams (Richter et al., 2016), J. Lucas provided an
inventory of countries implementing CCD (Lucas,
2017). CCD had been integrated into 19 countries,
in a variety of delivery settings, and translated into
17 languages. Both models are designed to be
incorporated into existing services, such as primary
health care and nutrition services.
Despite these reports, systematic evaluations of
programmatic efficacy at scale have been scarce.
Although at least four large-scale programs (Chile
Crece Contigo; Boo, Perez-Escamilla, Torres,
Segura-Perez, & Cetin, 2017), India’s Integrated
Child Development Services (Rao & Kaul, 2017),
South Africa’s Preschool Year (Richter, Samuels, &
Perez-Escamilla, 2017), and Bangladesh’s Shishu
Bikash Kendras (Child Development Centers;
Khan & Darmstadt, 2017), are included in an
appendix to the same article with reports on key
components to scale, no efficacy data are pro-
vided regarding children’s cognitive or socioemo-
tional development during the scale-up, and most
of these programs have limited or no parenting
components. To address this evidence gap, we
evaluated the efficacy of an early child develop-
ment intervention—CASITA—as it was scaled in
Lima, Peru.
From Pilot to Scale: The Evolution of a Community-
Based Early Intervention
From 2014 to 2015, a community-based early
intervention (“CASITA”) using evidence-based
strategies deployed by well-trained, well-supervised
community health agents (CHAs) in the community
was developed and pilot-tested in Peru. CASITA
was derived from an approach developed by the
SPARK Center at Boston Medical Center (SPARK
Center, 2019) using the ADAPT-ITT model (Assess-
ment, Decision-making, Adaptation, Production,
Topical experts, Integration, Training, Testing) for
adaptation (Wingood & DiClemente, 2008). The
model also drew from other evidence-based inter-
ventions, effective, and sustainable in LMICs, where
structural factors play a strong role in one’s ability
to engage in health services (Castro, 2005). Most
importantly, the pilot phase incorporated input
from local families, providers, leaders, and CHAs
to develop a model that could be readily scaled if
proven effective (O’Fallon, Tyson, & Dearry, 2000).
CASITA was designed to target the central sphere
of the child’s environment—the family—and to ulti-
mately support the healthy neurodevelopment of
children at risk for delay through an intervention
designed to increase caregiver knowledge, capacity,
and self-efficacy at providing developmentally stim-
ulating interactions with their young child. Addi-
tionally, CASITA focuses on providing caregiver
CASITA Outcomes at Scale, Lima, Peru 3
socioemotional support. In its scale-up phase,
CASITA also targets wider social spheres—other
local parents as part of the group activities, and the
macrosystem of the community as a whole, through
the implementation of community CASITA rooms
and interactions with the District and National gov-
ernments.
The pilot experience with 60 children in Lima,
Peru demonstrated a significant impact on commu-
nication, motor, and personal-social development
among children ages 6–24 months who partici-
pated in the 3-month intervention, compared with
children who did not participate (Nelson et al.,
2018).
The Present Study
Based on these promising data, with the collabo-
ration of municipal leaders and the Ministry of
Health, CASITA has been scaled to the entire dis-
trict of Carabayllo, reaching 3,000 Peruvian chil-
dren at risk for developmental difficulties in
3 years. An interprofessional team conducted a ran-
domized controlled trial on a subset of these 3,000
children to evaluate the impact of participation
compared with control children who were wait-
listed to receive the intervention after 3 months’
delay. The objective of this confirmatory study was
to measure the impact of CASITA on neurodevel-
opmental delay and developmentally supportive
home environment when delivered under program
conditions of scale. This scale-up phase for CASITA
used a social enterprise model with public–private
collaboration. The key evaluation question was
whether CASITA would retain the same program
effectiveness as that observed in the pilot study
when delivered to thousands of families. We inter-
pret this as a primarily confirmatory analysis, as
the pilot study provided data that CASITA was
efficacious in a small sample at improving both
nurturing parenting practices and child develop-
ment in comparison to standard of care. The ran-
domized controlled trial to address this question
was conducted on a subset of 347 children. Our
primary hypothesis was that participation in an
early community-based intervention (“CASITA”)
would be associated with improved development
over the standard of care for young children at risk
or with delayed child development, as measured
by a significantly higher mean overall Extended
Ages and Stages Questionnaire (EASQ) score. Our
secondary hypothesis was that participation would
be associated with improvements in developmen-
tally stimulating environment as measured by the
Home Observation for Measurement of Environ-
ment (HOME).
Method
Participants
Eligibility criteria for participation of dyads in
the randomized controlled trial were as follows: the
child’s age was between 6 and 20 months; the child
was either at risk for or had developmental delay
based on evaluation using the Escala de evaluacion
del desarollo psicomotor or “Psychomotor Develop-
ment Evaluation Scale”(EEDP; Peru’s national stan-
dard development tool—see Intervention section
below for details); the child did not have a medical
condition (e.g., cerebral palsy, tuberculosis) that
limited his or her ability to participate to the inter-
vention program; the caregiver was free of medical
conditions (e.g., psychosis, tuberculosis) that would
limit his or her ability to participate in the interven-
tion program; and finally, that the caregiver/child
dyad lived within the geographic jurisdiction of the
program (Caraballyo Region of Lima, Peru) without
plans to move for the duration of CASITA.
Through health establishment referrals from
Peru’s national child growth and development pro-
gram (“CRED”) staff and active community out-
reach, approximately 3,000 children with
developmental difficulties (specifically, delay or
risk) were identified and enrolled in the scale-up of
CASITA. Based on our pilot experience, our team
anticipated screening approximately 6,000 (73.6%)
of the 8,150 estimated children aged 6–24 months
in the region in order to enroll 3,000 eligible dyads
in the scale-up of CASITA. Of these, a small subset
was enrolled into the trial.
Of 362 children screened and invited to join, 347
(96%) were enrolled in the study (see Figure 1). 173
were enrolled into the intervention arm (immediate
CASITA) and 174 into the control arm (delayed
CASITA). Of the 173 enrolled into intervention, 144
received CASITA. All were included in the analysis.
Demographics are shown in Table 1. The median
age of the children at the time of enrollment was
12 months (interquartile range [IQR] [9–15]), and
median gestational age was 39 weeks [38–40]. All
caregivers identified as of mixed Spanish and
indigenous background and all families spoke
Spanish at home. Forty-one (12%) caregivers
reported symptoms of depression. Almost all chil-
dren (97%) were breastfed, and the median (mean)
duration of exclusive breastfeeding was 6 months
(M=4.8 months).
4 Miller et al.
Study Setting
This study took place in Carabayllo, a peri-urban
community in northern Lima, Peru. The largest and
oldest district in Lima, Carabayllo in 2017 was
home to approximately 333,000 people, of whom
16% live in poverty and approximately 17,100 are
under the age of three (Ministry of Health of Peru,
2014; Pessah, Ugarte, Suarez, & Tarazona, 2018).
