10.1177/1524838005283696TRAUMA, VIOLENCE, & ABUSE / January 2006Johnson et al. / YOUNG CHILDREN IN INSTITUTIONALCARE

YOUNG CHILDREN IN INSTITUTIONALCARE AT RISK OF HARM

REBECCA JOHNSONKEVIN BROWNECATHERINE HAMILTON-GIACHRITSISUniversity of Birmingham

Arecent survey has revealed a large number of young children in institutional careacross Europe. Young children placed in institutional care without parents may beat risk of harm. This review considers systematically the research evidence on theimpact of institutional care on brain growth, attachment, social behavior, and cog-nitive development. Analytical epidemiological study designs (i.e., including acontrol/comparison group) show that young children placed in institutional careare at risk of harm in terms of attachment disorder and developmental delays in so-cial, behavioral, and cognitive domains. Delays in physical growth, neural atro-phy, and abnormal brain development have also been implicated. The findingssuggest that the lack of a one-to-one relationship with a primary caregiver is a ma-jor cause of harm to children in residential care. Evidence indicates that infantswho are placed in institutional care will suffer harm to their development if they arenot moved to family-based care by the age of 6 months. The neglect and damagecaused by early privation and deprivation is equivalent to violence and policymakers should work to ensure that every child has the opportunity to grow up in afamily environment.

Key words: young children; residential settings; institutional care; abuse and neglect

A RECENT EUROPEAN COMMISSION“Daphne” project surveyed 32 European coun-tries (not including Russian speaking countries)and mapped the number and characteristics ofchildren under the age of 3 years in residentialcare (Browne et al., 2004). It was reported thatthere were 23,099 children, of a population of20.6 million under 3, in institutions for morethan 3 months without a parent. This represents11 children in every 10,000 under 3 years inresidential care institutions.

There was great variation between differentcountries for the proportion of children under 3

in institutional care. Four countries had none orless than 1% of children under 3 in institutions,12 countries had institutionalized between 1and 10 children per 10,000, seven countries hadbetween 11 and 30 children per 10,000 ininstitutions and alarmingly, eight countries hadbetween 31 and 60 children per 10,000 in institu-tions. Luxembourg could not provide informa-tion on the rate of children in institutions.

A UNICEF survey (2004) of Russian-speakingcountries formerly in the USSR (New Independ-ent States) showed that most have 30 or morechildren per 10,000 in “infant homes” (0-3

34

TRAUMA, VIOLENCE, & ABUSE, Vol. 7, No. 1, January 2006 34-60DOI: 10.1177/1524838005283696© 2006 Sage Publications

years). Comparable data for North America isdifficult to identify. On 30 September 2001,542,000 children (0-18 years) were in public(“foster”) care in the United States and approxi-mately one quarter (130,857) of these are under5 years (U.S. Department of Health and HumanServices, 2003). Furthermore, a report on “ChildWelfare Outcomes 2000” states that across the50 States, a median of 9% (range 1.3 Hawaii to27.2% Arizona) of children under 12 years inpublic (foster) care were placed in residentialchildren’s homes (U.S. Department of Healthand Human Services, 2001). Therefore, it can beestimated for the United States that approxi-mately 11,777 children under 5 years resided inresidential care institutions. Outside the devel-oped world of Europe and North America, theproblem of instititutionalized young children isvast and accurate statistics are unavailable.

The damaging consequences of institutionalcare have been known for many years. The pub-lications of Goldfarb (1945) and Bowlby (1951)were particularly influential and highlighted anumber of emotional, behavioral, and cognitiveimpairments that characterized individualswho had been raised in institutional care. Theseindividuals were reported to be intellectuallyretarded with specific difficulties in languagedevelopment. In addition, they had problemsconcentrating, forming emotional relation-ships, and were often described as attention-seeking. The lack of an attachment with amother figure during infancy was attributed asthe cause of these problems.

The influence of attachment theory (Bowlby,1969) emphasized the negative consequences ofinstitutional care compared to family-basedcare and the importance of a primary caregiverfor normal development. This led to a decline inthe use of institutional care or large children’shomes in some parts of world. In other parts ofworld, child care policy has been less influencedby the writings of Bowlby in terms of meetingthe psychosocial needs of children. Instead, anemphasis has been placed on the physical needsof children and controlling their environment.In such countries this has led to a reliance on in-stitutions rather than the development of substi-tute parenting, such as foster care and adoption(Browne, 2002). Furthermore, in some coun-

tries, advances in child-protection policy andprocedures that can remove parental rightshave sometimes progressed at a faster rate thanthe development of community services tomaintain children’s rights to be supported and/or rehabilitated into their families of origin oroffered alternative family-based care (UnitedNations, 1989). Therefore, children have beenplaced in hospital or residential care institutionsas a place of safety, often on a long-term basis.

In the 1990s, Romanian orphanages attractedthe attention of the media and researchers alikebecause of the devastatingly impoverished con-ditions in which many children were placed(Johnson, 2000). These unfortunate children

Johnson et al. / YOUNG CHILDREN IN INSTITUTIONAL CARE 35

KEY POINTS OF THE RESEARCH REVIEW• There are a significant number of young children

in residential care for more than 3 months with-out a parent in the European region at risk ofharm in terms of attachment disorder, develop-mental delay, and neural atrophy to the develop-ing brain.

• Children in institutional care have limited oppor-tunities to form selective attachments comparedto children in family-based care, especially wherethere are large numbers of children, small num-bers of staff, and a lack of consistent care throughshift work and staff rotation. Even apparently“good” institutional care can have a detrimentaleffect on children’s ability to form relationshipslater in life.

• Analytical epidemiological study designs (i.e.,including a control and/or comparison group)show that many of the problems observed insamples of severely deprived children, such asquasi-autistic behaviors and delays in cognitivedevelopment, show improvement once the childis removed from institutional care and placed in asupportive family environment. However, thequality of the subsequent family environment isan important factor in the outcome of institution-ally reared children.

• Research suggests that children who are movedfrom residential care before the age of 6 monthscan still reach optimal development. The pres-ence of attachment disorder is more common inchildren who have spent more than 6 months ininstitutional care.

• The neglect and damage caused by early priva-tion of parenting is equivalent to violence to ayoung child, as it inhibits optimal developmentand neglects the rights of the child to grow up in afamily.

provided a natural experiment on the effects ofsevere deprivation and provided a unique op-portunity for researchers to investigate whetherthe effects of such conditions can be reversed byplacing these children in the family-based carethey were deprived of in infancy. MichaelRutter and the English and Romanian AdopteesStudy Team at the Institute of Psychiatry in Lon-don (O’Connor et al., 1999, 2000a, 2000b; Rutteret al., 1998, 1999) have followed a large sampleof Romanian children who were adopted intothe United Kingdom after having spent theearly part of their life in such an institution. Sim-ilarly, research in British Columbia and Ontariohas reported on the progress of Romanian or-phans (although these so-called orphans aretypically social orphans rather than true or-phans) subsequently adopted into Canada(Fisher, Ames, Chisholm, & Savoie, 1997;Marcovitch et al., 1997).

The aim of this article is to review the litera-ture in a systematic way to consider the impactof early institutional care on neural develop-ment, attachment, social and/or behavioral de-velopment, and cognitive development. Ascoping exercise revealed that articles coveringthe neurobiological consequences of institu-tional care did not fit the inclusion criteria forthe review. Therefore, this article begins by con-sidering recent theoretical developments in thefield of neurobiology as background informa-tion to add to our understanding of why institu-tional care in the first few years of life isparticularly damaging for development.

NEUROBIOLOGICALCONSEQUENCES

The development ofthe brain in the postnatalperiod is truly stagger-ing; the human infant isborn with some 100-bil-lion neurons and eachneuron forms about15,000 synapses duringthe first few years of life(Balbernie, 2001). By theage of 3 the child hasformed about 1,000 tril-

lion synapses, this corresponds “to a rate of 1.8million new synapses per second between twomonths of gestation and two years after birth!”(Eliot, 2001, p. 27).

The overabundance of synapses and neuronsin the infant’s brain allows the adaptation ofthe brain in response to the environment(neuroplasticity). Synapses that are frequentlyused are reinforced, whereas redundant synap-ses are “pruned.” Thus, early experience deter-mines which neural pathways will become per-manent and which will be eliminated. However,for this process to result in normal brain devel-opment, the infant “must interact with a livingand responsive environment” (Balbernie, 2001).Specifically, a strong case has been proposed forthe maturation of the brain being “embedded inthe attachment relationship between the infantand the primary caregiver” (Schore, 2001a, p. 10).

The human infant is genetically predisposedto respond to a caregiver who will respond to,talk to, and handle them in a sensitive way andintroduce new stimuli in a manner that is safe,predictable, repetitive, gradual, and appropri-ate to the infant’s stage of development (Perry &Pollard, 1998). Thus, a sensitive caregiver and asecure attachment promote brain growth anddevelopment, whilst an impoverished environ-ment has the opposite effect and will suppressbrain development. Neglect and abuse in theearly years of life have the potential to affectsubsequent brain functioning; “neglect leads todeprivation of input needed by the infant brainat times of experience-expectant maturation,while abusive experiences affect brain develop-ment at experience-dependent stages” (Glaser,2000, p. 106).

Although abuse is certainly an issue with re-gard to children in residential care, this occursmore frequently in a family setting (Wolfe, Jaffe,& Jetté, 2003). Neglect, on the other hand, couldbe considered as a feature of typical institu-tional practice; institutional culture is primarilyconcerned with the physical care of childrenand the establishment of routine with little pro-vision for interaction with children (Giese &Dawes, 1999). Among other negative conse-quences, a lack of interaction and deprivation ofinput in the early years has obvious conse-quences for language development. Observa-

36 TRAUMA, VIOLENCE, & ABUSE / January 2006

The human infant isgeneticallypredisposed torespond to acaregiver who willrespond to, talk to,and handle them in asensitive way andintroduce new stimuliin a manner that issafe, predictable,repetitive, gradual,and appropriate tothe infant’s stage ofdevelopment.

tional studies, such as Giese and Dawes (1999)suggest that it is the “regulative” style of staff-child interactions that contributes primarily tolanguage delays seen in some institutionallyraised children. They observed interactionswithin the institutional setting and found thatmost interactions (83%) were highly regulativeand generally commands of a short duration(three seconds on average) that did not encour-age further interactions with the child. Tizard,Cooperman, Joseph, and Tizard (1972) similarlyreported that level of development was relatedto the quality of staff-child interaction.

A child raised in institutional care is typicallydeprived of the supportive, intensive, one-to-one relationship with a primary caregiver that isessential for optimal development. Without acaregiver to “scaffold“ infant learning, there isno process to guide synaptic connections andthe development of neural pathways. Schore(2001b) suggests that neglect leads to excessivepruning, which will result in neural and behav-ioral deficits. The neglect that is intrinsic in mostinstitutional settings is, therefore, damaging tobrain development and can cause regions of thebrain to atrophy (Balbernie, 2001).

Although these questions about critical peri-ods and recovery require more research, com-munication between developmental psycholo-gists and developmental neuroscientists hasbegun to lead to a greater understanding of howneglect and deprivation influence child devel-opment (Nelson et al., 2002). Future research,which takes a multidisciplinary approach (e.g.,Zeanah et al., 2003) offers the best chance foranswering these questions.

What is already clear is that the most sensi-tive period for brain development is the first 3years of life when the brain is in an unparalleledtime of developmental change (Schore, 2001a,2001b). Also, there is strong evidence that hu-man infants are born with a readiness to relate toothers and that engagement with sensitive oth-ers is essential for normal development(Trevarthen & Aitken, 2001). Neglect and abusein the early years of life, therefore, have the po-tential to affect adversely subsequent brainfunctioning (Glaser, 2000). Unfortunately, ne-

glect is typical of institutional practice and theroutine nature of institutional care does not en-courage the development of appropriate socialinteraction, language development, andautonomy (Giese & Dawes, 1999).

METHOD

The methods used for the review followed asystematic approach, with a search strategy andinclusion and/or exclusion criteria based onpopulation, intervention, comparator, and out-come (PICO). The approach, however, did notconsider the execution of the studies and noevaluation of bias or confounding influenceswas conducted.

Search Strategy

A search of published material (EMBASE1996 to 2003; MEDLINE 1996 to 2003; ISI Web ofScience 1998 to 2003; SOSIG 1998 to 2003; Sci-ence Direct 1998 to 2003) was conducted usingthe following terms: children and residentialcare, children and deprivation, children andprivation, early deprivation, early privation,children and orphanages, children andinstitutionalization, or institutionalization. Inaddition, the reference lists of relevant articleswere hand checked, and publications from au-thors known in the field were sought with visitsto experts in Romania. This generated 2,624hits, with research going back as far as 1944identified via reference lists. After taking intoaccount duplicates and inclusion criteria, therewere 27 hits.

Inclusion Criteria

The following PICO was developed for theidentification of studies to be included in thereview:

• Population—children 0-17 years• Intervention—children exposed to residential care

in an institution under the age of 5 years without aprimary caregiver for varying lengths of time

• Comparator—children exposed to family-basedcare with a primary caregiver for varying lengths oftime

Johnson et al. / YOUNG CHILDREN IN INSTITUTIONAL CARE 37

(text continues on page 41)

38

TAB

LE

1:

Su

mm

ary

of

Res

earc

h S

tud

ies

Inve

stig

atin

g A

ttac

hm

ent

in C

hild

ren

Rai

sed

in In

stit

uti

on

al C

are

Stu

dyIn

stitu

tion

Sam

ple

Inst

itutio

nal C

hara

cter

istic

sD

escr

iptio

n of

Stu

dyF

indi

ngs

Gol

dfar

b (1

944)

n=

15

Mal

e-fe

mal

e 1.

1:1

Age

10-

14 y

ears

IT 2

7-47

mon

ths

Not

des

crib

ed (

see

Gol

dfar

b,19

45)

Cap

acity

for

rela

tions

hips

was

mea

sure

d in

adol

esce

nts

who

had

spe

nt th

eir

early

in-

fanc

y in

inst

itutio

nal c

are

but w

ho h

ad s

ub-

sequ

ently

bee

n fo

ster

ed.T

hese

chi

ldre

nw

ere

com

pare

d w

ith a

mat

ched

com

paris

ongr

oup

who

wer

e in

fost

er c

are

and

had

been

in fa

mily

-bas

ed c

are

sinc

e bi

rth.

