Content uploaded by Robert J Ludwig
Author content
All content in this area was uploaded by Robert J Ludwig on Mar 02, 2019
Content may be subject to copyright.
Complementary Therapies in Clinical Practice (2006) 12,3–12
Outcomes of Prolonged Parent–Child Embrace
Therapy among 102 children with
behavioral disorders
Martha G. Welch
a,
, Robert S. Northrup
b
, Thomas B. Welch-Horan
c
,
Robert J. Ludwig
a
, Christine L. Austin
d
, Judith S. Jacobson
e
a
The Martha G. Welch Center, New York, NY, USA
b
Project Hope, Millwood, VA, USA
c
Department of Psychiatry, College of Physicians & Surgeons, Columbia University, New York, NY, USA
d
KC Mothering Center, Kansas City, MO, USA
e
Department of Epidemiology, Mailman School of Public Health, Columbia University, USA
KEYWORDS
Adoption;
Treatment;
Reactive attachment
disorder;
Oppositional defiant
disorder;
Attention deficit
hyperactive disorder
Summary A growing body of research in neuroscience points to the impact of
variations in maternal nurturing on child development and provides a rationale for
interventions that target stress adaptation conditioning through natural family
nurturing. This pilot study was collected within the course of private practice to
assess the progress of children with severe behavioral disorders who were treated
effectively with a multiple family therapy prototype, Prolonged Parent–Child
Embrace (PPCE) Therapy. Subjects were a consecutive series of 102 patient children
aged 4–18 years and their families. Children and their family members were guided
for 16 h over two consecutive days through intense PPCE Therapy. Families were
instructed to continue PPCE Therapy at home for at least 1 year. Scores were
compared statistically using t-tests and analysis of variance. For 96 children scores
declined on two written measures by approximately 50% between baseline and
follow-up (Po0.001). Results showed that PPCE Therapy resulted in significant and
prolonged improvements in symptomatic behavior in a majority of children.
&2005 Elsevier Ltd. All rights reserved.
Introduction
A growing body of research in neuroscience points
to the impact that variations in maternal nurturing
can have on development.
1–3
Absence of sufficient
maternal nurture adversely influences the mechan-
isms of development that regulate the expression
ARTICLE IN PRESS
www.elsevierhealth.com/journals/ctnm
1744-3881/$ - see front matter &2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ctcp.2005.09.004
Corresponding author. NYPI Unit 42, Columbia University,
1051 Riverside Dr., New York, NY 10021, USA.
Tel.: + 1 212 543 5101; fax: + 1 212 861 6816.
E-mail address: mgw13@columbia.edu (M.G. Welch).
Welch MG, Northrup RS, Welch-Horan TB, Ludwig RJ, Austin CL,
Jacobson JS. Outcomes of Prolonged Parent-Child Embrace
Therapy among 102 children with behavioral disorders.
Complement Ther Clin Pract. 2006 Feb;12(1):3-12.
of behavioral, endocrine, and autonomic responses
to stress.
4,5
Unless reversed, these early negative
influences can adversely impact the child’s stress
reactivity and behavior throughout life.
The literature supports the efficacy of maternal
nurturing actions that prevent or reverse behavior-
al and developmental disorders caused by nurtur-
ing deprivation. Hippocampal deficits in rat pups
deprived of maternal nurturing are reversed with
the reinstatement of nurturing interactions.
6
A
recently identified genetic program determines
stress reactivity and is differentially activated
environmentally by the amount of maternal nurtur-
ing a neonate receives.
7
Prolonged Parent–Child Embrace (PPCE) Therapy
is a family intervention that was developed within a
private practice.
8
PPCE combines aspects of family
systems theory and developmental psychopathol-
ogy.
9,10
The therapy focuses on the stress responses
and stress adaptation networks of the dysregulated
child, which we theorize can be positively condi-
tioned over time through specific kinds of family
interaction. For biological and physiological rea-
sons, the primary stress modulating relationship is
between the mother and child dyad, although the
intervention recognizes that individual behavior is
affected by and regulated within a complex net-
work of family relationships including grandpar-
ents, siblings and close relatives, and especially the
father.
11
Thus to the degree possible, the extended
family is treated simultaneously as a whole.
Previous use of the treatment with of a small
number of autistic children over a prolonged period
of time produced significant behavioral improve-
ment.
8,12
The current study used a brief 2-day
intensive intervention followed by family PPCE at
home for 1 year to treat a broader range of
behavioral pathology in a group of disorders that
are frequently resistant to treatment, including
oppositional defiant disorder (ODD), conduct dis-
order (CD), attention deficit hyperactivity disorder
(ADHD) and reactive attachment disorder (RAD).