Approximately 71% of the population self-identified
as “mestizo”on the 2017 census. Although most of
Carabayllo’s communities are urban, many of the
outermost communities are rural. Many families in
the community are first or second-generation immi-
grants from rural provinces of Peru (Plan de Desar-
rollo Concertado Para el Distrito de Carabayllo:
2017–2021 (“Concerted Development Plan for Cara-
bayllo District: 2017–2021”), 2016). The region is
served by 11 Ministry of Health establishments,
and the closest hospital is located in the adjacent
district, a minimum of 20 km for most community
members in Carabayllo. The study sample consisted
of children ages 6–20 months who met eligibility
criteria for CASITA and their primary caregivers,
all of whom were female. The study sample
included participants from catchment areas of nine
of the MOH health centers in Carabayllo, and con-
sisted exclusively of Spanish-speakers who self-
identified as mestizo, the majority of whom (>80%)
were not employed outside the home.
Socios En Salud (“Socios”), the nongovernmental
organization that designed and implemented
CASITA, has worked in early child health to serve
the surrounding community since its inception in
1996, with services including nutritional surveil-
lance to assess the growth and clinical status of all
children in the community. Socios’team of commu-
nity health workers typically identifies children
through community-based “health fairs”and home
visits, and assesses them for low body mass index,
as well as symptoms of diarrhea or cough. For chil-
dren who are identified to be high-risk, the team
assists the family to enroll in Social Security and
access primary care services for further evaluation
and management of common diseases such as
tuberculosis, parasitosis, and anemia.
Assessed for eligibility:
362 dyads
•Assigned to intervenon (n=173)
•Received CASITA (n=144)
•Did not receive CASTA (n=29)
ANALYSIS
Analyzed for baseline data (n=169)
•Missing baseline EASQ (n=4)
•Lost to Follow-up (n=34)
Analyzed for outcome data (n=138)
•Missing EASQ and HOME (n=31)
•Missing EASQ but not HOME (n=4)
Assigned to control (n=174)
•ANALYSIS
•Analyzed for baseline data (n=172)
•Missing EASQ (n=2)
•Lost to Follow-up (n=29)
•Analyzed for outcome data (n=145)
•Missing EASQ and HOME (n=29)
•Missing EASQ but not HOME (n=1)
•Excluded (n=15)
•Not meeng inclusion criteria (n=11)
•Declined parcipaon (n=0)
•Other reasons (n= 4)
Figure 1. Study flow chart.
CASITA Outcomes at Scale, Lima, Peru 5
Ministry of Health Establishments (health centers
or health posts) offer primary care services for free.
Peru’s national child growth and development pro-
gram (“CRED”) aims to monitor the health and
developmental status of all children up to age five.
At these visits, children are assessed for overall
health and immunizations, growth, development,
and nutritional status when they visit the clinic.
CRED evaluations are done every month for the
first year of life, and then every 3 months until
reaching 2 years of age.
Prior to 2012 and again since 2017, the Peruvian
Ministry of Health has used the EEDP (Rodriquez,
1996) to screen young children for developmental risk.
The EEDP takes approximately 20 min to administer
and assesses language, social, coordination, and gross
motor skills in children 0–24 months of age. Norms
and cutoffs are stratified into three groups: normal
(scores of 85 or higher), at risk (scores of 70–85), and
delayed (scores below 70). However, busy clinics face
significant staffing challenges to meet national norms
of systematically testing all children at CRED visits.
Therefore, typically, only children with severe symp-
toms of developmental delay are identified. These chil-
dren are seen in follow-up at their health center or
referred to a secondary or tertiary health establishment
to receive additional specialist care, such as physical
therapy, occupational therapy, or speech-language
therapy. Most families would need to travel at least an
hour each way to receive such specialty care, posing
significant barriers for frequent or ongoing therapeutic
interventions.
Procedure
Ethical Considerations
This study was reviewed and approved by the
Ethics Committee of the Universidad Peruana Caye-
tano Heredia and the Mass General Brigham (for-
merly Partners HealthCare) Institutional Review
Board.
Intervention Description
CASITA has been described elsewhere in detail
(Munoz et al., 2016; Nelson et al., 2018). Briefly,
weekly group sessions are led by a trained CHA to
provide training and support for the caretaker to
better stimulate the development of their children,
as well as social support to motivate the caregiver
in her own personal life and cultivate a sense of
self-efficacy as a parent. Each session follows a for-
mat of sequential activities:
Table 1
Baseline Demographic and Health Factors, by Study Arm, CASITA Scale Study, Carabayllo, Peru (n=347 Unless Otherwise Noted)
Variable
Control
n(%) unless noted
Intervention
n(%) unless noted pvalue
Child characteristics
Male sex 96 (55.2%) 87 (50.3%) .39
Age in months, MSD 12.1 3.6 11.6 3.4 .21
Age in months, median [IQR] 12 [9–15] 11 [9–14] .09
Birthweight in kg, MSD,n=329 3.39 0.47 3.36 0.4 .63
Missing birthweight 12 10 .67
Gestational age in weeks, median [IQR], n=350 39 [38–40] 39 [38–40] .31
Family characteristics
Region of origin .23
Costa 124 (70.9%) 109 (62.6%)
Sierra 38 (21.7%) 50 (28.7%)
Selva 13 (7.4%) 15 (8.6%)
Child’s biological father no longer with mother 27 (15.4%) 20 (11.5%) .29
Maternal depression .52
No symptoms 144 (83.2%) 151 (86.8%)
Mild symptoms 13 (7.5%) 13 (7.5%)
Moderate 7 (4%) 1 (1.1%)
Moderately severe 3 (1.7%) 3 (1.7%)
Severe 0 (0%) 0 (0%)
Missing 6 (2.4%) 5 (2.9%)
Note. IQR =interquartile range.
6 Miller et al.
1. Connecting with caregivers and children: CHAs
begin each session with an “ice-breaker”song
that all of the mothers and children have
learned at the beginning of the program, estab-
lishing a familiar routine for the group. The
CHA then starts by asking the caregivers
about their well-being and experiences since
the last session, inviting them to share any
successes and challenges in their parenting
practices and personal experiences over the
past week.
2. Child observation and knowledge-sharing about
general child development and its relation to the
child: Caregivers and CHAs observe children
together and reflect collaboratively on each
child’s behavior as it relates to general devel-
opmental expectations and to the child’s own
emerging developmental skills. CHAs engage
in formal conversation with caregivers about
general child development knowledge related
to specific developmental domains.
3. Demonstration and initiation of reciprocal attention
focusing (“Serve and Return”) and social interac-
tion activities tailored to child’s individual develop-
mental profile: Through demonstration and
gentle initiation using locally familiar objects
and toys, the CHA uses a structured curricu-
lum in flipchart format to share at least two
developmentally stimulating activities, inviting
participants to spend at least 15 min per activ-
ity. Activities are tailored to the age and devel-
opmental needs of each child and selected
with parents to maximize contingent and
reciprocal behaviors that elicit enjoyment
between parent and child and thus enhance
the parental incentive for subsequent uptake in
the home. The same curriculum and flipcharts
were provided by Socios En Salud and used in
all groups.