In c

ompa

rison

to th

e “f

oste

r”gr

oup

the

in-

stitu

tiona

lly r

aise

d ad

oles

cent

s w

ere

emot

iona

lly w

ithdr

awn

in r

elat

ions

hips

and

unpo

pula

r bu

t cra

ving

affe

ctio

n.

Rhe

ingo

ld a

nd B

ayle

y(1

959)

n=

14

Mal

e-fe

mal

e 1:

1A

ge 1

7-22

mon

ths

IT 4

-18

mon

ths

Not

des

crib

ed.

Two

grou

ps o

f chi

ldre

n in

inst

itutio

nal c

are

wer

e co

mpa

red

in a

n ex

perim

enta

l situ

a -tio

n.H

alf t

he c

hild

ren

rece

ived

2 m

onth

s of

care

from

a s

ingl

e ca

regi

ver.

The

com

pari -

son

grou

p w

as c

ompl

etel

y re

ared

und

er in

-st

itutio

nal r

outin

e.T

he s

ocia

l res

pons

ive -

ness

of t

he c

hild

ren

was

test

ed a

fter

the

expe

rimen

tal p

hase

and

18

mon

ths

late

r.T

he s

ocia

l res

pons

iven

ess

of c

hild

ren

sub-

sequ

ently

ado

pted

or

rest

ored

was

als

oco

mpa

red.

Afte

r th

e 2

mon

ths

of h

avin

g a

sing

le c

are -

give

r, th

e ex

perim

enta

l gro

up w

as r

ated

as m

ore

soci

ally

res

pons

ive.

How

ever

,th

is w

as n

ot m

aint

aine

d 18

mon

ths

late

r.A

dopt

ed c

hild

ren

gave

mor

e po

sitiv

e re

-sp

onse

s th

an r

esto

red

child

ren

did

but

this

diff

eren

ce w

as n

ot s

igni

fican

t.T

hech

ildre

n st

ill in

inst

itutio

nal c

are

wer

efr

iend

lier

to s

tran

gers

than

chi

ldre

n w

hoha

d be

en m

oved

from

inst

itutio

nal t

ofa

mily

-bas

ed c

are.

Tiz

ard

and

Jose

ph(1

970)

n=

30

Mal

e-fe

mal

e 1:

1A

ge 2

yea

rsIT

4-2

4 m

onth

s

Sta

ff to

chi

ld r

atio

1:3

but

hig

hst

aff t

urno

ver.

Boo

ks, t

oys,

pla

yfa

cilit

ies.

Mix

ed a

ge g

roup

s.H

ome

visi

ts a

nd o

utin

gs.P

er-

sona

l rel

atio

nshi

ps d

isco

urag

ed.

Goo

d st

anda

rd o

f phy

sica

l car

e.

Res

pons

e to

str

ange

rs a

nd to

sep

arat

ion

was

mea

sure

d in

a s

ampl

e of

chi

ldre

nra

ised

in a

UK

res

iden

tial n

urse

ry.A

sam

ple

of “

wor

king

-cla

ss”c

hild

ren

was

use

d as

aco

ntro

l gro

up.

The

chi

ldre

n ra

ised

at h

ome

wer

e si

gnifi

-ca

ntly

mor

e fr

iend

ly to

a s

tran

ger

than

the

resi

dent

ial n

urse

ry c

hild

ren.

Whe

n le

ftal

one

with

a s

tran

ger,

the

resi

dent

ial

nurs

ery

child

ren

wer

e m

ore

likel

y to

run

out o

f the

roo

m;n

one

of th

e ho

me

chil-

dren

did

this

.

Wol

kind

(19

74)

n=

92

Mal

e-fe

mal

e 1.

7:1

Age

5-1

2 ye

ars

IT 6

mon

ths-

6 ye

ars

Not

des

crib

ed.C

hild

ren’

s ho

me

inth

e U

nite

d K

ingd

om.

A p

sych

iatr

ic s

tudy

of c

hild

ren

who

wer

elo

ng-s

tay

resi

dent

s in

a U

K in

stitu

tion

was

carr

ied

out.

Sym

ptom

s of

chi

ldre

n w

how

ere

adm

itted

bef

ore

the

age

of 2

yea

rsw

as c

ompa

red

with

thos

e of

chi

ldre

n ad

mit-

ted

afte

r th

is a

ge.

The

re w

ere

diffe

renc

es b

etw

een

the

two

grou

ps fo

r di

sinh

ibiti

on;t

he c

hild

ren

ad-

mitt

ed b

efor

e th

e ag

e of

2 y

ears

wer

e“o

verf

riend

ly.”

Tiz

ard

and

Ree

s (1

975)

n=

26

Mal

e-fe

mal

e 2.

3:1

Age

4 a

nd a

hal

f yea

rsIT

24-

48 m

onth

s

Sta

ff to

chi

ld r

atio

1:3

but

hig

hst

aff t

urno

ver.

Boo

ks, t

oys,

pla

yfa

cilit

ies.

Mix

ed a

ge g

roup

s.H

ome

visi

ts a

nd o

utin

gs.P

er-

sona

l rel

atio

nshi

ps d

isco

urag

ed.

Goo

d st

anda

rd o

f phy

sica

l car

e.

Affe

ctio

nal b

onds

wer

e as

sess

ed in

a s

am-

ple

of c

hild

ren

rais

ed in

a U

K r

esid

entia

lnu

rser

y w

as m

easu

red.

Chi

ldre

n w

ere

ei-

ther

stil

l in

nurs

ery,

ado

pted

, or

rest

ored

toth

eir

natu

ral p

aren

ts.A

sam

ple

of “

wor

king

-cl

ass”

child

ren

was

use

d as

a c

ontr

ol g

roup

.

The

chi

ldre

n st

ill in

nur

sery

car

e w

ere

de-

scrib

ed b

y st

aff a

s sh

allo

w, e

mot

iona

llyde

tach

ed, a

nd w

ere

thou

ght n

ot to

car

ede

eply

abo

ut a

nyon

e.T

he a

dopt

ed c

hil-

dren

wer

e th

ough

t by

thei

r ad

opte

d m

oth-

ers

to b

e de

eply

atta

ched

to th

em.

39

Tiz

ard

and

Hod

ges

(197

8)n

= 5

1M

ale-

fem

ale

1.8:

1A

ge 8

yea

rsIT

24-

48 m

onth

s

Sam

e as

abo

ve.

Atta

chm

ent b

ehav

ior

and

pare

nt r

atin

gs o

f atta

ch-

men

t in

a sa

mpl

e of

chi

ldre

n ra

ised

in a

UK

res

i -de

ntia

l nur

sery

wer

e m

easu

red.

Som

e ch

ildre

nw

ere

still

in th

e nu

rser

y bu

t mos

t had

bee

nad

opte

d or

res

tore

d to

thei

r na

tura

l par

ents

.Asa

mpl

e of

“w

orki

ng-c

lass

”chi

ldre

n w

as u

sed

as a

cont

rol g

roup

.

Com

pare

d to

the

cont

rol g

roup

, the

ex-

inst

itutio

nal c

hild

ren

wer

e m

ore

ofte

nra

ted

as o

ver-

frie

ndly

.84%

of t

hem

othe

rs o

f ado

pted

chi

ldre

n be

lieve

dth

eir

child

to b

e at

tach

ed (

50%

of r

e -st

ored

mot

hers

).T

he e

nviro

nmen

t of

the

rest

ored

chi

ldre

n is

des

crib

ed a

sm

uch

less

favo

rabl

e as

that

exp

eri -

ence

d by

the

adop

ted

child

ren.

Hod

ges

and

Tiz

ard

(198

9a)

n=

42

Mal

e-fe

mal

e 2:

1A

ge 1

6 ye

ars

IT 2

4-48

mon

ths

Sam

e as

abo

ve.

The

fam

ily r

elat

ions

hips

of e

x-in

stitu

tiona

l ado

les -

cent

s ra

ised

unt

il at

leas

t the

age

of 2

in a

UK

res

i -de

ntia

l nur

sery

was

mea

sure

d at

16

year

s.T

head

just

men

t of t

he c

hild

ren

who

had

sub

sequ

ently

been

ado

pted

and

res

tore

d w

as c

ompa

red

and

also

with

a c

ontr

ol g

roup

of a

dole

scen

ts.

Clo

se a

ttach

men

ts w

ere

muc

h m

ore

likel

y am

ong

the

adop

ted

child

ren

and

thei

r ad

optiv

e pa

rent

s th

an b

etw

een

the

rest

ored

chi

ldre

n an

d th

eir

natu

ral

pare

nts.

Mar

covi

tch

et a

l.(1

997)

n=

56

Mal

e-fe

mal

e 1:

1.1

Age

3-5

yea

rsIT

0-<

6 m

onth

s, 6

-48

mon

ths

Sev

erel

y de

priv

ed e

nviro

n -m

ent.

Nut

ritio

nal a

nd p

sy-

chol

ogic

al p

rivat

ion.

Har

shph

ysic

al c

ondi

tions

.

Atta

chm

ent s

tatu

s w

as m

easu

red

in a

sam

ple

ofR

oman

ian

orph

ans

who

had

sub

sequ

ently

bee

nad

opte

d in

Can

ada.

The

out

com

e fo

r ch

ildre

n w

hoha

d sp

ent l

ess

than

6 m

onth

s in

an

orph

anag

ew

as c

ompa

red

with

that

of c

hild

ren

who

had

spe

ntlo

nger

than

6 m

onth

s in

an

inst

itutio

n.A

con

trol

sam

ple

from

ano

ther

stu

dy o

f hea

lthy

4-ye

ar-o

lds

was

use

d as

a c

ompa

rison

for

the

atta

chm

ent

mea

sure

.

The

ado

ptee

s w

ere

very

diff

eren

t to

the

cont

rol s

ampl

e.T

he r

ate

of s

ecur

e at

-ta

chm

ent w

as s

igni

fican

tly lo

wer

for

the

adop

tees

(30

% v

s.42

%).

Thi

s di

f-fe

renc

e m

ay b

e gr

eate

r as

ther

e m

ayha

ve b

een

“fal

se s

ecur

es”a

mon

g th

ead

opte

es.A

void

ant a

ttach

men

t, w

hich

was

the

mos

t com

mon

form

of i

nse-

cure

atta

chm

ent i

n th

e co

ntro

l gro

up,

was

abs

ent a

mon

g th

e ad

opte

es, w

how

ere

ambi

vale

nt o

r di

sorg

aniz

ed in

thei

r at

tach

men

t.

Chi

shol

m (

1998

)n

= 4

6M

ale-

fem

ale

1:1.

2A

ge 5

3-55

mon

ths

IT 8

-53

mon

ths

Sta

ff to

chi

ld r

atio

1:1

0-20

.S

ever

ely

depr

ived

env

i-ro

nmen

t.N

utrit

iona

l and

psyc

holo

gica

l priv

atio

n.H

arsh

phy

sica

l con

ditio

ns.

Atta

chm

ent a

nd in

disc

rimin

ate

frie

ndlin

ess

wer

eas

sess

ed a

t 30

mon

ths

and

54 m

onth

s in

a s

am-

ple

of R

oman

ian

orph

ans

(RO

) w

ho h

ad s

ubse

-qu

ently

bee

n ad

opte

d in

Can

ada.

The

se c

hild

ren

wer

e co

mpa

red

with

a m

atch

ed g

roup

of C

anad

ian

born

(C

B)

child

ren

(non

adop

ted)

and

a g

roup

of

Rom

ania

n ch

ildre

n w

ho h

ad b

een

adop

ted

befo

re4

mon

ths

(EA

).

RO

chi

ldre

n di

d no

t diff

er fr

om C

B a

ndE

A o

n at

tach

men

t sec

urity

(pa

rent

alre

port

) bu

t the

y di

d di

spla

y m

ore

inse

-cu

re a

ttach

men

t pat

tern

s th

an th

eot

her

two

grou

ps.R

O c

hild

ren

wer

esi

gnifi

cant

ly m

ore

likel

y to

sho

w in

dis-

crim

inat

ely

frie

ndly

beh

avio

r th

an th

eC

B a

nd E

A c

hild

ren;

the

latte

r tw

ogr

oups

did

not

sho

w a

ny s

igni

fican

tdi

ffere

nces

.

(con

tinu

ed)

40

O’C

onno

r et

al.

(199

9)n

= 1

11M

ale-

fem

ale

1:1.

2A

ge 4

yea

rsIT

< 6

mon

ths,

6-<

24m

onth

s

Sev

erel

y de

priv

ed e

nviro

n -m

ent.

Nut

ritio

nal a

nd p

sy-

chol

ogic

al p

rivat

ion.

Har

shph

ysic

al c

ondi

tions

.

Atta

chm

ent d

isor

der

(dis

inhi

bite

d be

havi

or, e

.g.,

wou

ld r

eadi

ly g

o of

f with

a s

tran

ger)

and

beh

av-

iora

l and

em

otio

nal p

robl

ems

wer

e m

easu

red

in a

sam

ple

of R

oman

ian

orph

ans

(RO

) w

ho h

ad s

ub-

sequ

ently

bee

n ad

opte

d in

the

Uni

ted

Kin

gdom

.D

urat

ion

of d

epriv

atio

n w

as c

ompa

red

with

out

-co

me.

A g

roup

of U

K-a

dopt

ed c

hild

ren

not e

x -po

sed

to d

epriv

atio

n w

as u

sed

as a

com

paris

ongr

oup.

A s

tron

g re

latio

nshi

p w

as fo

und

be-

twee

n du

ratio

n of

dep

rivat

ion

and

at-

tach

men

t dis

orde

r be

havi

ors,

but

70%

of th

e ch

ildre

n ex

pose

d to

mor

e th

an2

year

s of

dep

rivat

ion

did

not e

xhib

itse

vere

atta

chm

ent d

isor

der

whe

reas

som

e ch

ildre

n on

ly d

epriv

ed in

the

early

mon

ths

did.