13
Standardized written measures were used to test
the efficacy of PPCE Therapy. Parental oral assess-
ments were obtained for 1 year to further monitor
the progress of the children.
Methods
Study group
Subjects were an unselected, consecutive group
of 102 children referred over a 2 year period
by psychiatric professionals, state and private
agencies and treated in the course of private
practice. Seventy-eight percent of the children
were adopted. Children were included whose
parents agreed to complete one or more written
measures of behavior prior to treatment. The
children were diagnosed prior to referral with two
or more of the following disorders: ODD, CD, ADHD,
and RAD. Prior treatments that had been unsuc-
cessful included parental and professional inter-
ventions and medications. PPCE Therapy was
funded through state aid, private insurance, sub-
sidies, and self-payment.
Comparison group
In order to provide a benchmark for the behavior of
the children in our study, we compared their scores
on the Child Behavior Checklist (CBCL) before and
after treatment to the CBCL age- and gender-
matched normative data derived from a large
national (USA) sample (n¼4220) of referred
(disturbed) and non-referred (healthy) children.
14
Scores of the study group at baseline before
treatment were compared to the referred or
‘‘disturbed’’ scores in the normative group. Scores
of the group at follow-up were then compared
to the non-referred or ‘‘healthy’’ scores of the
normative group.
Personnel
A psychiatrist trained and supervised a team of
therapeutic staff that included psychiatrists, psy-
chologists, social workers, licensed professional
counselors and volunteer mentor families. Thirty-
four (33%) of the families brought their child’s local
therapist to the intervention. Volunteer parents
and their children aged 8–18 who had previously
received PPCE Therapy mentored the new families.
One lead therapist and three to five co-therapists
and mentors were assigned to each family for the 2-
day intervention. The chief psychiatrist supervised
interventions in each group in rotation.
Two-day intensive PPCE intervention
The intensive component of the intervention con-
sisted of two consecutive 8-h days of PPCE Therapy
in a single large room. Four to seven families were
treated simultaneously but independently in each
session on separate cushioned mats with coaching
from the therapist teams and mentor families. Up to
60 people attended each session, including thera-
pists, mentor families, identified patients, parents,
ARTICLE IN PRESS
M.G. Welch et al.4
siblings, grandparents, and others close to the
child.
The 2-day sessions consisted of (1) repeated PPCE
Therapy between mother and child; (2) PPCE
Therapy with other family members; (3) educa-
tional discussion on behavior and parenting; (4)
motivational discussion; and (5) planning for PPCE
at home.
The purpose of PPCE Therapy is to create
attunement between parent and child, thereby
modulating the stress response of the child and
mitigating aberrant behavior (Fig. 6). A typical
PPCE interaction between mother and child con-
sisted of the mother embracing the child face to
face for 20–90 min. Such an interaction progressed
through a sequence of four stages (with associated
emotional states): (1) Confrontation (discomfort,
irritability); (2) Conflict (expressing fears, anger, or
hurt); (3) Resolution of conflict (reciprocal em-
brace, caressing, kissing, conversation); (4) Syn-
chrony (attunement, reciprocity, state of calm
arousal).
15
During stage 1, the child typically
expressed discomfort or irritability and tried to
reject the parents’ closeness. During stage 2,
parents were encouraged to tolerate their child’s
strongest emotional outbursts, to persist through
the rejection, and to express their own intense
feelings of fear, anger, or hurt. Stages 1 and 2
typically resolved in 20–90 min, depending on the
degree of conflict. A successful PPCE cycle ended in
stages 3 and 4, with resolution of emotional conflict
and a state of synchrony and attunement between
the pair. These latter stages were characterized by
breathing in unison, deep mutual gaze, relaxation,
reciprocal pleasure in each other’s embrace, and
open verbal and non-verbal communication.
Therapists and mentor families functioned as
coaches and facilitators to ensure that the dyads
persisted through all four stages of the cycle. The
therapists and mentors did not participate physi-
cally in PPCE Therapy with the child or family
members and did not interpose themselves be-
tween family members. Sequential PPCE between
various family dyads and repeated mother–child
PPCE continued for most of the 2 days. At the end
of the second day, the therapeutic staff assisted
families in developing a course of action to be
followed at home.
Continued follow-up PPCE intervention by
parents
Upon returning home, the families implemented
the planned course of action. This plan typically
consisted of PPCE on a regular basis once or twice a
day and as needed for acute behavioral episodes
such as tantrums. Mentoring, therapist support and
parenting advice were given by phone in the course
of gathering follow-up data at defined intervals and
as needed.