4. Encouragement on parenting behavior and develop-
mental interactions: CHAs offer praise and
encouragement for child interactions and share
explicit information about how parents can
improve the course of their child’s develop-
mental progress.
5. Parent socioemotional support through referral
assistance, reassurance, and validation of parent’s
concerns: At each session, CHAs invite care-
givers to share their thoughts, including things
that they learned, concerns or worries, and
questions. This shared support is designed to
help parents feel understood and valued, and
to reduce feelings of isolation that accompany
financial and psychosocial stressors.
CASITA Implementation
After obtaining administrative approvals from
the National Health Directorate (Direcci
on de Redes
Integradas de Salud [DIRIS]) and the Peruvian Min-
istry of Health (MINSA), Socios’team of
community-based nurses and associates engaged
with medical directors of the regional and micro-
regional networks to collaborate with health clinics
located in Carabayllo district.
Groups typically consisted of eight to 10 dyads
(caregiver and child). Our team intentionally
grouped families who lived in the same region,
regardless of age or degree or domain of develop-
mental delay. This allowed our team to scale
quickly, beginning new sessions as soon as enough
dyads were enrolled in a region. Children still
received age-appropriate guidance by clustering
children by age during small-group activities. Each
CHA typically led two to four different groups
with each group attending 12 sessions over
3 months. Sessions were held in health establish-
ments, community centers, CHA homes, and other
public spaces. For families who missed a session,
CHAs made home visits to understand the reason
for nonattendance and encourage the family to
rejoin the group. CHAs were not paid; however,
during CASITA, they were compensated with food
vouchers (100 Peruvian soles per month per group
led), transportation costs, and field supplies (back-
pack, project shirt, office supplies, and sunscreen).
In addition to delivering CASITA, the Socios
team worked to ensure that families were con-
nected to primary care. For this reason, all screened
children were enrolled in CRED, if they were not
already followed. The primary caregiver (usually
the mother) was also screened for depression,
domestic violence, and TB symptoms, and referred
as needed for additional care.
Study staff conducted a home visit to families
meeting enrollment criteria to explain the CASITA
program and invite them to participate. Those who
agreed to participate were introduced to their local
CHA and provided a schedule of CASITA sessions.
Ultimately, families enrolled in the intervention arm
attended a median of 10 CASITA sessions [IQR 2–
11]. One hundred and four families completed at
least eight sessions, and 29 families did not attend
any CASITA session.
Randomization
Study participants were randomized in a 1:1
ratio to receive immediate (intervention) versus
CASITA Outcomes at Scale, Lima, Peru 7
delayed (control) CASITA. The study population
included dyads who were eligible to participate in
CASITA; however, the age (6–20 months) was nar-
rowed to ensure that children assigned to the con-
trol arm would still be able to participate in
CASITA after a 3-month wait. CASITA was deliv-
ered across the Caraballyo district. After the study
enrollment initiation date, families with children
meeting criteria for inclusion into CASITA were
consecutively enrolled into the study, and
caregiver-child dyads were randomly allocated to
intervention or control. A total of nine health cen-
ters within Carabayllo were identified as study
sites. Participants at these sites were assigned to the
intervention or control groups based on computer-
generated randomization, beginning with Stata’s set
seed function to ensure replicability, conducted by
study co-authors (ACM, SSS). Informed consent
was sought from caregivers to participate in the
study; nonetheless, if any caregiver declined study
participation, the family was still allowed to partici-
pate in CASITA under program auspices. All 3,000
participants in the program scale-up were adminis-
tered EASQ and HOME. Study enrollment and ran-
domization were blinded, such that the data
collectors and interviewers were not informed of
participants’enrollment into the study or study
arm assignment.
Data Collection
Data were collected between 08/2017 and 10/
2018. A separate team of trained study staff col-
lected data through participant interviews and
observations. Participant interviews took place in
homes, or elsewhere if preferred by the participant.
All study interviews were conducted in Spanish, as
all participants identified that as their language of
choice. Participants received a food packet for inter-
view compensation. Medical records at local health
establishments were also reviewed for relevant clin-
ical information.
Extended Ages and Stages Questionnaire. We
had two main prespecified outcomes. The primary
outcome was to determine the impact of CASITA
on developmental risk among at-risk children ages
9–24 months at measurement. For this analysis, we
used the parent-reported EASQ as a measure of
developmental risk. We administered the EASQ to
all children at baseline and at 3 months. The EASQ
was adapted from the second edition of the Ages
and Stages Questionnaire, and includes sections
related to communication, gross motor skills, and
personal-social development. The instrument has
been validated in a national sample of Peruvian
children aged 0–23 months, providing age-specific
normed data for the population (Fernald, Kariger,
Hidrobob, & Gertler, 2012). The current EASQ con-
tains all items from third edition of the Ages and
Stages Questionnaire (ASQ-3) in a continuous for-
mat, allowing for comparison across age groups
without relying on Western-established cut-off
scores.
Home Observation for Measurement of Environ-
ment. The secondary outcome was to determine
the impact of CASITA on developmentally stimu-
lating environment in the homes of these same at-
risk children. For this analysis, we used the Infant-
Toddler HOME questionnaire as a measure of the
home environment. The HOME-IT includes both
observation and parent interview and evaluates
parenting and home influences on child develop-
ment. We administered three sections (Acceptance,
Involvement, and Receptiveness) of the HOME at
baseline and 3 months to evaluate the quality and
quantity of stimulation and support available to the
child in the home environment (Caldwell & Brad-
ley, 2001). The HOME has been used by groups in
Latin America, including in urban Peru (Colombo
et al., 2014; Lozoff, Park, Radan, & Wolf, 1995).
Because of the length of the questionnaires in the
setting of the scale-up, the team decided to limit
the HOME to just the three sections that seemed
most likely to be affected by CASITA, which was
not designed to affect variety or organization.
Bayley Scale of Infant Development III. A ter-
tiary, exploratory outcome was to assess the associ-
ation between CASITA and NDD risk in a subset of
children using the Bayley Scales of Infant Develop-
ment, 3rd ed. (BSID III; Bayley, 2006). The BSID III
evaluates early infant development using cognitive,
motor, and language scales and been used in
research internationally, including among Peruvian
populations (Blouin, Casapia, Joseph, & Gyorkos,
2018; Colombo et al., 2014), although the BSID III
has not been normed in Peru. In this study, BSID
III scores were used to compare within the study
population rather than to compare to an outside
standard. The BSID III requires observation and
directed participation of the child, thus its inclusion
provides a complementary assessment to the EASQ,
which is primarily parent report. Because of its
complexity and the intensive training required to
successfully administer, we limited the BSID III to a
postintervention measurement (rather than pre-
post) and narrowed the eligibility criteria for its use
to children who were between 12 and 18 months
old at the time of enrollment. Of note, 1 year into
8 Miller et al.
the study, the number of children in this age group
was lower than originally projected; thus to meet
target enrollment numbers for BSID III, we trained
the team to expand the age group to include chil-
dren nine to 18 months of age.