RO

sho

wed

mor

e at

-ta

chm

ent d

isor

der

than

the

UK

com

-pa

rison

gro

up.

Sm

yke,

Dum

itres

cu, a

ndZ

eana

h (2

002)

n=

32

Mal

e-fe

mal

e ra

tio n

otst

ated

Age

4-6

8 m

onth

sIT

4-6

8 m

onth

s

Sta

ff to

chi

ld r

atio

1:10

.M

ultip

le c

areg

iver

s.C

hil -

dren

spe

nd m

ost o

f day

in o

ne la

rge

room

or

outs

ide.

Inhi

bite

d an

d di

sinh

ibite

d at

tach

men

t dis

orde

r w

asin

vest

igat

ed w

ere

mea

sure

d in

chi

ldre

n in

“st

an-

dard

”Rom

ania

n in

stitu

tiona

l car

e, in

chi

ldre

n re

-ce

ivin

g “p

ilot”

care

(m

ore

cons

iste

ncy

of c

are-

give

rs)

and

a co

ntro

l gro

up o

f chi

ldre

n in

day

car

ebu

t who

had

nev

er b

een

plac

ed in

an

inst

itutio

n.

Inhi

bite

d an

d di

sinh

ibite

d at

tach

men

tbe

havi

ors

wer

e ob

serv

ed s

igni

fican

tlym

ore

in th

e ch

ildre

n in

sta

ndar

d ca

reco

mpa

red

to th

e ot

her

two

grou

ps.

The

pilo

t gro

up d

ispl

ayed

mor

edi

sinh

ibite

d be

havi

or th

an th

e co

ntro

lgr

oup

thou

gh th

e di

ffere

nces

wer

e no

tsi

gnifi

cant

.

O’C

onno

r et

al.

(200

0a)

n=

165

Mal

e-fe

mal

e 1:

1.2

Age

6 y

ears

IT <

6 m

onth

s, 6

-< 2

4m

onth

s, 2

4-42

mon

ths

Sam

e as

abo

ve.

Atta

chm

ent d

isor

der

(dis

inhi

bite

d be

havi

or, e

.g.,

wou

ld r

eadi

ly g

o of

f with

a s

tran

ger)

and

beh

av-

iora

l and

em

otio

nal p

robl

ems

wer

e m

easu

red

in a

sam

ple

of R

oman

ian

orph

ans

(RO

) w

ho h

ad s

ub-

sequ

ently

bee

n ad

opte

d in

the

Uni

ted

Kin

gdom

.D

urat

ion

of d

epriv

atio

n w

as c

ompa

red

with

out

-co

me.

A g

roup

of U

K-a

dopt

ed c

hild

ren

not e

x-po

sed

to d

epriv

atio

n w

as u

sed

as a

com

paris

ongr

oup.

Atta

chm

ent d

isor

der

corr

elat

ed w

ithat

tent

iona

l and

con

duct

pro

blem

s bu

tap

pear

s to

be

a di

stin

ct s

et o

f beh

av-

iors

.A s

tron

g re

latio

nshi

p w

as fo

und

betw

een

dura

tion

of d

epriv

atio

n an

dat

tach

men

t dis

orde

r be

havi

ors,

but

70%

of t

he c

hild

ren

expo

sed

to m

ore

than

2 y

ears

of d

epriv

atio

n di

d no

t ex-

hibi

t sev

ere

atta

chm

ent d

isor

der

whe

re a

s so

me

child

ren

only

dep

rived

in th

e ea

rly m

onth

s di

d.R

O s

how

edm

ore

atta

chm

ent d

isor

der

than

the

UK

com

paris

on g

roup

.

NO

TE

:nre

fers

to th

e in

stitu

tiona

l sam

ple

subj

ects

onl

y;th

e st

udie

s ha

d a

sim

ilar

num

ber

of c

ompa

rison

or

cont

rol s

ubje

cts.

IT =

ran

ge o

f tim

e sp

ent i

n in

stitu

tiona

l car

e.

TAB

LE

1(c

on

tinu

ed)

Stu

dyIn

stitu

tion

Sam

ple

Inst

itutio

nal C

hara

cter

istic

sD

escr

iptio

n of

Stu

dyF

indi

ngs

• Outcome—child to primary caregiver attachmentpatterns, social and behavioural development, cog-nitive development

Studies were only selected where there wasevidence of a control or comparison group-study design. Attachment (12 studies), socialand/or behavioral (17 studies), and cognitive(13 studies) domains were addressed sepa-rately. Thus, some studies that researched morethan one domain are included in more than onetable.

For each domain, a table summarizes thefindings from a number of studies that have in-vestigated the consequences for children raisedin institutional care during their early yearscompared to a control or comparison sample. Inthe text, some of these studies are described inmore detail, focusing in particular on longitudi-nal studies with matched control groups.

CONSEQUENCES FOR ATTACHMENT

A summary of studies that have investigatedrelationships and attachment in children raisedin institutional care is provided in Table 1. Nineof these studies report specifically on indiscrim-inate friendliness, overfriendliness, and/ordisinhibited behavior. Eight of these nine stud-ies report that children raised in institutionalcare exhibited these features of a disordered at-tachment more than controls or children whowere admitted to institutional care after the ageof two years (Wolkind, 1974).

Tizard and Rees (1975) investigatedaffectional bonds in a sample of UK childrenwho had spent their early years in a residentialnursery. Although the conditions in the nurserywere good, there was a high turnover of staffand the staff group was discouraged from hav-ing close relationships with the children. At agefour, the staff reported that the children were“not deeply attached to anyone.” For the insti-tutional children who had been adopted by theage of 4, a third of them were reported asoverfriendly to strangers by their adoptive par-ents. Similarly, at age 8, ex-institutional childrenwere more often described as overfriendly incomparison to other children (Tizard & Rees,1975). The overfriendliness had attenuated by

the age of 16, although the ex-institutional teen-agers were still more oriented toward adultattention and approval. At age 16, the ex-insti-tutional children (children adopted and chil-dren restored to their natural family) showedmore problems with peer relationships and re-lationships with adults outside the family com-pared to other teenagers. In terms of family rela-tionships, only children restored to their naturalfamily were more likely to have difficulties andpoor family relationships. The adopted childrendid not differ in their family relationships fromother teenagers. Thus, institutional care with alack of secure attachments in the early years hadnot resulted in an inability to form close rela-tionships. However, the formation of subse-quent attachments does not occur automaticallyby placing the child in a family setting. The criti-cal factor appeared to be whether “the parentwanted the child and was able to put a lot intothe relationship” (Hodges & Tizard, 1989a).

Later research has sought to investigate rela-tionships in institutionally raised childrenwithin the framework of attachment theory. Forexample, Marcovitch et al. (1997) investigatedchild-parent attachment in a sample of childrenwho had been adopted from Romanian institu-tions into Canada and who had experienced“poor” conditions of institutional care. The chil-dren had been deprived of basic physical, emo-tional, and nutritional needs. The opportunityfor these children to form any sort of relation-ship with a caregiver was extremely limited.Marcovitch et al. (1997) assessed attachment inthese children with the strange situation test(Ainsworth, Blehar, Waters, & Wall, 1978) usingthe classification scheme developed by Cassidy,Marvin, and Attachment Working Group of theMacArthur Network on the Transition from In-fancy to Early Childhood (1987, 1992) for usewith preschoolers. The rate of secure attach-ment in the adoptee group was significantlylower than in the comparison group (30% vs.42%). However, the actual difference may bemuch greater as the previously institutional-ized children may have been incorrectly catego-rized as secure. The coding system used for as-sessing attachment was based on parent-childreunion and did not consider response tostrangers. However, the indiscriminate friendli-

Johnson et al. / YOUNG CHILDREN IN INSTITUTIONAL CARE 41

ness toward strangers seen in these children isincompatible with secure attachment status.This pattern of behavior suggests a disorganiza-tion of the attachment-behavioral system andmight suggest the presence of a disinhibited-at-tachment disorder (O’Connor et al., 1999). Chil-dren described as having a disinhibited-attach-ment disorder show indiscriminately friendlybehavior toward strangers and approach peo-

ple with whom they donot have a close relation-ship when distressed. Ithas been suggested thatthis represents a “disor-ganization” of the attach-ment system, which is adifferent and perhapsmore serious problemthan “insecure” attach-ment (O’Connor et al.,1999).

O’Connor et al. (1999)investigated attachment-disorder behaviors intheir sample of Romanianorphans who had beenadopted in the United

Kingdom. They found that the duration of thedeprivation experienced by their sample of 4-year-old Romanian orphans was positively as-sociated with attachment disorder behaviors(e.g., lack of checking with parents, clear indica-tion that child would readily go off with astranger). They also point out, however, that notall children who had experienced prolongeddeprivation display these behaviors.

Smyke, Dumitrescu, and Zeanah (2002) inves-tigated inhibited and disinhibited-attachmentdisorder in three groups of Romanian children.The first group received standard institutionalcare, which involved more than 20 differentstaff members caring for a large group of chil-dren in rotating shifts. The second group of chil-dren was in the same institution but receivedcare on a “pilot unit.” In the pilot unit, a smallerpool of staff was used so that instead of 20 in-consistent caregivers there were 4 consistentcarers. Children in the pilot unit were alsohoused in smaller groups (10 to 12 rather than30 to 35 in the standard unit), so that each group

had one main consistent caregiver. The thirdgroup was a control group of children who wereattending day care but who had never beenplaced in an institution. The group receiving thestandard institutional care had significantlyhigher scores for the signs of both inhibited anddisinhibited attachment disorders than theother two groups did. There were no significantdifferences between the pilot care group and thecontrol group for inhibited behaviors but therewere some significant differences for measuresof indiscriminate behavior; the pilot group hadhigher scores and exhibited more indiscrimi-nate behavior.

In terms of attachment, even apparently“good” institutional care can have a detrimentaleffect on children’s ability to form relationshipslater in life. The lack of a warm and continuousrelationship with a sensitive caregiver can pro-duce children who are desperate for adult at-tention and affection. Superficially, the behav-ior of these children can seem “normal,” andsome earlier classifications of attachmentmight label them as secure rather than disorga-nized and/or disorientated (Carlson, Cicchetti,Barnett, & Braunwald, 1989; Zeanah, 2000).However, their lack of discrimination is indica-tive of an attachment disorder (Carlson et al.,1989; Zeanah, 2000). The presence of attachmentdisorder is more common in children who havespent more than 6 months in institutional care(O’Connor et al., 1999, 2000a). However, thispattern is not an inevitable consequence of earlydeprivation and there are mediating factors thatcan ameliorate negative effects, such as the childbeing a particular favorite of a residential careworker and as a result receiving sensitive caregiving. Nevertheless, children in institutionalcare clearly have limited opportunities to formselective attachments compared to children infamily-based care, especially where there arelarge numbers of children, small numbers ofstaff, and a lack of consistent care through shiftwork and staff rotation.

CONSEQUENCES FOR SOCIAL ANDBEHAVIORAL DEVELOPMENT

Research investigating the development ofchildren who have been raised in institutions

42 TRAUMA, VIOLENCE, & ABUSE / January 2006

(text continues on page 48)

In terms ofattachment, evenapparently “good”institutional care canhave a detrimentaleffect on children’sability to formrelationships later inlife. The lack of awarm and continuousrelationship with asensitive caregivercan produce childrenwho are desperatefor adult attentionand affection.

43

TAB

LE

2:

Su

mm

ary

of

Res

earc

h S

tud

ies

Inve

stig

atin

g t

he

So

cial

an

d B

ehav

iora

l Dev

elo

pm

ent

of

Ch

ildre

n R

aise

d in

Inst

itu

tio

nal

Car

e

Stu

dyIn

stitu

tion

Sam

ple

Inst

itutio

nal C

hara

cter

istic

sD

escr

iptio

n of

Stu

dyF

indi

ngs

Gol

dfar

b (1

944)

n=

15

Mal

e-fe

mal

e 1.

1:1

Age

10-

14 y

ears

IT 2

7-47

mon

ths

Not

des

crib

ed (

see

Gol

dfar

b,19

45).

Per

sona

lity,

pro

blem

beh

avio

r, an

d so

cial

mat

urity

wer

e m

easu

red

in a

dole

scen

tsw

ho h

ad s

pent

thei

r ea

rly in

fanc

y in

inst

itu-

tiona

l car

e bu

t who

had

sub

sequ

ently

bee

nfo

ster

ed.T

hese

chi

ldre

n w

ere

com

pare

dw

ith a

mat

ched

com

paris

on g

roup

who

wer

e in

fost

er c

are

and

had

been

in fa

mily

-ba

sed

care

sin

ce b

irth

.

In c

ompa

rison

to th

e “f

oste

r”gr

oup,

the

in-

stitu

tiona

lly r

aise

d ad

oles

cent

s w

ere

“ap -

preh

ensi

ve,”

“apa

thet

ic,”

rest

less

and

hy -

pera

ctiv

e, a

nd le

ss s

ocia

lly m

atur

e.

Gol

dfar

b (1

945)

n=

15

Mal

e-fe

mal

e 1.

5:1

Age

43

mon

ths

IT 4

-32

mon

ths

Not

des

crib

ed in

det

ail “

Adu

lt-ch

ild r

atio

is v

ery

low

so

that

ther

e is

a m

inim

um o

f adu

lt st

im-

ulat

ion,

”“th

e ch

ild’s

act

iviti

es a

reco

mpl

etel

y re

gula

ted

...H

e is

not e

ncou

rage

d to

par

ticip

ate

inth

e fo

rmul

atio

n of

his

ow

n da

y to

day

prog

ram

.”

The

beh

avio

r an

d so

cial

mat

urity

of c

hild

ren

rais

ed in

inst

itutio

nal c

are

in th

e U

nite

dK

ingd

om w

as te

sted

at 3

yea

rs.T

he c

hil -

dren

wer

e th

en p

lace

d in

fost

er h

omes

,an

d a

follo

w-u

p te

st c

ondu

cted

9 m

onth

saf

ter

the

first

test

.The

se c

hild

ren

wer

eco

mpa

red

with

a m

atch

ed c

ompa

rison

grou

p w

ho w

ere

in fo

ster

car

e an

d ha

dbe

en in

fam

ily-b

ased

car

e si

nce

birt

h.