Written assessment of outcome
Informed consent was obtained from parents for
anonymous analysis and presentation of the data.
Parents in each family completed one or two
baseline written measures 1–10 days prior to
treatment and were asked to complete at least
one measure during follow-up. These measures
included the Achenbach CBCL and the Randolph
Attachment Disorder Questionnaire (RADQ).
The follow-up written measures were received in
our office via mail and facsimile, or reported by
phone to a staff member. To evaluate the effect of
the intervention, we compared baseline scores on
the CBCL and RADQ instruments with subsequent
scores using independent sample t-tests. Among
children for whom we had repeated measurements,
we compared baseline to post-intervention scores
using paired sample t-tests. Because CD has been
reported to be more intractable among adolescents
than among younger children and among girls than
among boys,
16
we wanted to determine how
effective the therapy was by age and gender. We
stratified our repeated measurements by subgroup
of age and gender. We then compared their pre-
and post-intervention total problem scores to the
CBCL normative comparison group for correspond-
ing strata, based on referred (i.e. with behavior
problems) and non-referred (i.e. without behavior
problems) children. We then compared the pre- and
post-intervention scores within the strata to the
entire sample.
Measures
The parent-scored CBCL is a 120 item standardized
instrument that is completed by a parent. The CBCL
is designed to record children’s problems in a
standardized format. Parents rate the presence and
frequency of certain behaviors on a three-point
scale (0 ¼not true, 1 ¼somewhat or sometimes
true, and 2 ¼very true or often true). Examples of
behaviors assessed include ‘‘gets into many fights,’’
‘‘hurts animals,’’ etc. The CBCL has been shown to
distinguish ‘‘referred’’ or clinically diagnosed from
‘‘non-referred’’ or normal children. Scores are
based on percentiles of results from a normative
sample drawn from a subset of non-referred
subjects in a national (USA) sample (n¼4220).
ARTICLE IN PRESS
Outcomes of Prolonged Parent–Child Embrace Therapy 5
Research on this instrument reported excellent
test–retest reliability, inter-parent agreement, and
construct validity. Scores above 55 indicate patho-
logic range.
17
The RADQ is a 30-item instrument encompassing
the most common adverse symptoms of the CBCL
associated with this study’s population. It is widely
used by therapists treating families with adopted
children diagnosed with RAD together with ODD,
CD and/or ADHD. Validity has been constructed
through comparison of total RADQ scores to the
CBCL using the Pearson product–moment correla-
tion. Scores above 65 indicate pathologic range.
18
Mean total scores for CBCL and RADQ before
intervention (baseline) and at follow-up were
compared. Significance was assessed by t-tests
and analysis of variance.
Oral assessment of outcome
In addition to written measures, parents were also
asked to make an oral assessment of their child’s
behavior by phone at intervals of 10, 30, 60, 90,
180, and 360 days after the 2-day treatment.
Parent oral assessment was obtained via tele-
phone at defined intervals during follow-up to
determine improvement and/or problems in the
child’s behavior. Parents were asked to character-
ize their child’s behavioral problems on a three-
point scale as being ‘‘severe’’ (often out of control,
unmanageable), ‘‘moderate’’ (sometimes out of
control, but manageable), or ‘‘mild’’ (in control,
manageable).
Results
Written assessment
Table 1 shows demographic characteristics for the
entire cohort of 102 children and for a subgroup of
60 children. The subgroup of 60 children had all
three assessment measures completed by parents.
These assessments included a baseline CBCL and 2-
month follow-up, a baseline RADQ and 1-month
follow-up, and parent oral assessments for 1 year.
The mean age of the 102 children was 9 years;
69% were aged 4–11 years and 31% were 12–18
years. The sample was ethnically and geographi-
cally diverse.
A total of 93 CBCL and 101 RADQ measures were
completed at baseline. At follow-up, 66 CBCL and 96
RADQ measures were completed. For this group, the
mean CBCL score fell from 68 to 35 and the mean
RADQ score fell from 63 to 29 (Po0:001) (Fig. 1). For
the subgroup of 60 children, the mean CBCL score
fell from 70 to 37 and the mean RADQ score fell from
61 to 30 (Po0:001) (Fig. 2).
The families of 80 (78%) of the 102 children
received the following additional inputs: 21 re-
ceived extra telephone mentoring and advice; 28
received extra therapist coaching; 31 received both
extra mentoring and extra therapist coaching.
Severity of behavioral problems at the start of
the study did not appear to affect the outcome of
the treatment. In the subgroup of 60 children, 40
(67%) had CBCL scores in pathologic range at
baseline, while at 2-month follow-up that number
dropped to 7 (12%). In the subgroup of 60 children,
37 (61%) similarly had RADQ scores in pathologic
range at baseline, while at 1-month follow-up the
number in pathologic range was 8 (13%).