In addition to developmental assessments, we
collected the following baseline data: child’s age,
gender, race or ethnicity, and medical history; care-
giver’s age, gender, birthplace, language used at
home, educational, and marriage status, partner sta-
tus (living with caregiver, father of child), number
of children (number, ages, dependency), past medi-
cal history including HIV and tuberculosis history,
mental health and substance use (including during
pregnancy).
We trained study staff to administer the EEDP,
HOME, EASQ, and BSID III, implementing proto-
cols to ensure satisfactory proficiency and fidelity.
The BSID III was always administered prior to
EASQ. Given the rigor of the BSID III, we selected
two BSID assessors based on their experience in
research methods, data collection, and expertise in
working with very young children. Training began
6 months before the data collection with weekly
sessions by BSID experts (BK, SL). As part of the
training, each BSID task was assessed to determine
if it was appropriate for the population. All tasks
that needed adaptation were piloted. Prior to data
collection, the BSID experts provided a week of
intensive on-site training, including observed prac-
tice with young children. After the week, the asses-
sors had each of their trial administrations videoed
and the videos were shared with the trainers, who
discussed any incorrect procedures, all scoring, and
any questions raised by the assessors. The assessors
practiced on volunteer children until proficiency
was reached. Videos of assessments were also taken
periodically during the study to insure accurate
administration and scoring. The assessors initially
worked in pairs to administer the BSID III, and
then independently conducted the BSID III with a
trained assistant. Throughout the entire study, the
experts communicated weekly and via email to
address all testing questions that arose.
Data Analytic Plan
Primary Outcome and Hypothesis 1: Participation in
CASITA Will Be Associated With Improved
Development Over the Standard of Care
To assess the primary outcome, the impact of
CASITA on developmental risk among at-risk chil-
dren, we followed our stated analysis plan and
compared mean age-adjusted EASQ z-scores (com-
pared to the World Bank normed Peru data) after
participation (i.e. at month three) in the CASITA
intervention arm versus the control arm controlling
for baseline EASQ scores, using regression analysis.
We assessed residuals from the model to assure
assumptions of normality were met. We compared
mean baseline and follow-up EASQ scores using t-
tests or Mann–Whitney Utests as appropriate. We
also reported mean EASQ subtest scores (with stan-
dard deviations) in the communication, gross motor
and personal-social functional domains. We con-
ducted intention-to-treat analysis, including dyads
in the analysis based on assignation to CASITA.
We included maternal education (primary, sec-
ondary or postsecondary), gestational age of child
(40+weeks, 38–39 weeks, or <38 weeks) and sex of
child as covariates in the model, as these risk fac-
tors may have been associated with both enrollment
in intervention and response to CASITA a priori
sample size calculations indicated that 362 total
dyads (181 in each arm) would be needed to detect
a difference of 0.24 SDs (expected difference in this
age range from the pilot study), with an alpha level
of .05 and power of .80.
Secondary Outcome and Hypothesis: Participation in
CASITA Will Be Associated With Improvements in
Developmentally Stimulating Environment as Measured
by the IT-HOME
As a secondary outcome, we evaluated the
impact of CASITA on HOME scores. We used the
HOME Questionnaire to compare mean follow-up
scores, limited to responsivity, acceptance and
involvement subscales within an a priori defined
acceptable time frame (90 days from date of enroll-
ment). We compared mean HOME postintervention
scores in the intervention and control arms, while
controlling for baseline HOME scores using regres-
sion analysis. As a final, exploratory outcome, we
compared postintervention mean scaled BSID III
scores by study arm, including composite subscales
for cognitive, language and motor domains using
regression analysis with an alpha level of .05.
Because BSID was administered at only one time
postintervention, and because we would not neces-
sarily expect to see changes in these scores after this
short intervention, we included baseline EASQ
score as a covariate in the model as a proxy for
baseline developmental status, and the same three
covariates used in the primary outcome. We used
Stata 15 for all analyses (StataCorp., 2017; Stata Sta-
tistical Software: Release 15. College Station, TX).
CASITA Outcomes at Scale, Lima, Peru 9
Results
Table 2 presents baseline and follow-up results by
study arm in child development as measured by
EASQ, and stimulating environment as measured
by HOME. Tables 3 and 4 present regression coeffi-
cients and confidence intervals for postintervention
EASQ zscores and HOME scores adjusting for
covariates.
Primary Outcome and Hypothesis 1: Participation in
CASITA Will Be Associated With Improved
Development Over the Standard of Care
Three hundred and forty-one children had a
baseline EASQ, 172 (50%) in the control arm and
169 (50%) in the intervention arm. The mean base-
line EASQ z-score was lower (M=1.63,
SD =1.15) compared to the normed Peruvian pop-
ulation. Of the 283 who had an EASQ assessed at
CASITA completion, 145 were in the control arm
(51%) and 138 (49%) in the intervention arm. No
statistically significant differences in EASQ z-scores
existed between the intervention and control arms
at baseline (intervention M=1.67, SD =1.21 and
control M=1.60, SD =1.16, Mann–Whitney U
score z=.309, p=.76). At study completion, mean
EASQ z-scores were statistically significantly higher
in the CASITA arm than the control arm overall
and in every domain when covarying the baseline
EASQ scores and the other included covariates
(Tables 3 and 4). Mean overall z-scores at follow-up
were M=1.14, SD =1.15 in the intervention arm
and M=1.52, SD =1.03 in control arm, p=.001
(Mann–Whitney Utest). The regression coefficient
was b=.39, t(275) =3.12, p=.002, 95% CI [0.14,
0.63]. Cohen’sdeffect size for this regression (com-
parison of intervention group vs. control group’s
final EASQ scores when controlling for other fac-
tors) was .36, 95% CI [0.12, 0.59].
Of note, we compared children with and without
follow-up EASQ scores and found no significant
baseline differences with respect to study arm, gen-
der, maternal depression, or baseline EASQ; how-
ever, those who did not have follow-up EASQ
scores were significantly older than those who did,
with a baseline age of 11.6 months versus
12.6 months.
Table 2
Comparison of Mean Baseline and Postintervention EASQ z-Scores,
and HOME Scores by Study Arm (n=283)
Intervention
z-score
Control
z-score
p-valueMSDMSD
N138 145
Baseline EASQ 1.67 1.21 1.6 1.16 .75
Follow-up EASQ 1.14 1.15 1.52 1.03 .001
Baseline communication 1.81 1.06 1.66 1.08 .26
Follow-up communication 1.26 1.00 1.57 0.89 .001
Baseline motor 0.81 1.08 0.75 1.08 .83
Follow-up motor 0.28 0.95 0.46 0.90 .047
Baseline personal-social 1.54 1.35 1.59 1.30 .75
Follow-up personal-social 1.25 1.28 1.68 1.20 .002
Baseline total HOME score 15.9 2.92 16.4 2.82 .16
Follow-up total HOME score 17.07 2.94 16.3 3.14 .025
Note. EASQ =Extended Ages and Stages Questionnaire;
HOME =Home Observation for Measurement of Environment.