At t

he fi

rst t

estin

g th

e in

stitu

tiona

lly r

eare

dch

ildre

n an

d th

e fo

ster

car

e ch

ildre

n ha

dsi

mila

r sc

ores

for

soci

al m

atur

ity.A

t the

seco

nd te

st a

fter

the

inst

itutio

n ch

ildre

nha

d al

so b

een

fost

ered

, how

ever

, the

scor

es o

f the

inst

itutio

n gr

oup

impr

oved

.T

his

is in

terp

rete

d as

a tr

aum

a fo

llow

ing

sepa

ratio

n fr

om th

e fa

mili

ar in

stitu

tiona

len

viro

nmen

t.In

the

beha

vior

rat

ings

, the

fost

er c

are

child

ren

wer

e ra

ted

as m

ore

fa-

vora

ble

than

the

inst

itutio

n gr

oup,

but

at

the

seco

nd te

st th

ere

wer

e no

diff

eren

ces.

Wol

kind

and

Rut

ter

(197

3)n

= 7

8M

ale-

fem

ale

20:1

Age

10-

11 y

ears

IT a

t lea

st o

ne w

eek

Not

des

crib

ed.V

arie

ty o

f UK

res

i-de

ntia

l ins

titut

ions

.A

pop

ulat

ion

sam

ple

of 1

0- to

11-

year

-old

child

ren

in tw

o Lo

ndon

bor

ough

s w

assc

reen

ed u

sing

teac

her

mea

sure

s of

be-

havi

oral

pro

blem

s.A

ran

dom

sam

ple

of th

e“d

evia

nt”c

hild

ren

was

inve

stig

ated

furt

her

in c

ompa

rison

to a

con

trol

gro

up o

f “no

n-de

vian

t”ch

ildre

n.In

form

atio

n w

as c

olle

cted

abou

t any

per

iods

of p

aren

tal s

epar

atio

n(e

.g.,

plac

emen

t with

fost

er p

aren

ts o

r in

ach

ildre

n’s

hom

e).

Chi

ldre

n ex

perie

ncin

g sh

ort-

term

inst

itu-

tiona

l car

e w

ere

foun

d to

be

at r

isk

for

anti-

soci

al d

isor

der;

a si

gnifi

cant

ly la

rger

num

-be

r of

chi

ldre

n in

the

devi

ant g

roup

had

been

“in

car

e”th

an in

the

rand

omly

se-

lect

ed c

ontr

ol g

roup

.The

vas

t maj

ority

of

the

devi

ant g

roup

was

mal

e.P

erio

ds in

care

wer

e ty

pica

lly v

ery

brie

f, an

d in

stitu

-tio

nal c

are

is r

ejec

ted

by th

e au

thor

s as

lead

ing

to th

e pr

oble

ms

seen

in th

ese

chil-

dren

.Fam

ily d

isco

rd is

sug

gest

ed a

s a

mor

e lik

ely

expl

anat

ion,

and

it is

sug

gest

edth

at b

oys

are

mor

e su

scep

tible

to th

is ty

peof

str

ess.

Wol

kind

(19

74)

n=

92

Mal

e-fe

mal

e 1.

7:1

Age

5-1

2 ye

ars

IT 6

mon

ths-

6 ye

ars

Not

des

crib

ed.C

hild

ren’

s ho

me

inth

e U

nite

d K

ingd

om.

A p

sych

iatr

ic s

tudy

of c

hild

ren

who

wer

elo

ng-s

tay

resi

dent

s in

a U

K in

stitu

tion

was

carr

ied

out.

Sym

ptom

s of

chi

ldre

n w

how

ere

adm

itted

bef

ore

the

age

of 2

yea

rsw

as c

ompa

red

with

thos

e of

chi

ldre

n ad

-m

itted

afte

r th

is a

ge.

The

re w

as n

o di

ffere

nce

betw

een

child

ren

adm

itted

bef

ore

the

age

of 2

and

chi

ldre

nad

mitt

ed a

fter

this

age

for

“affe

ctio

nles

sps

ycho

path

y”(e

.g.,

antis

ocia

l dis

orde

r).I

tis

sug

gest

ed th

at th

is c

ondi

tion

is p

rimar

ilyth

e re

sult

of fa

mily

fact

ors.

(con

tinu

ed)

44

Tiz

ard

and

Ree

s(1

975)

n=

26

Mal

e-fe

mal

e 2.

3:1

Age

4 a

nd a

hal

f yea

rsIT

24-

48 m

onth

s

Sta

ff to

chi

ld r

atio

1:3

but

hig

hst

aff t

urno

ver.

Boo

ks, t

oys,

pla

yfa

cilit

ies.

Mix

ed a

ge g

roup

s.H

ome

visi

ts a

nd o

utin

gs.P

er-

sona

l rel

atio

nshi

ps d

isco

urag

ed.

Goo

d st

anda

rd o

f phy

sica

l car

e.

Beh

avio

r pr

oble

ms

wer

e m

easu

red

in a

sam

ple

of c

hild

ren

rais

ed in

a U

K r

esid

en-

tial n

urse

ry.C

hild

ren

wer

e ei

ther

stil

l in

nurs

ery,

ado

pted

, or

rest

ored

to th

eir

natu

-ra

l par

ents

.A s

ampl

e of

“w

orki

ng-c

lass

”ch

ildre

n w

as u

sed

as a

con

trol

gro

up.

The

res

iden

tial n

urse

ry c

hild

ren

and

the

cont

rol g

roup

sho

wed

diff

eren

t pat

tern

s of

beha

vior

pro

blem

s.N

urse

ry g

roup

:poo

rco

ncen

trat

ion,

pee

r pr

oble

ms,

tem

per

tan -

trum

s, a

nd c

lingi

ng.T

he “

wor

st”p

robl

emsc

ores

wer

e fr

om c

hild

ren

with

an

irreg

ular

or b

roke

n pa

rent

con

tact

.The

low

est

scor

es w

ere

from

the

adop

ted

child

ren.

Res

tore

d ch

ildre

n w

ere

rate

d as

the

mos

tat

tent

ion-

seek

ing.

Tiz

ard

and

Hod

ges

(197

8)n

= 5

1M

ale-

fem

ale

1.8:

1A

ge 8

yea

rsIT

24-

48 m

onth

s

Sam

e as

abo

ve.

Beh

avio

r pr

oble

ms

wer

e m

easu

red

in a

sam

ple

of c

hild

ren

rais

ed in

a U

K r

esid

en-

tial n

urse

ry.S

ome

child

ren

wer

e st

ill in

the

nurs

ery

but m

ost h

ad b

een

adop

ted

or r

e-st

ored

to th

eir

natu

ral p

aren

ts.A

sam

ple

of“w

orki

ng-c

lass

”chi

ldre

n w

as u

sed

as a

cont

rol g

roup

.

Com

pare

d to

the

cont

rol g

roup

the

ex-

inst

itutio

nal c

hild

ren

wer

e m

ore

ofte

n ra

ted

as a

ttent

ion

seek

ing,

in a

dditi

on th

e re

-st

ored

chi

ldre

n ha

d a

varie

ty o

f ner

vous

habi

ts.E

x-in

stitu

tiona

l chi

ldre

n w

ere

mor

elik

ely

to b

e de

scrib

ed b

y te

ache

rs a

s di

s-ob

edie

nt, r

estle

ss, a

nd p

oor

at p

eer

rela

-tio

ns th

an th

e co

ntro

l chi

ldre

n.O

f the

re-

stor

ed c

hild

ren,

66%

had

bee

n re

ferr

ed to

a C

hild

Gui

danc

e C

linic

com

pare

d to

12%

of a

dopt

ed c

hild

ren.

Hod

ges

and

Tiz

ard

(198

9a)

n=

42

Mal

e-fe

mal

e 2:

1A

ge 1

6 ye

ars

IT 2

4-48

mon

ths

Sam

e as

abo

ve.

The

soc

ial r

elat

ions

hips

of e

x-in

stitu

tiona

lad

oles

cent

s ra

ised

unt

il at

leas

t the

age

of

2 in

a U

K r

esid

entia

l nur

sery

was

mea

-su

red

at 1

6 ye

ars.

The

adj

ustm

ent o

f the

child

ren

who

had

sub

sequ

ently

bee

nad

opte

d an

d re

stor

ed w

as c

ompa

red

and

also

with

a c

ontr

ol g

roup

of a

dole

scen

ts.

Bot

h gr

oups

of e

x-in

stitu

tiona

l chi

ldre

n ha

dm

ore

diffi

culti

es w

ith p

eers

and

had

few

ercl

ose

rela

tions

hips

with

pee

rs th

an th

eco

ntro

l gro

up.

Hod

ges

and

Tiz

ard

(198

9b)

n=

42

Mal

e-fe

mal

e 2:

1A

ge 1

6 ye

ars

IT 2

4-48

mon

ths

Sam

e as

abo

ve.

The

beh

avio

ral a

djus

tmen

t of e

x-in

stitu

tiona

l ado

lesc

ents

rai

sed

until

at

leas

t the

age

of 2

in a

UK

res

iden

tial

nurs

ery

was

mea

sure

d at

16

year

s.T

head

just

men

t of t

he c

hild

ren

who

had

sub

se-

quen

tly b

een

adop

ted

and

rest

ored

was

com

pare

d an

d al

so w

ith a

con

trol

gro

up o

fad

oles

cent

s.

At 1

6 ye

ars,

the

ex-in

stitu

tiona

l ado

lesc

ents

still

sho

wed

pro

blem

s at

sch

ool a

ccor

ding

to te

ache

r ra

tings

.The

se c

hild

ren

tend

edto

be

mor

e re

stle

ss a

nd d

istr

actib

le, q

uar-

rels

ome

with

pee

rs, a

nd r

esen

tful i

f cor

-re

cted

by

adul

ts c

ompa

red

to c

ontr

ols.

Ado

pted

chi

ldre

n ha

d be

gun

to d

ispl

aysi

gns

of a

nxie

ty.R

esto

red

child

ren

tend

edto

be

mor

e an

tisoc

ial o

r ap

athe

tic.O

vera

ll,th

e re

stor

ed c

hild

ren

show

ed m

ore

prob

-le

ms

than

the

adop

ted

child

ren,

and

pro

b-le

ms

obse

rved

in th

is g

roup

at a

ge 8

had

not i

mpr

oved

.

TAB

LE

2(c

on

tinu

ed)

Stu

dyIn

stitu

tion

Sam

ple

Inst

itutio

nal C

hara

cter

istic

sD

escr

iptio

n of

Stu

dyF

indi

ngs

45

Kal

er a

nd F

reem

an(1

994)

n=

25

Mal

e-fe

mal

e 1.

5:1

Age

23-

50 m

onth

sIT

1-4

7 m

onth

s

Des

crib

ed a

s re

pres

enta

tive

exam

ple

of a

Rom

ania

nor

phan

age

The

soc

ial d

evel

opm

enta

l sta

tus

of R

oma -

nian

orp

hans

was

com

pare

d w

ith a

gro

upof

Rom

ania

n ki

nder

gart

en c

hild

ren.

The

chi

ldre

n fr

om th

e or

phan

age

wer

e de

-la

yed

on a

ll m

easu

res:

adap

tive

beha

vior

,so

cial

com

mun

icat

ion,

vis

ual s

elf-

reco

gni -

tion,

soc

ial r

efer

enci

ng, l

evel

of p

lay,

and

leve

l of i

nter

actio

n.T

he g

reat

est s

tren

gth

in th

e or

phan

age

child

ren

was

in th

eir

peer

soci

al in

tera

ctio

n bu

t the

ir be

havi

or s

ug-

gest

ed th

ey w

ere

rela

ting

to o

ne a

noth

erin

disc

rimin

atel

y.

Slo

utsk

y (1

997)

n=

52

Mal

e-fe

mal

e 1:

1A

ge 7

0-88

mon

ths

IT 2

7-70

mon

ths

Sta

ff to

chi

ld r

atio

1:8

-10.

Bas

icbu

t ade

quat

e fa

cilit

ies

(toy

s,bo

oks)

.No

pers

onal

pos

ses -

sion

s.P

erso

nal r

elat

ions

hips

dis

-co

urag

ed.G

ood

stan

dard

of

phys

ical

car

e.

Em

path

y an

d co

nfor

mity

wer

e m

easu

red

inch

ildre

n pl

aced

in a

Rus

sian

orp

hana

ge.

The

se c

hild

ren

wer

e co

mpa

red

with

agr

oup

of R

ussi

an k

inde

rgar

ten

child

ren.

The

chi

ldre

n ra

ised

in th

e or

phan

age

had

alo

wer

leve

l of e

mpa

thy

but a

hig

her

leve

l of

conf

orm

ity th

an th

e ki

nder

gart

en c

hild

ren.

Thi

s ef

fect

was

gre

ater

the

long

er th

e ch

il -dr

en h

ad b

een

in th

e in

stitu

tion.

Fis

her

et a

l.(1

997

n=

46

Mal

e-fe

mal

e 1:

1.3

Age

18-

76 m

onth

sIT

8-5

3 m

onth

s

Sev

erel

y de

priv

ed e

nviro

nmen

t.N

utrit

iona

l and

psy

chol

ogic

al p

ri-

vatio

n.H

arsh

phy

sica

lco

nditi

ons.

Beh

avio

r pr

oble

ms

wer

e m

easu

red

in a

sam

ple

of R

oman

ian

orph

ans

(RO

) w

hoha

d su

bseq

uent

ly b

een

adop

ted

in C

an-

ada.

The

se c

hild

ren

wer

e co

mpa

red

with

am

atch

ed g

roup

of C

anad

ian-

born

(C

B)

child

ren

(non

-ado

pted

) an

d a

mat

ched

grou

p of

Rom

ania

n ch

ildre

n (R

C)

who

had

been

ado

pted

in C

anad

a bu

t had

nev

erbe

en in

stitu

tiona

lized

.