Comparison group
The mean CBCL total problem scores for this study’s
subgroup of 60 were compared with the mean CBCL
total problem score for the normative sample
(n¼4220).
14
The mean CBCL score at baseline for
the 60 children was 70, considerably higher than
ARTICLE IN PRESS
Table 1 Demographics of PPCE study population.
Characteristics Total cohort Subgroup
N%N%
Total 102 100 60 100
Mean age in years
(SD)
9.7 (3.55) —9.4 (3.97) —
Age 4–11 70 69 40 67
Age 12–18 32 31 20 33
Gender
Male 52 51 31 52
Female 50 49 29 48
Adopted 80 78 46 77
Race/ethnicity
White 64 63 36 60
Black 10 10 6 10
Asian 6 6 4 7
Hispanic 6 6 5 8
Mixed/other 16 15 9 15
Geographical residence
Northeast 27 26 16 27
Southeast 6 6 4 7
Midwest 60 59 35 58
West 9 9 5 8
M.G. Welch et al.6
the score of 53 for the referred or so-called
‘‘disturbed’’ children in the comparison group.
Notably, the mean CBCL score at the 2-month
follow-up was well within one standard deviation of
the mean score for the non-referred or the so-
called ‘‘healthy’’ children in the normative data set
(Fig. 4).
Gender and age for the group of 102 children in
this study did not appear to make a difference
in the outcome of the treatment. CBCL scores
declined greatly between baseline and follow-up
for both males and females under age 12 and age 12
and older (Po0:001). Comparison with normative
data by gender and age for the entire group (boys
and girls 4–11, n ¼582 and 619; boys and girls
12–18, n ¼450 and 459) was similar for the data for
the subgroup of 60. Mean scores in all gender and
age categories were higher for the PPCE group
before treatment than the referred or ‘‘disturbed’’
normative group. Mean scores after PPCE treat-
ment of the group of 102 children for all gender and
age categories were well within one standard
deviation of the mean for the non-referred or
‘‘healthy’’ group (Fig. 5).
Oral assessment
Parents oral assessment of behavior showed 73
children with severe and 29 with moderate beha-
vioral problems at baseline. At the 10-day follow-
up, the parents reported 21 children with severe,
20 with moderate and 57 with mild behavioral
problems. At the end of 1 year parents reported 6
with severe problems, 6 with moderate and 84 with
mild behavioral problems (Fig. 3).
Loss to follow-up
Following the initial 2-day intervention, phone
contact was lost with the family of one child and
therefore no follow-up written measure was sub-
mitted. Two children had no follow-up written
measure, although regular phone contact was
maintained with their families throughout the
study period. There were 5 children from 5
different families with whom contact was comple-
tely lost after the follow-up. All 5 of these children
were adopted, 4 by single mothers. Although the
group of 5 children showed a mean drop of 34% in
their RADQ scores at follow-up, their adoptions
were terminated by the parent(s) for known
reasons in some cases and unknown reasons in
others.
Unsuccessful participants
Based on the parents’ reports, 6 children did not
show improvements in behavior following the PPCE
ARTICLE IN PRESS
Figure 1 Entire study group of 102 children showing two
measures, at baseline and follow-up, Po0:001 for both
measures.
Figure 2 Subgroup of 60 showing two measures, at
baseline and follow-up, Po0:001 for both measures.
Figure 3 Parent oral assessments of behavioral problems
for all 102 children for 1 year follow-up as reported by
phone. Number of children in each category is indicated
in each column.
Outcomes of Prolonged Parent–Child Embrace Therapy 7
Therapy. In the follow-up year, the parents of one
child divorced and the mother stopped doing the
therapy. Three of the children belonged to a single
mother who did not have a family support system;
she felt unable to continue PPCE on her own at
home as planned. No details are known about
events in the follow-up year for the remaining two
children.
Medications
At baseline, 48 of the 102 children were on 75
prescriptions for 20 different pharmaceutical
agents. At the end of 1 year, with the parents of
96 children reporting by phone interview, 14
children were on 18 prescriptions for 6 different
agents.
Discussion
The positive results obtained in this study with
PPCE Therapy are somewhat unexpected. The
initial therapist-facilitated component was brief
and the complex symptoms of the treatment group
(ODD, CD, RAD, ADD, and/or ADHD) are frequently
resistant to treatment and considered to be chronic
(Table 2). Despite the fact that many of the
children had experienced deprivation and neglect,
including maternal separation and abuse in their
early years, the intervention proved to be effective
from the parents’ perspective. Symptoms were
reported to improve rapidly and the effects
were substantial and consistent over time for most
of the children, regardless of gender, age, adoption
status, race, geographic region, and severity of
behavior at baseline.