Table 3
Summary of Regression Analyses Testing CASITA Effects (vs. Con-
trol) on EASQ zScores (n=277)
b95% CI SE t p
Outcome
EASQ zscore
a
.42 .17, .67 .001
EASQ zscore
b
.39 .14, .63 .12 3.12 .002
EASQ communications
domain score
c
.30 .08, .52 .11 2.71 .007
EASQ gross motor domain
score
d
.22 .007, .43 .11 2.04 .04
EASQ personal social
domain
e
.41 .13, .68 .14 2.91 .004
Note. EASQ =Extended Ages and Stages Questionnaire.
a
Controlling for baseline EASQ.
b
Controlling for baseline EASQ,
maternal education level, child sex, and gestational age.
c
Control-
ling for baseline EASQ, maternal education level, child sex, and
gestational age.
d
Controlling for baseline EASQ, maternal educa-
tion level, child sex, and gestational age.
e
Controlling for baseline
EASQ, maternal education level, child sex, and gestational age.
Table 4
Multivariable Regression of Effect of CASITA on EASQ Scores Versus
Control When Controlling for Other Factors
b95% CI SE t p
EASQ score
n=277
.39 .14, .63 .12 3.12 .002
Baseline EASQ .20 .10, .31 .05 3.86 .000
Maternal education above
primary school
.13 .34, .09 .11 1.16 .25
Female child .41 .16, .65 .12 3.30 .001
Gestational age
<40 weeks
.29 .49, .09 .10 2.93 .004
Note. EASQ =Extended Ages and Stages Questionnaire.
10 Miller et al.
Secondary Outcome and Hypothesis: Participation in
CASITA Will Be Associated With Improvements in
Developmentally Stimulating Environment as Measured
by the IT-HOME
With regard to HOME scores, analysis is con-
ducted on the results of 256 children (74%) with a
baseline and follow-up HOME within the a priori
defined time limits (132 in the intervention arm and
124 in the control arm); dyads missing one of the
HOME scores were not included in the analysis.
Baseline scores did not differ between study
arms for any domains (Table 2). At follow-up, the
intervention group had significantly higher HOME
scores when covarying the baseline scores with a
regression coefficient of b=.94, t(253) =2.57,
p=.011, 95% CI [0.22, 1.66] and a Cohen’sdeffect
size of .31 95% CI [0.06, 0.55].
Exploratory Outcome: Impact of CASITA on
Developmental Status as Measured by BSID-III
119 children were administered the BSID III test
at the time of CASITA completion, 72 in the control
arm and 47 in the intervention arm (Table 5). No
differences existed in mean age or gender between
the study arms when controlling for baseline EASQ
alone or with maternal education, child sex, or ges-
tational age. No differences were seen between the
study arms in the cognitive, language, or motor
composite scores.
Discussion
In this study, which took place during program-
matic expansion of a proven community-based
intervention, CASITA participants had better
postintervention EASQ and HOME scores, indica-
tive of child developmental status and developmen-
tally supportive home environments, when
compared to controls. In an earlier pilot study,
EASQ scores improved by more than 0.75 SDs from
baseline to follow-up with a Cohen’sdeffect size of
.87 (Nelson et al., 2018). In this subsequent cohort,
EASQ scores improved by approximately 0.37 SDs
from baseline to follow-up, suggesting that CASITA
continues to be effective even when scaled to reach
a large urban and peri-urban population. Given
that other studies of psychosocial parenting inter-
ventions have had success in the controlled research
context but failed to demonstrate effectiveness
when brought to scale (Andrew et al., 2018; Rockers
et al., 2018), these study findings highlight the
potential of CASITA as a robust, scalable interven-
tion to address the massive global burden of devel-
opmental delay. Testing the efficacy of programs at
scale provides a critical appraisal of ECD interven-
tions and informs how best to implement programs
that can successfully overcome the challenge of
reaching the millions of children at risk for delay.
Certain key design and implementation features
allowed the successful scale of the CASITA model.
In his 2018 commentary on components needed to
drive ECD programs to scale in LMICs, Tomlinson
highlights the dilemma that the “soft”elements of
programs (including recruitment, training, and
supervision) that may be important drivers of pro-
grammatic success are often dropped in the transi-
tion to scale (Tomlinson, 2018). Aboud and
Yousafzai in a 2019 commentary also suggest that
when considering scale, studies on psychosocial
stimulation programs need to provide information
on how these new programs are implemented into
existing services in a compatible way (Aboud &
Yousafzai, 2019). Since its pilot inception, CASITA
was intentionally designed for scale in three key
aspects: (a) testing different strategies (e.g. individ-
ual vs. group delivery) in the pilot to explore which
would be optimal for scale. (b) Engagement of pub-
lic and private sector partners at all stages helped
align the program as it developed with leadership
priorities and leverage resources. (c) Intensive
coaching and socioemotional support to CHAs
ensured fidelity and normed the practice of
empathic interactions with caregivers, as well as
reduced the likelihood of program staff turnover.
The investment of resources in the CHAs provided
a shared sense of CASITA as a psychosocial inter-
vention that builds self-efficacy, personal growth
and social capital among community participants.
Table 5
Mean Postintervention BSID III Composite Scores by Study Arm
(n=119 Unless Otherwise Specified)
Variable
Control
n=72
Intervention
n=47
p-valueMSDMSD
Cognitive composite
standard score
96.46 8.10 98.62 8.80 .18
Language composite
standard score
(n=118)
86.84 12.40 86.00 9.81 .69
Motor composite
standard score
100.32 11.05 102.25 8.82 .31
Note. BSID III =Bayley Scales of Infant Development, 3rd ed.
CASITA Outcomes at Scale, Lima, Peru 11
A few other differences between CASITA and
other similarly evidence-based ECD programs may
have helped in efficacy at scale. CASITA was con-
ducted in a group setting rather than in individual
home visits, unlike some similarly scaled programs
(Andrew et al., 2018; Luo, Emmers, Warrinnier,
Rozelle, & Sylvia, 2019). The group setting allowed
for a caregiver support segment of the intervention,
with support and encouragement from peer care-
givers as well as facilitators. Compared to other
ECD interventions (Aboud & Yousafzai, 2015),
CASITA has a relatively short duration in its inter-
action requirements with caregivers and children—
3 versus 6–20 months. On the other hand, resources
directed toward CHA training and retention are
comparatively intensive and require a dedicated,
specialized team of nurses, field supervisors, and
peer coaches.
The scale-up phase was challenging. Recruitment
and retention of participants was an ongoing con-
cern. To ramp up enrollment during the scale-up
phase, the CASITA team used quality improvement
methods, testing different strategies to increase
recruitment and enrollment rates, and then expand-
ing those strategies found to be most effective. In
most communities, community outreach, rather
than referrals from health care providers, proved to
be more efficacious for the recruitment of partici-
pants. Not surprisingly, CASITA retention declined
compared with the 100% retention in the pilot
phase; however, 81% of enrolled participants were
retained at follow-up, reflecting intentional program
efforts to address retention in the scale-up phase.
Factors associated with participant attrition—such
as work, health issues, conflict or controlling behav-
ior by the partner, and lack of parental awareness
of the importance of child development—were dis-
cussed in weekly CASITA team meetings. The team
employed mitigation strategies, such as providing
detailed initial orientation of what to expect during
the program; allowing parents to assign an alter-
nate caregiver if they could not attend; and con-
ducting home visits and offering make-up sessions
to families who missed a session.