The

RO

chi

ldre

n ha

d hi

gher

tota

l pro

blem

scor

es th

an th

e C

B a

nd R

C c

ompa

rison

grou

ps.T

he R

O c

hild

ren

also

had

sig

nifi-

cant

ly h

ighe

r “in

tern

aliz

ing”

scor

es (

e.g.

,de

pres

sion

, soc

ial w

ithdr

awal

) bu

t not

“ext

erna

lizin

g”sc

ores

(e.

g., a

ggre

ssio

n,hy

pera

ctiv

ity)

than

the

othe

r tw

o gr

oups

.65

% o

f RO

had

an

eatin

g pr

oble

m (

over

-ea

ting,

pro

blem

with

sol

id fo

ods)

, 44%

had

a sl

eepi

ng p

robl

em (

did

not s

igna

l wak

e-up

), 8

4% d

ispl

ayed

ste

reot

yped

beh

avio

rs.

The

se p

robl

ems

wer

e no

t typ

ical

in th

e C

Ban

d R

C g

roup

s.M

ore

sibl

ing

prob

lem

sw

ere

repo

rted

in th

e R

O c

hild

ren

and

the

RO

and

RC

chi

ldre

n ha

d m

ore

peer

pro

b-le

ms

than

the

CB

gro

up.I

mpr

ovem

ents

wer

e ob

serv

ed fo

r ea

ting

prob

lem

s an

dst

ereo

type

d be

havi

ors,

the

leas

t im

prov

e-m

ents

wer

e ob

serv

ed fo

r si

blin

g an

d pe

erpr

oble

ms.

Mar

covi

tch

et a

l.(1

997)

n=

56

Mal

e-fe

mal

e 1:

1.1

Age

3-5

yea

rsIT

< 6

-mon

ths,

6-4

8 m

onth

s

Sam

e as

abo

ve.

Beh

avio

r pr

oble

ms

wer

e m

easu

red

in a

sam

ple

of R

oman

ian

orph

ans

who

had

subs

eque

ntly

bee

n ad

opte

d in

Can

ada.

The

out

com

e fo

r ch

ildre

n w

ho h

ad s

pent

less

than

6 m

onth

s in

an

orph

anag

e w

asco

mpa

red

with

that

of c

hild

ren

who

had

spen

t lon

ger

than

6 m

onth

s in

an

inst

itutio

n.

Bot

h gr

oups

of c

hild

ren

scor

ed in

the

nor-

mal

ran

ge o

n m

easu

re o

f beh

avio

r pr

ob-

lem

s bu

t chi

ldre

n w

ho h

ad s

pent

long

erth

an 6

mon

ths

in th

e or

phan

age

cons

is-

tent

ly s

core

d hi

gher

than

chi

ldre

n in

inst

itu-

tiona

l car

e fo

r le

ss th

an 6

mon

ths.

(con

tinu

ed)

46

Vor

ria, R

utte

r, P

ickl

es,

Wol

kind

, and

Hob

sbau

m (

1998

)

n=

41

Mal

e-fe

mal

e 1:

1A

ge 9

-11

year

sIT

2-7

yea

rs

Sta

bilit

y of

car

e-gi

ving

sta

ff bu

tlo

w c

areg

iver

-chi

ld r

atio

.Car

e-gi

ving

“no

n-pe

rson

aliz

ed.”

Goo

dst

anda

rd o

f phy

sica

l car

e.

The

soc

ial a

nd b

ehav

iora

l adj

ustm

ent o

fG

reek

chi

ldre

n in

long

-ter

m r

esid

entia

lca

re w

as in

vest

igat

ed.A

lthou

gh th

e ch

il-dr

en w

ere

in lo

ng-t

erm

car

e, m

ost h

adsp

ent t

he fi

rst 2

yea

rs o

f life

with

thei

r fa

m-

ily.T

he o

utco

me

of th

ese

child

ren

was

com

pare

d w

ith a

mat

ched

con

trol

gro

up o

fG

reek

chi

ldre

n ra

ised

in tw

o-pa

rent

fam

ilies

.

The

res

iden

tial c

are

grou

p w

as m

ore

inat

-te

ntiv

e, le

ss p

artic

ipat

ory

and

mor

edi

stra

ctib

le a

t sch

ool t

han

the

cont

rol

grou

p.O

n pa

rent

and

teac

her

ratin

gs, t

hein

stitu

tiona

l chi

ldre

n sh

owed

mor

e ov

eral

ldi

stur

banc

e, h

ad le

ss h

arm

onio

us p

eer

re-

latio

ns a

nd w

ere

mor

e at

tent

ion-

seek

ing

with

teac

hers

.Boy

s sh

owed

poo

r ta

sk in

-vo

lvem

ent,

mor

e em

otio

nal d

iffic

ultie

s,co

nduc

t pro

blem

s, a

nd h

yper

activ

ity th

anco

ntro

ls.G

irls

show

ed p

oor

task

invo

lve-

men

t and

mor

e em

otio

nal d

iffic

ultie

s th

anco

ntra

sts.

Kre

ppne

r et

al.

(199

9)n

= 1

04M

ale-

fem

ale

2.3:

1A

ge 4

yea

rsIT

< 6

mon

ths,

6-<

24m

onth

s

Sev

erel

y de

priv

ed e

nviro

nmen

t.N

utrit

iona

l and

psy

chol

ogic

al p

ri-

vatio

n.H

arsh

phy

sica

lco

nditi

ons.

The

pre

tend

and

soc

ial r

ole

play

of a

sam

-pl

e of

Rom

ania

n or

phan

s (R

O)

who

had

subs

eque

ntly

bee

n ad

opte

d in

the

Uni

ted

Kin

gdom

was

inve

stig

ated

.A g

roup

of U

Kad

opte

es w

ere

also

obs

erve

d as

a c

om-

paris

on g

roup

.

The

UK

ado

ptee

s w

ere

muc

h m

ore

likel

y to

enga

ge in

inte

ract

ive

role

-pla

y, p

rete

ndpl

ay, r

efer

to o

ther

s’m

enta

l sta

tes

and

show

mor

e sh

ared

enj

oym

ent t

han

the

RO

.T

he d

iffer

ence

s co

uld

not b

e ex

plai

ned

byco

gniti

ve d

evel

opm

ent a

nd v

erba

l abi

lity.

The

re w

ere

no d

iffer

ence

s be

twee

n ea

rlier

or la

ter

plac

ed R

O, t

houg

h th

ere

was

atr

end

for

the

late

pla

ced

child

ren

to e

ngag

ein

less

pre

tend

and

rol

e-pl

ay.

Rut

ter

et a

l.(1

999)

n=

111

Mal

e-fe

mal

e 2.

3:1

Age

6 y

ears

IT 6

-24

mon

ths

Sev

erel

y de

priv

ed e

nviro

nmen

t.N

utrit

iona

l and

psy

chol

ogic

alpr

ivat

ion.

Poo

r ph

ysic

alco

nditi

ons.

The

pre

senc

e of

beh

avio

rs a

ssoc

iate

d w

ithau

tism

wer

e in

vest

igat

ed a

sam

ple

of R

o-m

ania

n or

phan

s (R

O)

who

had

sub

se-

quen

tly b

een

adop

ted

in th

e U

nite

d K

ing

-do

m.A

gro

up o

f UK

ado

ptee

s w

ere

also

obse

rved

as

a co

mpa

rison

gro

up.

“Aut

istic

”beh

avio

rs w

ere

obse

rved

in 1

2%of

the

RO

chi

ldre

n at

age

4;h

owev

er, t

heim

prov

emen

t see

n at

age

6, t

he e

qual

-sex

ratio

, and

the

norm

al h

ead

circ

umfe

renc

esu

gges

ted

that

thes

e ca

ses

diffe

red

to “

or-

dina

ry”a

utis

m.

TAB

LE

2(c

on

tinu

ed)

Stu

dyIn

stitu

tion

Sam

ple

Inst

itutio

nal C

hara

cter

istic

sD

escr

iptio

n of

Stu

dyF

indi

ngs

47

Bec

kett

et a

l.(2

002)

n=

144

Mal

e-fe

mal

e 1:

1.2

Age

6 y

ears

IT <

6 m

onth

s, 6

-<12

mon

ths,

12-

<24

mon

ths,

24-4

3 m

onth

s

Sam

e as

abo

ve.

A n

umbe

r of

beh

avio

r pa

ttern

s w

ere

inve

sti -

gate

d in

a s

ampl

e of

Rom

ania

n or

phan

sw

ho h

ad s

ubse

quen

tly b

een

adop

ted

inth

e U

nite

d K

ingd

om.D

urat

ion

of d

epriv

a -tio

n w

as c

ompa

red

with

out

com

e.T

he b

e -ha

vior

s m

easu

red

incl

uded

roc

king

, sel

f-in

-ju

ry, u

nusu

al s

enso

ry in

tere

sts,

and

eat

ing

prob

lem

s (d

iffic

ulty

with

sol

id fo

ods)

.

At a

dopt

ion,

47%

eng

aged

in r

ocki

ng b

e -ha

vior

, 18%

stil

l did

this

at a

ge 6

.At e

ntry

into

the

Uni

ted

Kin

gdom

, 24%

sel

f-in

jure

d;13

% s

till s

elf-

inju

red

at a

ge 6

.Sel

f-in

jury

was

ofte

n a

resp

onse

to b

eing

told

off.

All

the

abov

e w

ere

mor

e lik

ely

in c

hild

ren

who

had

been

in in

stitu

tions

for

a lo

nger

per

iod

of ti

me.

At p

lace

men

t, 11

% h

ad u

nusu

alse

nsor

y in

tere

sts,

and

som

e ch

ildre

n be

-ga

n di

spla

y af

ter

adop

tion

(too

imm

atur

e at

entr

y).P

robl

ems

with

che

win

g an

d sw

al-

low

ing

solid

food

s w

ere

mor

e lik

ely

in c

hil -

dren

who

had

rem

aine

d in

inst

itutio

nal

care

for

a ye

ar o

r lo

nger

.

Har

den

(200

2)n

= 3

5M

ale-

fem

ale

1.5:

1A

ge 9

-30

mon

ths

IT 9

-27

mon

ths

Sta

ff to

chi

ld r

atio

1:2

(bu

t diff

er-

ent w

eeke

nd s

taff)

.Wel

lpr

ovis

ione

d.C

areg

iver

-chi

ld in

-te

ract

ion

enco

urag

ed.G

ood

stan

dard

of p

hysi

cal c

are.

Ada

ptiv

e be

havi

or a

nd b

ehav

ior

prob

lem

sw

ere

mea

sure

d in

a s

ampl

e of

infa

nts

and

todd

lers

in U

.S.c

ongr

egat

e ca

re s

ettin

gs.

The

se c

hild

ren

wer

e co

mpa

red

with

agr

oup

of U

.S.c

hild

ren

fost

ered

in fa

mili

es.

The

chi

ldre

n ra

ised

in c

ongr

egat

e ca

refa

red

wor

se th

at th

e ch

ildre

n fo

ster

ed in

fam

ilies

on

mea

sure

s of

com

mun

icat

ion

and

soci

aliz

atio

n.T

here

wer

e no

diff

er-

ence

s in

rep

orte

d an

d ob

serv

ed b

ehav

ior

prob

lem

s.

NO

TE

:nre

fers

to th

e in

stitu

tiona

l sam

ple

subj

ects

onl

y;th

e st

udie

s ha

d a

sim

ilar

num

ber

of c

ompa

rison

or

cont

rol s

ubje

cts.

IT =

ran

ge o

f tim

e sp

ent i

n in

stitu

tiona

l car

e.

has highlighted a number of social and behav-ioral problems that are more prevalent in thatgroup compared to other children (see Table 2).In particular, research has highlighted problemswith behavior, social competence, play, andpeer and/or sibling interactions. Researchershave also reported “quasi-autistic” behaviors insome severely deprived children (Rutter et al.,1999). Of the 17 studies summarized in Table 2,16 reported some negative social or behavioralconsequences for children raised in institu-tional care compared to controls or childrenwho had spent less time in institutional care.However, the severity and duration of difficul-ties varied greatly across these studies, reflect-ing the different and changing situations andexperiences of the children studied.

Tizard and Rees (1975) described the prob-lems reported by their control sample of “Lon-don mothers” as disciplinary issues (e.g., dis-obedience and not settling down when put tobed). For children raised in institutions, how-ever, the problems were of a different nature.The institutional staff they interviewed re-ported few disciplinary problems with the 4-year-old children in their care. The most fre-quent problems reported by the institutionalstaff included poor peer relations, tempertantrums, clinging, and poor concentration.

Tizard and Hodges (1978) described the be-havioral and emotional development of thesechildren again at age 8. By this age, the majorityof the institutional children had been restored totheir natural parents or had been adopted, only8 of the 65 children described in earlier studies(Tizard & Rees, 1975) remained in institutionalcare. They concluded that behavioral and emo-tional problems were very much a function ofthe environment that the child had been placedin after institutional care. The adopted childrenwere faring much better at age 8 than the re-stored children were. Tizard and Hodges (1978)identified a number of differences between theadoptive and the natural parents. The adoptiveparents had very much wanted a child, whereasthe natural parents were often ambivalent or re-luctant to take their children back from the insti-tutional care they had placed them in. The re-stored children also tended to have more

siblings and particularly younger siblings whothe mother often expressed a preference for.Also, many of the restored children returned toa stepfather who was “indifferent or evenhostile to them, or showed an open preferencefor his own children.”

Although the adopted children in Tizard’ssample fared much better than the children whowere restored, at age 8 the ex-institutional chil-dren generally showed several differences to acomparison group of noninstitutionalized chil-dren. There were large and significant differ-ences between the ex-institutional children andthe comparison group on the teachers’ ratingsof problem behaviors. On total problem scoreand antisocial items, the ex-institutional chil-dren scored much higher and were more oftendescribed by their teachers as attention seekingthan the comparison group.