The mean CBCL and RADQ scores for the 102
children fell approximately 50% from baseline
to follow-up (Fig. 1). The parent’s oral assessment
confirmed a significant drop in severe behavior
that was sustained for the 1 year of follow-up
(Fig. 3).
In order to determine whether results were due
to incomplete data or to a bias in results from
those children with complete data, we analyzed a
subgroup of 60 children who had all three measures
at baseline and follow-up. We then compared the
subgroup to the entire data set of 102 children. The
demographics and results for the subgroup did not
differ significantly from the overall set of 102
(Table 1). For the set of 102, the CBCL mean score
dropped 49%, as compared with a 47% drop for the
subgroup of 60. For the entire set the RADQ score
dropped 54% compared to 51% for the subgroup
ARTICLE IN PRESS
Table 2 Comparison of our results with those of other interventions as measured by parent-scored CBCL total problem scores or Tscores.
First author Year NAge range
(years)
Intervention Type of CBCL
score
Mean
baseline
score
Mean interval Mean follow-
up score
Reduction in
score (%)
Welch This study 102 4–18 PPCE Therapy Total problem 70.1 2 mos 34.1 51
Armenteros
19
2002 11 7–15 Citalopram Total problem 74.6 1.5 mos 68.3 8
Visser
16
2003 1286 4–18 Multiple
psychiatric
services
Total problem 56.1 6.2 yr 40.5 28
Welch This study 102 4–18 PPCE Therapy T69.1 2 mos 51.9 25
Dalton
20
2000 33 5–12 Pharmacotherapy T69.7 3 mos 61.3 12
Dalton 2000 13 5–12 Psychotherapy T69.7 3 mos 67.1 4
Table summarizes results of 5 studies using different treatments for behavioral disorders in children aged 4–18. Results are given in CBCL total problem scores (raw) at baseline and
follow-up with % change.
CBCL Tscores are computed from the CBCL raw scores and are used to provide a common mean and standard deviation for subscales.
M.G. Welch et al.8
(Figs. 1 and 2). We concluded that the 60-children
subgroup was representative of the larger sample
from which it was drawn.
Parents tended to fill out the shorter RADQ
measure (30 questions) as opposed to the lengthier
CBCL measure (120 questions). Parents of 66 of
children a CBCL follow-up measure (65%); 93
completed a follow-up RADQ. The number of
written measures for all children declined sharply
after 2 months, while the number of periodic
parent oral assessments was constant for the entire
group at nearly 95% for the 1 year of follow-up
(Fig. 3). As children improved, parents chose to
give oral assessments, rather than fill out the
time consuming CBCL or RADQ written measures.
The loss to follow number of 5 (5%) and the
unsuccessful participants number of 6 (6%) were
low. The decline in scores from baseline to follow-
up was statistically significant for both CBCL and
RADQ (Po0:001).
This study had several methodological strengths
that helped validate the results. The set of 102 and
the subgroup of 60 children were demographically
identical (Table 1). While the numbers of subjects
at follow-up in subcategories of teenage boys and
girls could be considered low (12 and 12 respec-
tively), the size effect of the drop in mean CBCL
total problem scores overcomes this limitation:
80–35 (56% drop) and 90–39 (57% drop), respec-
tively. The percentage drop seen in our study was
much greater than in other studies using the CBCL
as their outcome measure (Table 2). Although this
clinical series lacked a control group, we were able
to compare the PPCE treatment data to the CBCL
normative data, which was drawn from a large
national sample that matched referred and non-
referred children (Figs. 4 and 5). The comparison
showed that at baseline the PPCE group was higher
than the referred sample. At follow-up the PPCE
group was well below pathologic range and was, in
fact, within one standard deviation of the non-
referred sample.
This case series report has several limitations.
Regression to the mean, the tendency for scores to
move toward the norm, may have accounted for a
portion of the improvement. However, regression
to the mean for the CBCL has been reported to be
only 5%,
17
while the improvement after PPCE was
49%. Selection factors cannot be ruled out entirely.
Although we entered every consecutive patient
during the period of the study, and the patients
were referred to us, it is possible that self-selection
by the participants may have produced some bias
in the results. Therefore, this source of bias
needs to be ruled out by a future randomized
controlled study.
Family members were required to devote two full
days to the initial intervention and most families
traveled from out of state to participate. The study
participants therefore represented families with a
high level of commitment. The study does not
account for other factors, such as the involvement
of other family members, the effects of group
interaction, the contribution of mentor families,
the follow-up phone support, the placebo effect
and the degree to which families implemented
PPCE at home.