Our exploratory use of BSID III allowed our
team to gain familiarity and experience using this
tool, which has been used with differing levels of
success in studies in many countries outside of the
US. BSID III scores were not different between the
study arms at the end of the study, but this was
not unexpected, as the sample size was small and
sensitivity of BSID III to short-term change is low.
An advantage is that the BSID III does not rely on
parent interview, however, reliance on the direct
participation of the child in the assessment process
(rather than the parent interview format) is depen-
dent on the child’s level of participation at that
time. Given that the results of the BSID III did not
show differences, it is possible that intervention
group parents exaggerated improvement on EASQ
because of a social-acceptance bias, or that mothers
in the intervention group merely paid more atten-
tion or had increased awareness of their children’s
developmental behaviors as a consequence of the
intervention, thus reporting higher EASQ scores.
However, improvements were also noted on
HOME, which uses observation by an examiner
who was not part of the CASITA intervention, so
the changes were not solely on parent-reported
items. Under-enrollment of BSID-eligible children
also resulted in underpowered analysis: per a priori
calculations, assessing 64 children using the BSID
III per study arm would have allowed detection of
a 0.5 SD difference between arms; however, the
intervention arm had BSID in only 47 children.
BSID III was not a primary endpoint for this study,
but rather a pilot test for the CASITA team. It
proved feasible for our field staff to conduct this
during home visits; future studies could incorporate
both baseline and follow-up BSID.
Our study has several limitations. Missing data
were a factor in both EASQ and HOME endpoints;
however, sensitivity analyses demonstrated that
baseline characteristics among participants who did
not have follow-up EASQ or HOME measures were
not significantly different from those in the analysis,
with the exception of child age for EASQ nonpartic-
ipants. Despite the similarities between the retained
and missing dyads, we cannot rule out the possibil-
ity of residual confounding influencing these
results. Given that this intervention took place in
one district and the context of a district-wide scale-
up, it is possible that contamination between study
arms occurred between neighbors assigned to dif-
ferent arms. However, the likely consequence of
this would be to reduce the observed effect of the
intervention. We also acknowledge that study find-
ings may not be generalizable to other settings;
however, it is notable that CASITA is also being
implemented and evaluated in several other coun-
tries (Mexico, Lesotho). Program assessments from
these sites will likely shed light on whether
CASITA is applicable and effective in other settings.
Longer follow-up could determine if the benefitof
CASITA is sustained over time.
Despite these limitations, our study has several
unique strengths. We believe the design of the
intervention—including the investment to build and
12 Miller et al.
support a strong CHA workforce and the use of
group sessions—made the program more amenable
to scale. Together with the pilot data, this study
provides compelling data to suggest that CASITA
can be successfully scaled to improve developmen-
tal outcomes in thousands of children in Lima,
Peru. This study contributes to the global literature
on the effectiveness of parenting stimulation pro-
grams in the context of scale. Programs that are
demonstrably effective in multiple domains at scale
are urgently needed to address the development
risks faced by millions of children worldwide.
In conclusion, the CASITA nutrition and parent-
ing support intervention, when implemented at
scale, was successful at improving children’s devel-
opment and creating developmentally supportive
home environments, as measured by the EASQ and
HOME. Programs such as CASITA that are demon-
strably effective in multiple domains at scale are
badly needed to address the development risks
faced by millions of children worldwide.
References
Aboud, F. E., & Yousafzai, A. K. (2015). Global health
and development in early childhood. Annual Review of
Psychology,66, 433–457. https://doi.org/10.1146/
annurev-psych-010814-015128
Aboud, F. E., & Yousafzai, A. K. (2019). Scaling up child
psychosocial stimulation programmes for young chil-
dren. The Lancet Global Health,7, e294–e295. https://
doi.org/10.1016/S2214-109X(19)30018-X
Andrew, A., Attanasio, O., Fitzsimons, E., Grantham-
McGregor, S., Meghir, C., & Rubio-Codina, M. (2018).
Impacts 2 years after a scalable early childhood devel-
opment intervention to increase psychosocial stimula-
tion in the home: A follow-up of a cluster randomised
controlled trial in Colombia. PLoS Medicine,15,
e1002556. https://doi.org/10.1371/journal.pmed.
1002556
Araujo, M. C., Dormal, M., & Rubio-Codina, M. (2018).
Quality of parenting programs and child development out-
comes: The case of Peru’s Cuna Mas. Retrieved from
https://publications.iadb.org/publications/english/doc
ument/Quality-of-Parenting-Programs-and-Child-Deve
lopment-Outcomes-The-Case-of-Peru-Cuna-Mas.pdf
Bayley, N. (2006). Bayley Scales of Infant and Toddler Devel-
opment (3rd ed.). San Antonio, TX: Harcourt Assess-
ment.
Blouin, B., Casapia, M., Joseph, L., & Gyorkos, T. W.
(2018). A longitudinal cohort study of soil-transmitted
helminth infections during the second year of life and
associations with reduced long-term cognitive and ver-
bal abilities. PLoS Neglected Tropical Diseases,12,
e0006688. https://doi.org/10.1371/journal.pntd.0006688
Boo, F. L., Perez-Escamilla, R., Torres, A., Segura-Perez,
S., & Cetin, Z. (2017). Web appendix 4: Case studies of
four scaled-up early child development programmes in
middle-income countries: Chile Crece Contigo.
Appendix to Richter et al. The Lancet,389, 103–118.
Bronfenbrenner, U. (1979). The ecology of human develop-
ment: Experiments by nature and design. Cambridge, MA:
Harvard University Press.
Caldwell, B., & Bradley, R. (Eds.). (2001). Home Inventory
Administration Manual (3rd ed.). Little Rock, AR:
University of Arkansas at Little Rock.
Castro, A. (2005). Adherence to antiretroviral therapy:
Merging the clinical and social course of AIDS. PLoS
Medicine,2, e338. https://doi.org/10.1371/journal.
pmed.0020338
Center on the Developing Child at Harvard University.
(2021). Serve and return. Retrieved from https://deve
lopingchild.harvard.edu/science/key-concepts/serve-
and-return/
Colombo, J., Zavaleta, N., Kannass, K. N., Lazarte, F.,
Albornoz, C., Kapa, L. L., & Caulfield, L. E. (2014).
Zinc supplementation sustained normative neurodevel-
opment in a randomized, controlled trial of Peruvian
infants aged 6–18 months. Journal of Nutrition,144,
1298–1305. https://doi.org/10.3945/jn.113.189365
Committee on Integrating the Science of Early Childhood
Development. (2000). From neurons to neighborhoods: The
science of early childhood development (J. P. Shonkoff, &
D. Phillips, Eds., 1st ed.). Washington, DC: National
Academy Press.