Hodges and Tizard (1989a, 1989b) also re-ported on the outcome of their sample at age 16.In adolescence, the ex-institutional groups, whohad spent at least the first 2 years of their life inresidential care, had more behavioral and emo-tional problems than their matched compari-sons. According to their teachers, between 35%and 50% of the ex-institutional children showedthe following difficulties to some degree: rest-less, distractible, quarrelsome with peers, irrita-ble, and resentful if corrected by adults. By age16, the adopted group were displaying moresigns of anxiety whereas the restored childrentended toward more antisocial types ofbehavior.

Fisher et al. (1997) investigated behaviorproblems in Romanian orphans aged 18 to 76months who had been adopted to Canada.Three groups of children were compared: a Ro-manian orphanage (RO) group who had spentat least 8 months in a Romanian orphanage, aCanadian-born (CB) group of nonadopted chil-dren matched to the RO group for sex and age,and a Romanian comparison (RC) group whowere adopted to Canada before the age of 4months but had not been placed in an orphan-age. On the CBCL, the RO had higher totalscores and higher internalizing scores (e.g., de-pression, social withdrawal) than the CB andRC matches. Parental reports supported this

48 TRAUMA, VIOLENCE, & ABUSE / January 2006

finding; RO children were described by theirparents as withdrawing from and avoiding sib-ling interaction. There were no differences be-tween the groups for externalizing (e.g., aggres-sion, hyperactivity) scores. Using parentalreports of problems, the RO children reporteddistinctly different types of problem than the CBparents. The RO children had more feedingproblems than the CB children. The problemscited by the RO parents were excessive eatingand dislike of solid foods, which were notreported by the CB parents.

The eating problems reported in the RO chil-dren reflect the conditions in the orphanageswhere the children were malnourished andgiven all of their food in a bottle up to the age of2 years. The number of sleep problems was thesame for the CB and the RO groups but againthe type of problems were different for the twogroups. The RO children did not signal waking,though this was not described as a problem bythe parents. The RO children also slept exces-sively though this may have been a misinterpre-tation of the fact that they did not indicate whenthey had woken. Again this reflects the orphan-age experience where lying quietly in bed wasthe most common activity.

The parents of the RO children also reporteda high prevalence of stereotyped behaviors(84%). These stereotypies have frequently beenobserved in institutionalized children problems(e.g., Beckett et al., 2002), however, they havealso been observed to a lesser extent innoninstitutionalized samples of children(Smyke et al., 2002). Stereotyped behaviors in-clude body rocking, hand rocking, and rhythmi-cal head shaking (Thelen, 1979). These behav-iors are thought to be precursors to movementthat have not been allowed to develop further inthe confines of a crib. The behaviors may serveas a means of self-stimulation in an unrespon-sive environment or as a means to soothe intimes of distress. The stereotyped behaviorproblems of the RO children showed the mostimprovement or complete resolution after timein an adoptive home.

Rutter et al. (1999) and Beckett et al. (2002) de-scribe a set of autistic-like patterns of behaviorobserved in their sample of Romanian

adoptees. These quasi-autistic patterns in-cluded stereotyped behaviors, repetitivebehaviors, a lack of boundaries, difficultiesforming selective relationships. Only a smallsubsample of the adoptees displayed these be-haviors; 6% showed autistic patterns and a fur-ther 6% showed milder (usually isolated) fea-tures of autism (Rutter et al., 1999). Thesepatterns of behavior were more likely amongchildren who had spent longer in institutionalcare.

Although the clinical features observed inthese children were similar to “ordinary” au-tism, there was an equal sex ratio, a degree of so-cial interest, and there was a great improvementseen in these children between the ages of 4 and6 in these Romanian “autistics.” Rutter et al.(1999) conclude that this quasi-autistic patternof behavior is associated with prolonged experi-ential and perceptual deprivation, cognitive im-pairment, and a lack of opportunity to developclose attachment. However, these behaviorswere only observed in a minority of Romanianadoptees and the etiology of these symptoms isunclear.

Institutional care in early life predisposeschildren to behavioral and social problems laterin life. Many of the problems observed in sam-ples of severely deprived children, such as ste-reotyped behaviors andeating problems, showrapid improvement oncethe child is removed frominstitutional care andplaced in a supportivefamily environment.However, placementwith a family is notenough by itself to over-come difficulties; pooroutcome of some childrenrestored to their naturalfamily (Hodges & Tizard,1989a) shows that thequality of the subsequentfamily environment is animportant factor in the outcome of institution-ally reared children. Whilst subsequent place-ment in a supportive family can result in the for-

Johnson et al. / YOUNG CHILDREN IN INSTITUTIONAL CARE 49

(text continues on page 53)

Placement with afamily is not enough

by itself to overcomedifficulties; poor

outcome of somechildren restored totheir natural family

shows that the qualityof the subsequent

family environment isan important factor in

the outcome ofinstitutionally reared

children.

50

TAB

LE

3:

Co

gn

itiv

e D

evel

op

men

t o

f C

hild

ren

Rai

sed

in In

stit

uti

on

al C

are

Stu

dyIn

stitu

tion

Sam

ple

Inst

itutio

nal C

hara

cter

istic

sD

escr

iptio

n of

Stu

dyF

indi

ngs

Gol

dfar

b (1

944)

n=

15

Mal

e-fe

mal

e 1.

1:1

Age

10-

14 y

ears

IT 2

7-47

mon

ths

Not

des

crib

ed (

see

Gol

dfar

b,19

45)

Inte

llige

nce

and

spee

ch d

evel

opm

ent w

asm

easu

red

in a

dole

scen

ts w

ho h

ad s

pent

thei

r ea

rly in

fanc

y in

inst

itutio

nal c

are

but

who

had

sub

sequ

ently

bee

n fo

ster

ed.

The

se c

hild

ren

wer

e co

mpa

red

with

am

atch

ed c

ompa

rison

gro

up w

ho w

ere

info

ster

car

e an

d ha

d be

en in

fam

ily-b

ased

care

sin

ce b

irth

.

The

chi

ldre

n w

ho h

ad e

xper

ienc

ed in

stitu

tiona

lca

re in

ear

ly in

fanc

y w

ere

infe

rior

to th

e “f

os-

ter”

grou

p on

a n

umbe

r of

cog

nitiv

e m

easu

res;

100%

of t

he “

inst

itutio

n”gr

oup

com

pare

d w

ith40

% o

f the

“fo

ster

”gro

up h

ad b

elow

ave

rage

IQ.A

lso,

the

spee

ch o

f the

inst

itutio

n gr

oup

was

infe

rior

to th

at o

f the

fost

er g

roup

.

Gol

dfar

b (1

945)

n=

15

Mal

e-fe

mal

e 1.

5:1

Age

43

mon

ths

IT 4

-32

mon

ths

Not

des

crib

ed in

det

ail “

adul

t-ch

ild r

atio

is v

ery

low

...

a m

ini -

mum

of a

dult

stim

ulat

ion,

”“th

ech

ild’s

act

iviti

es a

re c

ompl

etel

yre

gula

ted.

”

The

inte

llige

nce

and

lang

uage

dev

elop

-m

ent o

f chi

ldre

n ra

ised

in in

stitu

tiona

lca

re in

the

Uni

ted

Kin

gdom

was

test

ed a

t3

year

s.T

he c

hild

ren

wer

e th

en p

lace

d in

fost

er h

omes

and

a fo

llow

-up

test

car

ried

out 9

mon

ths

afte

r th

e fir

st te

st.T

hese

child

ren

wer

e co

mpa

red

with

a m

atch

edco

mpa

rison

gro

up w

ho w

ere

in fo

ster

care

and

had

bee

n in

fam

ily-b

ased

car

esi

nce

birt

h.

On

both

mea

sure

s of

IQ, t

he fo

ster

-car

e ch

il -dr

en s

core

d hi

gher

than

the

inst

itutio

nally

rear

ed c

hild

ren.

Thi

s w

as a

lso

the

case

at t

hefo

llow

-up

visi

t whe

n al

l the

chi

ldre

n w

ere

info

ster

car

e.A

t bot

h te

stin

g tim

es, t

he fo

ster

-ca

re c

hild

ren

also

had

sup

erio

r la

ngua

gesk

ills;

the

seco

nd te

st s

core

s of

the

inst

itutio

nch

ildre

n w

ere

still

low

er th

an th

e fir

st te

stsc

ores

of t

he fo

ster

gro

up.

Prin

gle

and

Tann

er(1

958)

n=

18

Mal

e-fe

mal

e 1:

0.8

Age

4 a

nd a

hal

f yea

rsIT

6-4

8 m

onth

s

Not

des

crib

ed.R

esid

entia

l nur

s-er

y in

Uni

ted

Kin

gdom

.Goo

dst

anda

rd o

f phy

sica

l car

e.

Lang

uage

dev

elop

men

t in

youn

g ch

ildre

nra

ised

in U

K r

esid

entia

l nur

serie

s w

asco

mpa

red

with

that

of a

gro

up o

fm

atch

ed c

ontr

ols.

For

mal

asp

ects

of

spee

ch, v

ocab

ular

y, a

nd s

ente

nce

stru

c-tu

re a

nd c

hild

ren’

s ab

ility

to u

nder

stan

dan

d ex

pres

s th

emse

lves

wer

ein

vest

igat

ed.

Res

iden

tial n

urse

ry c

hild

ren

wer

e re

tard

ed in

form

al a

spec

ts o

f lan

guag

e an

d ha

d a

poor

ervo

cabu

lary

than

the

cont

rol g

roup

(th

ey c

ould

not n

ame

the

item

s of

per

sona

l pos

sess

ions

).S

peec

h de

velo

pmen

t was

nor

mal

.

Rhe

ingo

ld a

nd B

ayle

y(1

959)

n=

14

Mal

e-fe

mal

e 1:

1A

ge 1

7-22

mon

ths

IT 4

-18

mon

ths

Not

des

crib

ed.

Two

grou

ps o

f chi

ldre

n in

inst

itutio

nal c

are

wer

e co

mpa

red

in a

n ex

perim

enta

l situ

a-tio

n.H

alf t

he c

hild

ren

rece

ived

2 m

onth

sof

car

e fr

om a

sin

gle

care

give

r.T

he c

on-

trol

gro

up w

as c

ompl

etel

y re

ared

und

erin

stitu

tiona

l rou

tine.

The

IQ o

f the

chi

l-dr

en w

as m

easu

red

18 m

onth

s la

ter.

The

IQ o

f chi

ldre

n su

bseq

uent

ly a

dopt

ed o

rre

stor

ed to

thei

r fa

mily

was

als

oco

mpa

red.

The

exp

erim

enta

l gro

up d

id n

ot fa

re a

ny b

ette

rth

an th

e gr

oup

rais

ed u

nder

inst

itutio

nal c

are

alon

e.T

he c

hild

ren

who

wer

e su

bseq

uent

lyad

opte

d ha

d a

high

er IQ

and

a la

rger

voc

abu-

lary

than

thos

e ch

ildre

n w

ho w

ere

rest

ored

toth

eir

natu

ral f

amily

, but

thes

e di

ffere

nces

wer

eno

t sig

nific

ant.

51

Tiz

ard

and

Jose

ph(1

970)

n=

30

Mal

e-fe

mal

e 1:

1A

ge 2

yea

rsIT

4-2

4 m

onth

s

Sta

ff to

chi

ld r

atio

1:3

but

hig

hst

aff t

urno

ver.

Boo

ks, t

oys,

pla

yfa

cilit

ies.

Mix

ed a

ge g

roup

s.H

ome

visi

ts a

nd o

utin

gs.P

er-

sona

l rel

atio

nshi

ps d

isco

urag

ed.

Goo

d st

anda

rd o

f phy

sica

l car

e.

Cog

nitiv

e de

velo

pmen

t and

spo

ntan

eous

lang

uage

was

mea

sure

d in

a s

ampl

e of

child

ren

rais

ed in

a U

K r

esid

entia

l nur

s -er

y.A

sam

ple

of “

wor

king

-cla

ss”c

hild

ren

was

use

d as

a c

ontr

ol g

roup

.

The

men

tal a

ge o

f the

res

iden

tial n

urse

ry c

hil -

dren

was

2 m

onth

s be

hind

the

norm

.The

con

-tr

ol g

roup

voc

aliz

ed m

ore,

had

a la

rger

voc

ab-

ular

y, a

nd u

sed

long

er s

ente

nces

than

the

resi

dent

ial n

urse

ry g

roup

.

Tiz

ard

and

Ree

s(1

974)

n=

65

Mal

e-fe

mal

e 2.

3:1

Age

4 a

nd a

hal

f yea

rsIT

24-

48 m

onth

s

Sam

e as

abo

ve.

Cog

nitiv

e de

velo

pmen

t in

a sa

mpl

e of

chi

l -dr

en r

aise

d in

a U

K r

esid

entia

l nur

sery

was

mea

sure

d.C

hild

ren

wer

e ei

ther

stil

lin

nur

sery

, ado

pted

, or

rest

ored

to th

eir

natu

ral p

aren

ts.A

sam

ple

of “

wor

king

-cl

ass”

child

ren

was

use

d as

a c

ontr

olgr

oup.

No

evid

ence

of c

ogni

tive

reta

rdat

ion.

Chi

ldre

nad

opte

d ha

d hi

gher

IQ s

core

s th

an c

hild

ren

who

wer

e st

ill in

inst

itutio

nal c

are

or w

ho h

adbe

en r

esto

red

to th

eir

natu

ral f

amily

.

Tiz

ard

and

Hod

ges

(197

8)n

= 5

1M

ale-

fem

ale

1.8:

1A

ge 8

yea

rsIT

24-

48 m

onth

s

Sam

e as

abo

ve.

Cog

nitiv

e de

velo

pmen

t in

a sa

mpl

e of

chi

l -dr

en r

aise

d in

a U

K r

esid

entia

l nur

sery

was

mea

sure

d.S

ome

child

ren

wer

e st

illin

the

nurs

ery

but m

ost h

ad b

een

adop

ted

or r

esto

red

to th

eir

natu

ral p

ar-

ents

.A s

ampl

e of

“w

orki

ng-c

lass

”chi

ldre

nw

as u

sed

as a

con

trol

gro

up.