We have made a number of observations that
may help account for the positive results of the
study. First, PPCE Therapy appears to replicate or
reinstate the effects of normal early nurturing on
the infant, which serves to modulate behavior.
21,22
It is not generally thought that such nurturing
activities can be applied to older children. How-
ever, we observed throughout this study that it is
possible for a mother who is supported by her
husband and/or extended family at home to
effectively modulate the behavior of even the most
difficult child over an extended period of time using
nurturing activities that are common in interactions
with infants.
Winnicott states that the mother provides con-
tinuity by ‘‘holding’’ the infant in an environment
she creates, thereby facilitating development.
23,24
A child can use other relationships only in so far as
he or she has been able to use the mother-child
relationship properly to foster development.
25
We
would say further that the mother–infant and
father–infant relationships create the models for
all future relating. When the home nurturing
ARTICLE IN PRESS
Figure 4 Comparison of PPCE study subgroup of 60 to
CBCL normative group by referral status using mean CBCL
total problem scores. Note: Compare PPCE baseline
(prior to treatment) to the normative referred (dis-
turbed) group and PPCE follow-up to normative non-
referred (healthy).
Outcomes of Prolonged Parent–Child Embrace Therapy 9
relationships were missing, as with orphans in this
study who were reared in inadequate orphanages,
the normal stress adaptation network of the child
was impaired and symptomatic behavior ensued.
Development was also impaired in the case of birth
children in this study, who for reasons of environ-
mental and/or biological insults could not benefit
from nurturing parental interactions. We believe
that PPCE Therapy was effective because it
organized and focused family nurturing activities
so that that the child could benefit from them.
It has been observed that the lack of involvement
of the father is a limitation in many family
therapies.
26
The importance of the mother to
development is widely accepted. It is supported
by substantial animal research, as well as by
considerable human data demonstrating the im-
portance of the mother–infant dyad. It is also
supported by our clinical observations. Nonethe-
less, we found that working with the family system
was critical in enabling the mother to fulfill the
emotional and physical needs of the child. In some
cases, issues leading to family resistance necessi-
tated beginning with the parent–parent dyad and/
or parent/grandparent dyads. In one case, the
entire first day was focused on the mother–grand-
mother dyad. In another case, the therapy was only
able to proceed after working for 5 h with the
mother/mother-in-law dyad to resolve conflict.
When the mother was supported by her husband
and/or her own parents or parents-in-law in her
efforts to modulate the stress responses of a
dysregulated child, the task was much easier.
There were several single mothers in this study.
Those that had grandparent support did well,
whereas those that did not have the support of
either a husband or grandparents did not do as
well. Of the five children who were lost to follow-
up, four had single mothers with no family support.
We believe that contact between mother and
infant—breast feeding, holding, kissing, vocaliza-
tion, odor cues, eye contact, and facial expression-
s—leads to mother–child synchronous attunement.
PPCE Therapy appears to enable the mother,
whether biological or adoptive, to modulate the
stress responses of the child at any age. We theorize
that the nurturing activities return the child’s
physiology and emotions to a state of balance,
thus freeing the child from symptomatic behavior
(Fig. 6).
During PPCE, we observe that both mother and
child quickly learn or relearn to associate the
comfort of synchronous attunement in each other’s
embrace with the accompanying stress-free state
of homeostasis or calm arousal. After many PPCE
nurturing cycles, the child’s response to stress is
positively conditioned and the child is able to
maintain calm arousal for longer and longer periods
of time. With repeated achievement of a synchro-
nous and modulated state, the child also develops
empathy. He or she learns to give as well as receive
comfort.
The fact that PPCE Therapy ameliorated a wide
range of pathology over the timeframe of the
study suggests that a common physiological me-
chanism may underlie many behavioral disorders.
ARTICLE IN PRESS
Figure 5 Comparison of mean CBCL total problem scores
in entire PPCE study group and CBCL normative group
according to gender and age. Note: Compare PPCE
baseline (prior to treatment) to the CBCL normative
referred (disturbed) group and PPCE follow-up to CBCL
normative non-referred (healthy) group.
Figure 6 The Stress-response Conditioning Cycle of PPCE
Therapy that serves to modulate the child’s behavior.
With repeated cycles the child learns to control sympto-
matic responses to stress.
M.G. Welch et al.10
We hypothesize that the common behavioral
mechanism involves neuropeptides normally re-
leased by mother–infant nurturing interactions. In
our laboratory we have been investigating the
mechanisms by which neuropeptides of nurture
may lead to the amelioration of stress-related
inflammation in the gut and the brain. Systemic
administration of neuropeptides in an animal model
of peripheral and central dysregulation is the
subject of our current research.