Cooper, P. J., Tomlinson, M., Swartz, L., Landman, M.,
Molteno, C., Stein, A., . . . Murray, L. (2009). Improving
quality of mother–infant relationship and infant attach-
ment in socioeconomically deprived community in
South Africa: Randomised controlled trial. British Medi-
cal Journal,338, b974. https://doi.org/10.1136/bmj.b
974
Farmer, P., L
eandre, F., Mukherjee, J. S., Claude, M. S.,
Nevil, P., Smith-Fawzi, M. C., . . . Kim, J. Y. (2001).
Community-based approaches to HIV treatment in
resource-poor settings. The Lancet,358, 404–409.
https://doi.org/10.1016/S0140-6736(01)05550-7
Fernald, L., Kariger, P., Hidrobob, M., & Gertler, P.
(2012). Socioeconomic gradients in child development
in very young children: Evidence from India, Indone-
sia, Peru, and Senegal. Proceedings of the National Acad-
emy of Sciences of the United States of America,109,
17273–17280. https://doi.org/10.1073/pnas.1121241109
First, L. R., & Palfrey, J. S. (1994). The infant or young
child with developmental delay. New England Journal of
Medicine,330, 478–483. https://doi.org/10.1056/
NEJM199402173300708
Golinkoff, R. M., Can, D. D., Soderstrom, M., & Hirsh-
Pasek, K. (2015). (Baby)talk to me: The social context of
infant-directed speech and its effects on early language
acquisition. Current Directions in Psychological Science,
24, 339–344. https://doi.org/10.1177/096372141559
5345
CASITA Outcomes at Scale, Lima, Peru 13
Grantham-McGregor, S. M., Fernald, L. C., Kagawa, R.
M., & Walker, S. (2014). Effects of integrated child
development and nutrition interventions on child
development and nutritional status. Annals of the New
York Academy of Sciences,1308,11–32. https://doi.org/
10.1111/nyas.12284
Grassi, J. R., & La Morto-Corse, A. (1979). Identification
and remediation of basic cognitive deficits in disadvan-
taged children. Journal of Learning Disabilities,12, 483–
487. https://doi.org/10.1177/002221947901200709
Jeong, J., Franchette, E., & Yousafzai, A. (2018). World
Health Organization recommendations on caregiving inter-
ventions to support early child development in the first three
years of life: Report on the systematic review of evidence.
Retrieved from https://www.who.int/maternal_chil
d_adolescent/guidelines/SR_Caregiving_interventions_
ECD_Jeong_Final_Nov2018.pdf?ua=1
Jeong, J., Pitchik, H. O., & Yousafzai, A. K. (2018). Stimu-
lation interventions and parenting in low- and middle-
income countries: A meta-analysis. Pediatrics,141.
https://doi.org/10.1542/peds.2017-3510
Kalinauskiene, L., Cekuoliene, D., Van Ijzendoorn, M. H.,
Bakermans-Kranenburg, M. J., Juffer, F., & Kusakovs-
kaja, I. (2009). Supporting insensitive mothers: The Vil-
nius randomized control trail of video-feedback
intervantion to promote maternal sensitivity and infant
attachment security. Child: Care, Health and Development,
35, 613–623.
Katz, L. C., & Shatz, C. J. (1996). Synaptic activity and
the construction of cortical circuits. Science,274, 1133–
1138. https://doi.org/10.1126/science.274.5290.1133
Khan, N. Z., & Darmstadt, G. L. (2017). Web Appendix 4:
Case studies of four scaled-up early child development
programmes in middle-income countries: Shishu Bikash
Kendras (Child Development Centres) in Bangladesh.
The Lancet,389, 103–118.
Knudsen, E. I. (2004). Sensitive periods in the develop-
ment of the brain and behavior. Journal of Cognitive
Neuroscience,16, 1412–1425. https://doi.org/10.1162/
0898929042304796
Law, J., Garrett, Z., & Nye, C. (2004). The efficacy of
treatment for children with developmental speech and
language delay/disorder: A meta-analysis. Journal of
Speech, Language, and Hearing Research,47, 924–943.
https://doi.org/10.1044/1092-4388(2004/069)
Levin, A. R., Zeanah Jr., C. H., Fox, N. A., & Nelson, C.
A. (2014). Motor outcomes in children exposed to early
psychosocial deprivation. Journal of Pediatrics,164, 123–
129.e1. https://doi.org/10.1016/j.jpeds.2013.09.026
Lozoff, B., Park, A. M., Radan, A. E., & Wolf, A. W.
(1995). Using the HOME inventory with infants in
Costa Rica. International Journal of Behavioral Develop-
ment,18, 277–295. https://doi.org/10.1177/
016502549501800206
Lu, C., Black, M. M., & Richter, L. M. (2016). Risk of poor
development in young children in low-income and
middle-income countries: An estimation and analysis at
the global, regional, and country level. The Lancet Global
Health,4, e916–e922. https://doi.org/10.1016/S2214-
109X(16)30266-2
Lucas, J. (2017). Web appendix 2: Implementing care for
child development: An Inventory of a play and com-
munication intervention to support child development
and responsive caregiving. In Investing in the founda-
tion of sustainable development: Pathways to scale up
for early childhood development. The Lancet,389, 103–
118.
Luo, R., Emmers, D., Warrinnier, N., Rozelle, S., & Sylvia,
S. (2019). Using community health workers to deliver a
scalable integrated parenting program in rural China:
A cluster-randomized controlled trial. Social Science and
Medicine,239, 112545. https://doi.org/10.1016/j.socsc
imed.2019.112545
McEwen, B. S., Gray, J. D., & Nasca, C. (2015). 60 years
of neuroendocrinology: Redefining neuroendocrinology:
Stress, sex and cognitive and emotional regulation.
Journal of Endocrinology,226, T67–T83. https://doi.org/
10.1530/JOE-15-0121
Ministry of Health of Peru. (2014). Analisis de Situacion de
Salud del Distrito de Carabayllo: Proyeccion de la Poblacion
de Carabayllo (“Health Situation Analysis of Carabayllo
District: Population Projection for Carabayllo”). Lima,
Peru: Peruvian Ministry of Health.
Munoz, M., Nelson, A. K., Johnson, M., Godoy, N., Ser-
rano, E., Chagua, E., . . . Shin, S. (2016). Community-
based needs assessment of neurodevelopment, care-
giver, and home environment factors in young children
affected by HIV in Lima, Peru. Journal of the Interna-
tional Association of Providers of AIDS Care,16, 161–167.
National Scientific Council on the Developing Child.
(2004). Young children develop in an environment of rela-
tionships: Working paper no. 1. Cambridge, MA: Center
on the Developing Child at Harvard University.
National Scientific Council on the Developing Child.
(2020). Connecting the brain the rest of the body: Early
childhood development and lifelong health are deeply inter-
twined: Working paper number 15. Cambridge, MA: Har-
vard University.
Nelson, A. K., Miller, A. C., Munoz, M., Rumaldo, N.,
Kammerer, B., Vibbert, M., . . . Shin, S. S. (2018).
CASITA: A controlled pilot study of community-based
family coaching to stimulate early child development in
Lima, Peru. BMJ Paediatrics Open,2, e000268. https://
doi.org/10.1136/bmjpo-2018-000268
Nelson, C. A., Bos, K., Gunnar, M. R., & Sonuga-Barke, E.