The

chi

ldre

n st

ill in

inst

itutio

nal c

are

at a

ge 8

and

thos

e w

ho h

ad b

een

rest

ored

to th

eir

nat -

ural

fam

ily h

ad a

vera

ge IQ

s.T

he c

hild

ren

adop

ted

befo

re th

e ag

e of

4 a

nd a

hal

f yea

rsha

d ab

ove-

aver

age

IQs.

IQ h

ad r

emai

ned

sta-

ble

for

each

gro

up s

ince

the

age

of 4

and

aha

lf.C

hild

ren

in in

stitu

tiona

l car

e ha

d th

e lo

w-

est I

Qs

but w

ere

still

in th

e no

rmal

ran

ge.

Hod

ges

and

Tiz

ard

(198

9b)

n=

42

Mal

e-fe

mal

e 2:

1A

ge 1

6 ye

ars

IT 2

4-48

mon

ths

Sam

e as

abo

ve.

The

cog

nitiv

e de

velo

pmen

t of e

x-in

stitu

-tio

nal a

dole

scen

ts r

aise

d un

til a

t lea

st th

eag

e of

2 in

a U

K r

esid

entia

l nur

sery

was

mea

sure

d at

16

year

s.T

he IQ

of t

he c

hil-

dren

who

had

sub

sequ

ently

bee

nad

opte

d an

d re

stor

ed to

thei

r fa

mily

was

com

pare

d an

d al

so w

ith a

con

trol

gro

upof

ado

lesc

ents

.

The

gro

up o

f chi

ldre

n ad

opte

d be

fore

the

age

of 4

and

a h

alf y

ears

stil

l had

the

high

est I

Q.

The

oth

er e

x-in

stitu

tiona

l gro

ups

did

not d

iffer

sign

ifica

ntly

.Non

e of

the

grou

ps h

ad a

mea

nIQ

of l

ess

than

94.

Kal

er a

nd F

reem

an(1

994)

n=

25

Mal

e-fe

mal

e 1.

5:1

Age

23-

50 m

onth

sIT

1-4

7 m

onth

s

Des

crib

ed a

s re

pres

enta

tive

ex-

ampl

e of

a R

oman

ian

orph

an-

age

(i.e.

, dep

rived

env

ironm

ent

with

nut

ritio

nal a

nd p

sych

olog

i-ca

l priv

atio

n an

d po

or p

hysi

cal

cond

ition

s)

The

cog

nitiv

e de

velo

pmen

t of R

oman

ian

orph

ans

was

com

pare

d w

ith a

gro

up o

fR

oman

ian

kind

erga

rten

chi

ldre

n.

Non

e of

the

child

ren

from

the

orph

anag

e w

ere

func

tioni

ng a

t age

leve

l;20

out

of 2

5 w

ere

func

tioni

ng a

t lev

els

less

than

hal

f the

ir ch

ro-

nolo

gica

l age

.The

con

trol

gro

up s

how

ed n

ode

velo

pmen

tal d

elay

.

Slo

utsk

y (1

997)

n=

52

Mal

e-fe

mal

e 1:

1A

ge 7

0-88

mon

ths

IT 2

7-70

mon

ths

Sta

ff to

chi

ld r

atio

1:8

-10.

Bas

icbu

t ade

quat

e fa

cilit

ies

(toy

s,bo

oks)

.No

pers

onal

pos

ses-

sion

s.P

erso

nal r

elat

ions

hips

disc

oura

ged.

Goo

d st

anda

rd o

fph

ysic

al c

are.

The

cog

nitiv

e de

velo

pmen

t of c

hild

ren

plac

ed in

a R

ussi

an o

rpha

nage

was

com

-pa

red

with

a g

roup

of R

ussi

an k

inde

rgar

-te

n ch

ildre

n.

The

chi

ldre

n ra

ised

in th

e or

phan

age

had

low

erIQ

sco

res

than

the

kind

erga

rten

chi

ldre

n.

(con

tinu

ed)

52

Rut

ter

et a

l.(1

998)

n=

111

Mal

e-fe

mal

e 2.

3:1

Age

4 y

ears

IT <

6 m

onth

s, 7

-24

mon

ths

Sev

erel

y de

priv

ed e

nviro

nmen

t.N

utrit

iona

l and

psy

chol

ogic

alpr

ivat

ion.

Poo

r ph

ysic

alco

nditi

ons.

The

cog

nitiv

e de

velo

pmen

t of t

wo

grou

psof

Rom

ania

n or

phan

s (R

O)

who

had

sub

-se

quen

tly b

een

adop

ted

to th

e U

nite

dK

ingd

om w

ere

com

pare

d w

ith a

sam

ple

of U

K a

dopt

ees.

Rom

ania

n ch

ildre

n w

hoha

d be

en a

dopt

ed b

efor

e th

e ag

e of

6m

onth

s w

ere

com

pare

d w

ith th

ose

adop

ted

afte

r th

is a

ge.

In c

ompa

rison

to U

K a

dopt

ees,

RO

cat

ch-u

p in

cogn

itive

leve

l was

com

plet

e at

age

4 fo

rth

ose

adop

ted

befo

re th

e ag

e of

6 m

onth

s.G

ood

prog

ress

but

not

com

plet

e re

cove

ry fo

rth

ose

adop

ted

afte

r 6

mon

ths.

O’C

onno

r et

al.

(200

0b)

n=

165

Mal

e-fe

mal

e 1:

1.2

Age

6 y

ears

IT <

6 m

onth

s, 7

-24

mon

ths,

25-4

2 m

onth

s

Sev

erel

y de

priv

ed e

nviro

nmen

t.N

utrit

iona

l and

psy

chol

ogic

alpr

ivat

ion.

Poo

r ph

ysic

alco

nditi

ons.

The

cog

nitiv

e de

velo

pmen

t of t

he R

oma -

nian

orp

hans

(R

O)

stud

ied

by R

utte

r et

al.

(199

8) w

as te

sted

aga

in a

t age

6 y

ears

.Afu

rthe

r gr

oup

of “

late

-pla

ced”

adop

tees

wer

e al

so in

clud

ed.T

hese

gro

ups

wer

eco

mpa

red

with

a s

ampl

e of

UK

ado

ptee

s.

Dur

atio

n of

priv

atio

n w

as th

e m

ost i

mpo

rtan

tpr

edic

tor

of c

ogni

tive

outc

ome.

In c

ompa

rison

to U

K a

dopt

ees,

RO

rec

over

y at

age

4 w

asm

aint

aine

d at

age

6 b

ut s

ome

defic

its a

t age

4fo

r th

ose

adop

ted

afte

r 6

mon

ths

wer

e pr

esen

tat

age

4 w

ere

also

pre

sent

at a

ge 6

yea

rs.

Tim

e sp

ent i

n ad

optiv

e ho

me

beyo

nd a

per

iod

of 2

yea

rs d

id n

ot im

prov

e de

velo

pmen

t

Har

den

(200

2)n

= 3

5M

ale-

fem

ale

1.5:

1A

ge 9

-30

mon

ths

IT 9

-27

mon

ths

Sta

ff to

chi

ld r

atio

1:2

(bu

t diff

er-

ent w

eeke

nd s

taff)

.Wel

l pro

-vi

sion

ed.C

areg

iver

-chi

ld in

ter-

actio

n en

cour

aged

.Goo

dst

anda

rd o

f phy

sica

l car

e.

The

men

tal d

evel

opm

ent o

f inf

ants

and

todd

lers

in U

.S.c

ongr

egat

e ca

re s

ettin

gsw

as c

ompa

red

with

a g

roup

of U

.S.c

hil-

dren

fost

ered

in fa

mili

es.

The

chi

ldre

n ra

ised

in c

ongr

egat

e ca

re fa

red

wor

se th

at th

e ch

ildre

n fo

ster

ed in

fam

ilies

inth

eir

men

tal d

evel

opm

ent.

NO

TE

:nre

fers

to th

e in

stitu

tiona

l sam

ple

subj

ects

onl

y;th

e st

udie

s ha

d a

sim

ilar

num

ber

of c

ompa

rison

or

cont

rol s

ubje

cts.

IT =

ran

ge o

f tim

e sp

ent i

n in

stitu

tiona

l car

e.

TAB

LE

3 (

con

tinu

ed)

Stu

dyIn

stitu

tion

Sam

ple

Inst

itutio

nal C

hara

cter

istic

sD

escr

iptio

n of

Stu

dyF

indi

ngs

mation of close attachments within that familyunit, many institutionally raised children willstill have problems interacting with peers andadults outside the family unit (Hodges &Tizard, 1989a).

CONSEQUENCES FOR COGNITIVEDEVELOPMENT

A summary of the 13 studies that have inves-tigated the cognitive development of childrenraised in institutional care is provided in Table 3.A total of 12 of the 13 studies illustrate a nega-tive effect of institutional care on cognitive de-velopment although some of these studies alsosuggest that early intervention, (i.e., removal tofamily-based care) can result in recovery.Rheingold and Bayley (1959) did not find anysignificant difference in IQ between childrenwho were raised in institutional care and chil-dren who were raised in institutional care butwho received 2 months of care from a singlecaregiver.

The early research into the cognitive develop-ment of children raised in institutions sug-gested that infants who were raised in institu-tions would be severely retarded, with specificdifficulties in language development and atten-tion, and that these difficulties would be perma-nent (e.g., Goldfarb, 1945). Subsequent researchby Barbara Tizard and her colleagues (Hodges& Tizard, 1989b; Tizard & Hodges, 1978; Tizard& Joseph, 1970, 1974), which followed a groupof children who were raised in institutions, gavea more optimistic prognosis for cognitive devel-opment. At 2 years of age, the nursery group(children who had been placed in institutionalcare before the age of 4 months) were 2 monthsbehind the contrast group (noninstitutionalizedbut from a working-class background) for men-tal age (Tizard & Joseph, 1970). The nurserygroup also had lower verbal competence scores,a smaller vocabulary, and made fewer wordcombinations than the contrast group did(Tizard & Joseph, 1970). But by 4 years of age,the children who were still in institutional caredid not show any signs of retardation and it wasthe children who had been restored to their bio-logical family who scored the poorest on mea-sures of intelligence (Tizard & Rees, 1974). At

age 8, the children still in institutional care hadaverage IQ scores (Tizard & Hodges, 1978) andby the time the children reached the age of 16,Hodges and Tizard (1989b) concluded that in-stitutional rearing does not have the “devastat-ing long-term effects described in some earlystudies.”

Although the longitudinal research by Tizardand her colleagues (Hodges & Tizard, 1989b;Tizard & Hodges, 1978; Tizard & Joseph, 1970,1974) suggest that institutional care does nothave a detrimental effect on cognitive develop-ment, there are some important points that needto be considered before conclusions can bedrawn. First, the institutional care that the chil-dren in Tizard’s study received was of a highstandard. The nursery environment for thesechildren was well equipped with plenty of toysand books, the children were read to daily andthe children were taken on outings and occa-sionally made weekend visits to the homes ofstaff members. The children lived in “familygroups” of six children, each group had its ownsuite of rooms and two assigned nurses. Al-though the children who remained in the insti-tution had average IQ scores, it was the childrenwho were adopted from the institution beforethe age of 4 and a half years old, who made thelargest gains in IQ, and these gains were main-tained over the subsequent 12 years (Hodges &Tizard, 1989b). Being adopted after this age didnot have the same effect; only one child out ofthe five adopted after the age of 4 and a half hadincreased in IQ by age 8.

Although the sample size from these studieswas small, the results from Tizard’s work(Hodges & Tizard, 1989b; Tizard & Hodges,1978; Tizard & Joseph, 1970, 1974) suggest thatchildren who are raised in small well-staffedand well-equipped institutions will not be se-verely cognitively delayed. However, unlesschildren are placed with families before the ageof 4, they will be at a cognitive disadvantagecompared with children who have spent theirearly years in a family setting.

Of course, one of the problems of trying toconsider the impact of institutional care on chil-dren is that standards and practices of institu-tional care vary enormously. Whilst the researchby Tizard demonstrates that retardation is not

Johnson et al. / YOUNG CHILDREN IN INSTITUTIONAL CARE 53

an inevitable consequence of institutional care,the conditions in the nurseries studied byTizard were of a high standard and not all insti-tutions can be described as such. This becameall too apparent when the fall of the Ceausescuregime in Romania brought the attention of theworld to more than 100,000 children who hadeffectively been “warehoused” typically with-out sufficient food, clothing, heat or caregivers(see Johnson, 2000).

Kaler and Freeman (1994) set out to describe arepresentative group of such children and con-ducted a number of tests on a group of 25 chil-dren. As with many studies that have sought toinvestigate the children from Romanian institu-tions, a lack of systematic records made it im-possible to rule out the influence of genetic fac-tors or the possibility that children had beenplaced into an institution because of a handicap.However, anecdotal material from the recordsthat were available suggested that the majorityof the children were not true orphans but the in-fants of adolescent mothers or the youngestchildren from a large family. Therefore, the pri-mary reasons for child placement were socio-economic factors. The children studied by Kalerand Freeman (1994) were aged between 23 to 50months and the mean length of time spent in theorphanage was approximately 26 months. Thecognitive development of the “orphanage”sample was compared with a group of childrenof a similar age who were attending a local kin-dergarten. The kindergarten group was func-tioning at chronological age level whereas 20children from the sample of 25 orphanagechildren were functioning at levels less thanhalf their chronological age.

The plight of children in Romanian orphan-ages attracted international media attentionand subsequently, many of these children were“rescued” and adopted internationally. Thisprovided a unique opportunity for researchersto study the effects of early deprivation and toinvestigate whether the effects of such depriva-tion in early life on cognitive development werereversible. Michael Rutter and his team at theInstitute of Psychiatry in London have followeda large sample (n = 111) of Romanian childrenwho were adopted into the United Kingdomfollowing severe early global privation. The

children in this sample had all been brought tothe United Kingdom before the age of 2 yearsand their level of cognitive functioning wasmeasured at age 4 (Rutter et al., 1998) and age 6(O’Connor et al., 2000b). On entry to the UnitedKingdom the children were severely develop-mentally impaired; the mean score for the groupon the Denver Scales was 63 (59% had a devel-opmental quotient below 50) and 51% werebelow the 3rd percentile in weight.