27–31
Our clinical observations, research hypotheses
and lab findings support the concept of equifinality,
that is, that multiple causes and processes can
result in one symptom complex. In terms of
treatment, however, the results of this study
support the concept of multifinality, i.e. multiple
symptom complexes can originate from a single
factor or mechanism.
9
Thus, therapies that target
the anti-stress mechanisms that underlie family
nurture systems may provide effective treatments
for a range of childhood developmental disorders.
The view is now widely shared that neurobiolo-
gical development and experience are mutually
influencing.
32–34
Though this study did not assess
the impact of PPCE Therapy on children past 1 year,
clinical observation of other patients in our
practice who received PPCE Therapy from their
families throughout childhood development sup-
ports the concept that family nurture can have a
dramatic and positive impact on behavior and
development.
In summary we conclude that, within the limita-
tions of this study, PPCE Therapy in children with
severe behavioral symptoms led to significantly
improved behavior. The results suggest that a
multiple family group, intergenerational family
system treatment model that targets stress-re-
sponse conditioning can be effective in ameliorat-
ing even severe symptomatic behavior in children
for extended periods of time. The results also
challenge a widespread belief that severe beha-
vioral disorders are chronic. We believe that out-
comes offer promise for treatment of a wide range
of childhood developmental disorders and justify
further study at the bedside and at the bench.
Acknowledgments
This treatment program was conducted during the
course of private practice. We wish to thank the
following for their contributions to this work: D.A.
Ruggiero Ph.D, M.G. Opler Ph.D, MPH, C.I. Berlin
Ph.D, and all the families who participated in this
study. Special appreciation is extended to S.A.
Power for close reading. Dr. Welch wishes to
acknowledge the late Nobel Laureate Niko Tinber-
gen’s participation and support in the development
of the therapy and his encouragement to pursue
the underlying basic science.
Reference
1. Pedersen CA, Boccia ML. Oxytocin antagonism alters rat
dams’ oral grooming and upright posturing over pups.
Physiol Behav 2003;80:233–41 [SmmaryPlus | Full Text +
Links | PDF (239 K)].
2. Francis DD, Young LJ, Meaney MJ, Insel TR. Naturally
occurring differences in maternal care are associated with
the expression of oxytocin and vasopressin (V1a) receptors:
gender differences. J Neuroendocrinol 2002;14:349–53
[Abstract and Full Text].
3. Anisman H, Zaharia MD, Meaney MJ, Merali Z. Do early-life
events permanently alter behavioral and hormonal re-
sponses to stressors? Int J Dev Neurosci 1998;16:149–64
[SummaryPlus | Full Text + Links | PDF (299 K)].
4. Suchecki D, Rosenfeld P, Levine S. Maternal regulation of the
hypothalamic–pituitary–adrenal axis in the infant rat: the
roles of feeding and stroking. Brain Res Dev Brain Res
1993;75:185–92 [Abstract | Abstract + References | PDF
(762 K)].
5. Uvnas-Moberg K. Antistress Pattern Induced by Oxytocin.
News Physiol Sci 1998;13:22–5 [Abstract and Full Text].
6. Liu D, Diorio J, Day JC, Francis DD, Meaney MJ. Maternal
care, hippocampal synaptogenesis and cognitive develop-
ment in rats. Nat Neurosci 2000;3:799–806 [Abstract | Full
text | PDF (262K)].
7. Weaver IC, Cervoni N, Champagne FA, D0Alessio AC, Sharma
S, Seckl JR, et al. Epigenetic programming by maternal
behavior. Nat Neurosci 2004;7:847–54 [Abstract | Full text |
PDF (650K) | Supplementary Information].
8. Welch MG. Retrieval from autism through mother–child
holding therapy. In: Tinbergen N, Tinbergen EA, editors.
‘Autistic’ children: new hope for a cure. London: George
Allen & Unwin; 1983. p. 322–36.
9. Davies PT, Cicchetti D. Toward an integration of family
systems and developmental psychopathology approaches.
Dev Psychopathol 2004;16:477–81 [Abstract and Full Text].
10. Cicchetti D. Developmental psychopathology: reactions,
reflections, projections. Dev Rev 1993;13:471–502 [Abstract
| PDF (1123 K)].
11. Cox MJ, Paley B. Families as systems. Annu Rev Psychol
1997;48:243–67 [Abstract | Full Text |].
12. Welch MG, Chaput P. Mother–child holding therapy and
autism. Pa Med 1998;91:33–8 [PMID: 3226740].