J. (2011). The neurobiological toll of early human depri-
vation. Serial No 5. Monographs of the Society for
Research in Child Development,76, 127–146. https://doi.
org/10.1111/j.1540-5834.2011.00630.x
Nelson, C., de Haan, M., & Thomas, K. M. (2006). Neural
bases of cognitive development. In W. Damon, R. Ler-
ner, D. Kuhn, & R. Seigler (Eds.), Handbook of child psy-
chology (Vol. 2, pp. 15–26). Hoboken, NJ: Wiley.
Nelson, H. D., Nygren, P., Walker, M., & Panoscha, R.
(2006). Screening for speech and language delay in pre-
school children: Systematic evidence review for the US
14 Miller et al.
Preventive Services Task Force. Pediatrics,117, e298–
e319. https://doi.org/10.1542/peds.2005-1467
O’Fallon, L. R., Tyson, F. L., & Dearry, A. (2000). Success-
ful models of community-based participatory research.
Retrieved from http://www.niehs.nih.gov/translat/cb
r-final.pdf
Pessah, S. E., Ugarte, C. M. T., Suarez, L. A. N., & Tara-
zona, A. M. (2018). Analisis de Situation de Salud 2018:
Direccion de Redes Integradas de Salud Lima Norte
(“Health Situation Analysis: Directive of Integrated Health
Networks Northern Lima”). Lima, Peru: Ministry of
Health. Retrieved from http://www.dirislimanorte.gob.
pe/wp-content/uploads/2019/01/asis-2018.pdf
Pianta, R. C., Nimetz, S. L., & Bennett, E. (1997). Mother–
child relationships, teacher–child relationships, and
school outcomes in preschool and kindergarten. Early
Childhood Research Quarterly,12, 263–280. https://doi.
org/10.1016/S0885-2006(97)90003-X
Plan de Desarrollo Concertado Para el Distrito de Cara-
bayllo: 2017–2021 (“Concerted Development Plan for
Carabayllo District: 2017–2021”). (2016). Municipalidad
Distrital de Carabayllo (“District Municipality of Cara-
bayllo”). Lima, Peru: Carabayllo Municipality. Retrieved
from https://www.scribd.com/document/383808327/
Plan-de-Desarrollo-Carabayllo-2017-2021
Pollak,S.D.,Nelson,C.A.,Schlaak,M.F.,Roeber,B.J.,
Wewerka, S. S., Wiik, K. L., . . . Gunnar, M. R. (2010).
Neurodevelopmental effects of early deprivation in
postinstitutionalized children. Child Development,81,224–
236. https://doi.org/10.1111/j.1467-8624.2009.01391.x
Rao, N., & Kaul, V. (2017). Web appendix 4: Case studies
of four scaled-up early child development programmes
in middle-income countries: Integrated Child Develop-
ment Services; appendix to Richter et al. The Lancet,
389, 103–118.
Richter, L. M., Daelmans, B., Lombardi, J., Heymann, J.,
Boo, F. L., Behrman, J. R., . . . Darmstadt, G. L. (2016).
Working Group and the Lancet Early Childhood Devel-
opment Series Steering Committee. Investing in the foun-
dation of sustainable development: Pathways to scale up
for early childhood development. The Lancet,389,103–
118. https://doi.org/10.1016/S0140-6736(16)31698-1
Richter, L., Samuels, M., & Perez-Escamilla, R. (2017). Web
appendix 4: Case studies of four scaled-up early child
development programmes in middle-income countries:
South Africa’s preschool year (grade R or reception year).
Appendix to Richter et al. The Lancet,389, 103–118.
Rockers, P. C., Zanolini, A., Banda, B., Chipili, M. M.,
Hughes, R. C., Hamer, D. H., & Fink, G. (2018). Two-
year impact of community-based health screening and
parenting groups on child development in Zambia:
Follow-up to a cluster-randomized controlled trial.
PLoS Medicine,15, e1002555. https://doi.org/10.1371/
journal.pmed.1002555
Rodriquez, S. (Ed.). (1996). Escala de evaluacion del desarollo
psicomotor: 0 a 24 meses (“Psychomotor Development Eval-
uation Scale: 0 to 24 months”)(12th ed.). Santiago, Chile:
Galdoc.
Shin, S., Furin, J., Bayona, J., Mate, K., Kim, J. Y., &
Farmer, P. (2004). Community-based treatment of
multidrug-resistant tuberculosis in Lima, Peru: 7 years
of experience. Social Science and Medicine,59, 1529–1539.
https://doi.org/10.1016/j.socscimed.2004.01.027
Shonkoff, J. P., Garner, A. S., Siegel, B. S., Dobbins, M. I.,
Earls, M. F., Garner, A. S., . . . Wood, D. L. (2012). The
lifelong effects of early childhood adversity and toxic
stress. Pediatrics,129, e232–e246. https://doi.org/10.
1542/peds.2011-2663
SPARK Center (Supporting Parents and Resilient Kids).
(2019). Retrieved from https://www.bmc.org/progra
ms/spark-center
Steele, R. G., Nelson, T. D., & Cole, B. P. (2007). Psy-
chosocial functioning of children with AIDS and HIV
infection: Review of the literature from a socioecologi-
cal framework. Journal of Developmental and Behavioral
Pediatrics,28,58–69. https://doi.org/10.1097/DBP.0b
013e31803084c6
Tomlinson, M. (2018). From surviving to thriving: What
evidence is needed to move early child-development
interventions to scale? PLoS Medicine,15, e1002557.
https://doi.org/10.1371/journal.pmed.1002557
UNESCO. (2012). Expanding equitable early childhood care
and education is an urgent need: Policy paper 03. Paris,
France: Author.
UNICEF and World Health Organization. (2012). Care for
child development package. Geneva, Switzerland: World
Health Organization.
Walker, S. P., Chang, S. M., Smith, J. A., & Barker-
Henningham, H. (2018). The reach up early childhood
parenting programme: Origins, content and implemen-
tation. Zero to Three,38,37–43.
Walker, S. P., Chang, S. M., Younger, N., & Grantham-
McGregor, S. M. (2010). The effect of psychosocial stim-
ulation on cognition and behaviour at 6 years in a
cohort of term, low-birthweight Jamaican children.
Developmental Medicine and Child Neurology,52, e148–
e154. https://doi.org/10.1111/j.1469-8749.2010.03637.x
Walker, S. P., Wachs, T. D., Grantham-McGregor, S.,
Black, M. M., Nelson, C. A., Huffman, S. L., . . . Rich-
ter, L. (2011). Inequality in early childhood: Risk and
protective factors for early child development. The Lan-
cet,378, 1325–1338. https://doi.org/10.1016/S0140-
6736(11)60555-2
Wingood, G., & DiClemente, R. (2008). The ADAPT-ITT
model: A novel method of adapting evidence-based
HIV interventions. Journal of Acquired Immune Deficiency
Syndromes,47(Suppl. 1), S40–S46. https://doi.org/10.
1097/QAI.0b013e3181605df1
CASITA Outcomes at Scale, Lima, Peru 15