By the age of 4, the children had made sub-stantial physical and developmental catch-up;2% were under the 3rd percentile for weight andthe mean score on the Denver Scales rose to 107.The Romanian children who were adopted intothe United Kingdom before the age of 6 months(0-6 months) appeared to have made a completerecovery and were no different from compari-son samples of within–United Kingdomadoptees or Romanian children who had notbeen institutionalized. However, the catch-upin children who were adopted into the UnitedKingdom after the age of 6 months (7-24months), although still promising, suggestedthat the recovery in these children was not yetcomplete. Therefore, at age 4, there was a dose-response link between duration of deprivationand cognitive functioning.

At age 6, the Romanian adoptees were testedagain (O’Connor et al., 2000b). In addition to thesample of 111 children tested at age 4, a secondgroup of Romanian adoptees, “late-placedadoptees” (n = 48), who entered the UnitedKingdom between 24 and 42 months of agewere also tested at age 6. The late-placed groupallowed further testing of the dose-response hy-pothesis of deprivation and cognitive develop-ment and also, after more than 2 years of severedeprivation, this group provided a unique op-portunity to test resilience. All three groups ofadoptees (0-6 months, 7-24 months, and 25-42months) were equally delayed at entry into theUnited Kingdom).

At age 6, after between nearly 3 to 4 years inan adoptive home, the late-placed adoptees hadmade significant progress. On entry into theUnited Kingdom, more than 90% of the late-placed group had Denver scores below 70,where as at age 6, only 18% were below 70 onthis measure. Comparing the other groups on

54 TRAUMA, VIOLENCE, & ABUSE / January 2006

measures of cognitive ability at age 6, the earlyRomanian adoptees (0-6 months) did not differfrom the UK adoptee comparison group,though both of these groups scored signifi-cantly higher than the other samples (6-24months and 25-42 months). The strongest pre-dictor of cognitive ability at age 6 years was ageat entry into the United Kingdom and this wasalso the case when only the longitudinal sample(0-6 months, 7-24 months) was considered. Fur-ther analysis revealed that it was duration ofprivation rather than length of time in the adop-tive home (beyond a period of approximately 2years) that was the most important predictor ofcognitive level. In general, all of the adopteesmade remarkable progress in their cognitivefunctioning suggesting a resilience of develop-ment to early deprivation. However, the effectsof deprivation in the early months of life werestill apparent at age 6 and O’Connor et al.(2000b) conclude that the data provides strongevidence that early deprivation doescompromise long-term development.

One area that is typically reported as beingdelayed in children raised in institutional care islanguage development. Goldfarb (1944, 1945)investigated speech sounds, intelligibility ofspeech and language organization in early in-fancy, at 6 to 8 years of age and in adolescence.At all three age levels, institutionalized childrendisplayed a clear deficiency in language devel-opment as compared with a group of fosteredchildren. Numerous other researchers havesince reported deficits in the language skills ofchildren raised in institutions. These deficits in-clude poorer vocabulary and less spontaneouslanguage (Tizard & Joseph, 1970) and retarda-tion in formal aspects of language and languagedevelopment (Pringle & Tanner, 1958).

The degree of cognitive and language delayvaries depending on the standard of care pro-vided by different institutions, which explainsthe variations in findings across research stud-ies. Tizard and Joseph (1970) describe twoextreme types of child-care environment:institution-oriented facilities that result in de-layed development and child-oriented settingsthat promote normal development. Withinchild-oriented facilities, the staff do not adhereto a strict routine and tend to spend more time

interacting with and scaffolding the develop-ment of children. An institution-orientedapproach typically occurs under conditions ofscarce personnel resources and staff withinthese types of facilities are primarily concernedwith the physical care of the children.

Other features of institutional care that havealso been suggested as contributing to delayedlanguage development include poor provisionof books and play equipment, low staff-child ra-tios, staff experience, staff autonomy, children’slack of personal possessions, and children’s lackof “everyday” experience (Pringle & Tanner,1958; Tizard & Joseph, 1970).

Overall, the evidence suggests that institu-tional care is typically detrimental to the cogni-tive development of children. Severe depriva-tion, such as that encountered by children inRomanian orphanages after the collapse of theCeausescu regime, has a profound effect on cog-nitive development and complete recovery hasonly been observed, so far, in children who wereplaced in family-based care before the age of 6months. Children who were placed later madesignificant improvements in their developmentafter leaving institutional care but were still at acognitive disadvantage some years later(O’Connor et al., 2000b). However, not all chil-dren raised in institutional care display the se-verely delayed development observed in sam-ples of “Romanian orphans.” A child-orientedapproach, with adequate personnel and re-sources, can result in a similar cognitive out-come for both children who remain in less rigidresidential care routines and children who arerestored to their biological families who maystill be high risk for child abuse and neglect.However, children who are raised in a familysetting with parents who do have the personalresources to nurture them, have a better cogni-tive outcome than children in institutional careand the sooner a child is moved frominstitutional care to a family setting, the betterthe cognitive outcome will be.

DISCUSSION

The evidence for the detrimental effects of ex-posure to institutional care without a primarycaregiver on children is overwhelming when

Johnson et al. / YOUNG CHILDREN IN INSTITUTIONAL CARE 55

compared to the expo-sure of family-basedcare with a primarycaregiver. Of 12 stud-ies on attachment inchildren raised in in-stitutional care onlyone found no sup-porting evidence foran increase in attach-ment difficulties. Of17 studies on socialand behavioral devel-opment of children,again only one foundinconclusive evi-dence in relation toage of exposure toinstitutional care. Of13 studies on cogni-tive development, allexcept one report apoorer cognitive per-formance associatedwith insti tutionalcare.

When consideringthe consequence of in-stitutional care onemajor difficulty is thatthe standards of carethat children receive

in institutions varies enormously. To try and ad-dress this issue, two bodies of research havebeen the focus of this review. First, the researchwork that has described the development ofchildren who have been raised in “good” resi-dential care, and second, the research outliningthe development of children raised in ex-tremely poor institutional settings.

Comparing these two bodies of researchhighlights the fact that some of the detrimentaleffects of “institutionalization” are the result ofa lack of resources rather than institutional careper se. However, when “good” institutionalcare leads to a poor outcome this suggests thatthere are aspects of institutional culture, whichare fundamentally damaging to a developingchild. On a practical level, it is important to con-sider which aspects of institutional care are the

most damaging and how damage can be lim-ited, rather than simply write off institutionalcare completely. For many children institutionalcare is the only care available and attempts tode-institutionalize children without adequatesupport may be more damaging thaninstitutional care itself.

In the absence of foster care services and suit-able adopting parents, it is difficult to judgewhether children are better off in a residentialcare institution that may provide at least aphysically safe environment rather than reinte-grating them into potentially abusive and ne-glectful homes; for example, with parents whomay not have been adequately rehabilitatedfor mental health problems, substance abuse,and violent outbursts. Therefore, a comprehen-sive-assessment process is necessary before re-turning a child to their family. Nevertheless itmust be recognized that children in institutionalcare are not immune from maltreatment by theirpeers (e.g., bullying) and abuse and neglect bystaff. Indeed, physical and sexual abuse hasbeen reported as existing in a number of institu-tions worldwide (e.g., Barter, 2003; UNICEF,2002).

In a nonabusive environment with reason-able provision of resources, institutional carecan result in adequate cognitive developmentwithin the normal range. Nevertheless, childrenwho have been raised entirely in a supportivefamily will show higher IQs on average.

In terms of behavioral problems, the distinctset of behavioral problems seen in the severelydeprived Romanian orphans (e.g., difficultieswith solid foods, quasi-autistic behaviors) areextremely responsive to intervention and typi-cally disappear once the child is removed frominstitutional care and placed with a family.However, there do seem to be a number of prob-lems that persist in children who have spent theearly years in institutional care and these prob-lems seem to be due to a lack of close attach-ments during the early years. The Romanianadoptees studied by Fisher et al. (1997) showedthe least improvement in peer and sibling rela-tions. The children studied by Hodges andTizard (1989a, 1989b) also were still havingproblems with peers at age 16. Difficulties withpersonal relationships may be a lasting conse-

56 TRAUMA, VIOLENCE, & ABUSE / January 2006

The evidence for thedetrimental effects ofexposure to institutionalcare without a primarycaregiver on children isoverwhelming whencompared to theexposure of family-based care with aprimary caregiver. Of 12studies on attachment inchildren raised ininstitutional care onlyone found no supportingevidence for anincrease in attachmentdifficulties. Of 17 studieson social andbehavioraldevelopment ofchildren, again only onefound inconclusiveevidence in relation toage of exposure toinstitutional care. Of 13studies on cognitivedevelopment, all exceptone report a poorercognitive performanceassociated withinstitutional care.

quence of early institutional care and this wouldbe predicted from the neurobiological perspec-tive. The long-term consequences of institu-tional care on attachment have yet to be investi-gated fully but again the neurobiologicalperspective would suggest that these problemswill be ongoing for a number of children whospent their early years without a one-to-one sen-sitive caregiver. Many of these children will beemotionally vulnerable and their craving foradult attention and readiness to go off with astranger will make them obvious targets for sexoffenders (Elliott, Browne, & Kilcoyne, 1995).

In summary, the evidence clearly indicatesthat institutional care does not support the opti-mal development of children. Intervening earlywith children in institutional care and returning

them to a family is important for subsequent de-velopment. The international communityshould be encouraged to promote the humanrights of children in residential care and to pro-tect children from the abuse inherent in institu-tional care systems. According to the UnitedNations Convention on the Rights of the Child(United Nations, 1989), every child has the rightto grow up in a family and research suggeststhat children who are moved from residentialcare before the age of 6 months can still reachoptimal development (O’Connor et al., 2000b;Rutter et al., 1998). A longer period results insignificant risk of harm to physical andpsychological development and constitutesinstitutional maltreatment.

IMPLICATIONS FOR PRACTICE, POLICY, AND RESEARCH

Practice

Analytical epidemiological study designs(i.e., including a control and/or comparisongroup) show that young children placed in in-stitutional care without parents are at risk ofharm in terms of attachment disorder and de-velopmental delays in social, behavioral andcognitive domains. Delays in physical growth,neural atrophy and abnormal brain develop-ment have also been implicated in studies ofchildren in institutional care. The neglect anddamage caused by privation is equivalent toviolence to a young child.

Infants who are placed into residential carewill suffer harm to their development if notmoved to family-based care by the age of 6months.

Policy

Countries with young children in institu-tional care must develop alternative strategies

such as foster care and adoption. Rehabilitatingchildren to their families of origin is an optionbut this needs to be adequately supported andmonitored.

The international community should be en-couraged to promote the human rights of chil-dren in residential care and to support the de-velopment of family-based care alternatives.

Research

Research must identify good practices for thedeinstitutionalization of children in residentialcare that considers the needs of the child and re-duces the potential for trauma from change.

Alternative forms of family-based careshould be evaluated to identify the advantagesand disadvantages for the child, as well as fac-tors related to successful or unsuccessfulplacements.

REFERENCESAinsworth, M.D.S., Blehar, M., Waters, E., & Wall, S. (1978).

Patterns of attachment. Hillsdale, NJ: Lawrence Erlbaum.

Balbernie, R. (2001). Circuits and circumstances: Theneurobiological consequences of early relationshipexperiences and how they shape later behaviour. Jour-nal of Child Psychotherapy, 27, 237-255.

Johnson et al. / YOUNG CHILDREN IN INSTITUTIONAL CARE 57

Barter, C. (2003). Abuse of children in residential care: Informa-tion briefing, October 2003. London: National Society forthe Prevention of Cruelty to Children.

Beckett, C., Bredenkamp, D., Castle, J., Groothues, C.,O’Connor, T. G., Rutter, M., & The English and Roma-nian Adoptees Study Team. (2002). Behaviour patternsassociated with institutional deprivation: A study ofchildren adopted from Romania. Journal of Developmen-tal & Behavioral Pediatrics, 23(5), 297-303.

Bowlby, J. (1951). Maternal care and mental health. Geneva,Switzerland: World Health Organisation.

Bowlby, J. (1969). Attachment and loss, Vol. 1. Attachment.London: Hogarth.

Browne, K. D. (2002). Child protection. In M. Rutter & E.Taylor (Eds.), Child and adolescent psychiatry (4th ed.,chap. 70, pp. 1158-1174). Cambridge, MA: BlackwellScience.

Browne, K. D., Hamilton-Giacritsis, C. E., Johnson, R.,Agathonos, H., Anaut, M., Herczog, M., et al. (2004).Mapping the number and characteristics of children underthree in institutions across Europe at risk of harm (EuropeanUnion Daphne Programme, Final Project Report No.2002/017/C). Birmingham, UK: University Centre forForensic and Family Psychology.

Carlson, V., Cicchetti, D., Barnett, D., & Braunwald, K.(1989). Disorganized/disoriented attachment relation-ships in maltreated infants. Developmental Psychology,25, 525-531.

Cassidy, J., & Marvin, R. S., & Attachment Working Groupof the MacArthur Network on the Transition FromInfancy to Early Childhood. (1987, 1992). Attachmentorganization in three- and four-year-olds: Coding guidelines.Unpublished manual, Psychology Department, Uni-versity of Virginia, Charlottesville, VA. (Cited byMarcovitch et al., 1997)

Chisholm, K. (1998). A three year follow-up of attachmentand indiscriminate friendliness in children adoptedfrom Romanian orphanages. Child Development, 69(4),1092-1106.

Eliot, L. (2001). Early intelligence: How the brain and minddevelop in the first years. London: Penguin.

Elliot, M., Browne, K. D., & Kilcoyne, J. (1995). Child sexualabuse prevention: What offenders tell us. Child Abuse &Neglect: The International Journal, 19(5), 579-594.

Fisher, L., Ames, E. W., Chisholm, K., & Savoie, L. (1997).Problems reported by parents of Romanian orphansadopted to British Columbia. International Journal ofBehavioral Development, 20(1), 67-82.

Giese, S., & Dawes, A. (1999). Child care, developmentaldelay and institutio