13. Weisz JR, Jensen AL. Child and adolescent psychotherapy
in research and practice contests: review of the
evidence and suggestions for improving the field. Eur
J Adol Psychiatry 2001; 10:1/12–1/18. [doi:10.1007/
s007870170003] [Abstract and Full Text].
14. McConaughy SH, Stanger C, Achenbach TM. Three-year
course of behavioral/emotional problems in a national
sample of 4- to 16-year-olds: I. Agreement among
informants. J Am Acad Child Adol Psychiatry 1992;131:
932–40.
15. Welch MG. Holding time. New York: Simon & Schuster; 1988
[BOOK available].
16. Visser JH, van der Ende J, Koot HM, Verhulst FC. Predicting
change in psychopathology in youth referred to mental
ARTICLE IN PRESS
Outcomes of Prolonged Parent–Child Embrace Therapy 11
health services in childhood or adolescence. J Child Psychol
Psychiatry 2003;44:509–19 [Abstract and Full Text].
17. Achenbach TM, Edelbrock C. Manual for the Child Behavior
Checklist and Revised Child Behavior Profile. Burlington, VT:
University of Vermont; 1991.
18. Randolph EM. Randolph Attachment Disorder Questionnaire.
Evergreen, CO: The Attachment Center Press; 1997.
19. Armenteros JL, Lewis JE. Citalopram treatment for im-
pulsive aggression in children and adolescents: an open
pilot study. J Am Acad Child Adol Psychiatry 2003;41:
522–9.
20. Dalton R, Pellerin K, Carbone V, Theriot A, Thibodeaux D,
Stewart L, et al. Treatment outcome among child psychiatric
outpatients in a community mental health center. Commu-
nity Ment Health J 2000;36:195–203.
21. Klaus MH, Jerauld R, Kreger NC, McAlpine W, Steffa M,
Kennel JH. Maternal attachment. Importance of the first
post-partum days. N Engl J Med 1972;286:460–3.
22. Klaus M. Mother and infant: early emotional ties. Pediatrics
1998;102:1244–6 [Abstract and Full Text].
23. Winnicott DW. Mother and child. A primer of first relation-
ship. New York: Basic Books, Inc.; 1957.
24. Winnicott DW. The theory of the parent–child relationship.
Int J Psychoanal 1960;41:585–95.
25. Winnicott DW. The use of an object. Int J Psychoanal
1969;50:711–6.
26. Cowan PA, Cowan PC. Interventions as tests of family
systems theories: marital and family relationships in
children’s development and psychopathology. Dev Psycho-
pathol 2002;14:731–59 [Abstract and Full Text].
27. Welch MG, Keune JD, Welch-Horan TB, Anwar M, Anwar N,
Ruggiero DA. Secretin activates visceral brain regions in rat
including areas abnormal in autism. Cell Mol Neurobiol
2003;23:817–37 [Abstract and Full Text].
28. Welch MG, Welch-Horan TB, Anwar M, Keune JD, Anwar N,
Ludwig RJ, et al. Secretin: hypothalamic distribution and
hypothesized neuroregulatory role in autism. Cell Mol
Neurobiol 2004;24:219–41 [Abstract and Full Text].
29. Welch MG, Welch-Horan TB, Anwar M, Anwar N, Ludwig RJ,
Ruggiero DA. Brain effects of chronic IBD in areas abnormal
in autism and treatment by single neuropeptides secretin
and oxytocin. J Mol Neurosci 2005;25:259–74 [PMID:
15800379].
30. Welch MG, Ruggiero DA. Predicted therapeutic role of
secretin and oxytocin in autism: implications for treatment
of mental illness. In: Dhossche DM, editor. GABA in autism
and related disorders, vol. 71, International review of
neurobiology. San Diego, CA: Elsevier/Academic Press; 2005
in press.
31. Welch MG, Ruggiero DA. Welch method therapy. In: Myles BS,
Swanson TC, Holverstott J, editors. Autism spectrum
disorders: an encyclopedia for parents and educators.
Westport, CT: Greenwood Publishing Group, in press.
32. Cicchetti D, Tucker D. Development and self-regulatory
structures of the mind. Dev Psychopathol 1994;6:533–49.
33. Kandel ER. A new intellectual framework for psychiatry. Am
J Psychiatry 1998;155:457–69 [PMID:9545989].
34. Kandel ER. Biology and the future of psychoanalysis: a new
intellectual framework for psychiatry revisited. Am J
Psychiatry 1999;156:505–24 [PMID:10200728].
ARTICLE IN PRESS
M.G. Welch et al.12
