1. DEPRESSION AND ANXIETY 19:51–58 (2004)
Brief Report
TREATMENT OF DEPRESSED MOTHERS OF
DEPRESSED CHILDREN:
PILOT STUDY OF FEASIBILITY
Helen Verdeli, Ph.D.,n Tova Ferro, Ph.D., Priya Wickramaratne, Ph.D., Steven Greenwald, M.P.H.,
Carlos Blanco, M.D., Ph.D., and Myrna M. Weissman, Ph.D.
Numerous studies have shown that depression is highly familial and impairing
and that a history of depression in a parent is the strongest risk factor for
depression in a child. Many of the parents in these studies have never received
sustained treatment despite histories of recurrent depression. None of the studies
have examined the effects of maternal symptom remission on offspring
symptom or functioning. We sought to determine the feasibility of treating
depressed mothers who brought an offspring for the treatment of depression and
to examine the relationship between improved maternal depression and
symptomatic improvement and social functioning in their offspring. Nine
mothers bringing their offspring for treatment of depression, and who were
evaluated and found to be currently depressed, completed a 12-week open trial of
interpersonal psychotherapy. Mothers and their depressed offspring were
assessed by independent evaluators at weeks 0, 6, and 12 for depressive
symptomatology and social functioning. Although the rates of depression were
high among the mothers, few eligible mothers agreed to participate. Of the 12
who entered treatment, 9 (75%) completed it. Mothers and offspring improved
with regard to depressive symptomatology and global functioning over the course
of the trial. Improvement in maternal depression was significantly associated
with improvement in offspring functioning but not symptom reduction.
Improvement of maternal depression may be associated with improved outcomes
in depressed offspring. However, it is difficult to engage depressed mothers in
treatment for themselves if they come to the clinic to bring their child for
treatment of depression. It may be more feasible to study the effect of improved
maternal depression on offspring by sampling depressed mothers
coming for their own treatment and then assessing their children over
the course of maternal treatment. Depression and Anxiety 19:51–58, 2004.
& 2004 Wiley-Liss, Inc.
Key words: depressed mothers; IPT; depressed children
n
Divisions of Child Psychiatry and Clinical and Genetic Correspondence to: Dr. Helen Verdeli, Division of Clinical and
Epidemiology, New York State Psychiatric Institute, Depart- Genetic Epidemiology, New York State Psychiatric Institute-
ment of Psychiatry, Columbia University College of Physi- Columbia University, 1051 Riverside Drive, Unit 24, New York,
cians and Surgeons and Mailman School of Public Health, NY 10032. E-mail: verdelih@child.cpmc.columbia.edu
Columbia University, New York, New York
Received for publication 5 November 2002; Accepted 25 August 2003
Presented as an Invited Address at the New York State Office
DOI 10.1002/da.10139
of Mental Health Annual Meeting, December 5, 2000, Albany,
Published online in Wiley InterScience (www.interscience.wiley.com).
New York
& 2004 WILEY-LISS, INC.

2. 52 Verdeli et al.
INTRODUCTION for suicide, (3) were currently in treatment for
Epidemiologic depression, or (4) had a medical condition that could
studies across diverse cultures have
cause depression. Inclusion criteria for offspring were:
shown that the highest rate of depression occurs in
(1) age 6–18 years, (2) initially identified through case
women and the highest risk for first onset of major
records as having a diagnosis of a DSM-IV unipolar,
depression is in women of childbearing years [Kessler
nonpsychotic depression made by clinic staff (child
et al., 1993; Regier et al., 1987; Robins et al.,
psychiatrists, doctoral-level child psychologists, and
1984;Weissman et al., 1988, 1996]. Depression is also
clinical child social workers), (3) in current treatment,
associated with impairment in social and occupational
(4) an IQ above the 75 percentile as assessed with the
functioning, including parenting [Broadhead et al.,
Test of Nonverbal Intelligence, and (5) not participat-
1990; Johnson et al., 1992; Murray and Lopez, 1996,
ing in another research protocol. Comorbid diagnoses
Wells et al., 1989]. History of depression in a parent is
in offspring were not a basis for exclusion. All subjects
one of the strongest risk factors for depression in a
provided written informed consent, and their offspring
child, with a two- to threefold increased risk for
written informed assent, after they were given a
occurrence in offspring of depressed parents [Hammen
complete description of the study.
et al., 1990; Weissman et al., 1997].
Our recent studies found that in addition to the high
lifetime rates of depression in women and their PROCEDURE
offspring, high rates of concurrent depression exist in
Screen. Before entering the study, mothers were
mothers and their offspring brought for treatment of
screened for depression, anxiety disorders, and sub-
depression. However, only a minority of mothers who
stance use disorders with the Patient Problem Ques-
have an episode of depression concurrent with their
tionnaire [PPQ; Spitzer et al., 1994; Spitzer, 1997]
offspring (22%) are receiving psychiatric treatment
when they brought their offspring for treatment of
[Ferro et al., 2000]. This finding is consistent with
depression. The PPQ screening procedure and results
reports from community samples in which only about
of its use in a screening study have been previously
one third of depressed people receive treatment
described in detail [Ferro et al., 2000].
[Kessler, 1994; Leaf et al., 1985; Young et al., 2001].
Eligibility evaluation. If mothers screened positive
An important and unexamined issue is how treatment
for depression and negative for substance abuse/
of maternal depression af fects the symptomatology and
dependence on the PPQ, they were asked to participate
functioning of their depressed offspring who are also in
in an eligibility evaluation for the study which involved
treatment.
a structured diagnostic assessment with the 24-item
In this report, we examine the feasibility of identify-
Hamilton Rating Scale for Depression [HRSD; Ha-
ing the depressed mothers for treatment from among
milton, 1960] on which the mother had to score above
those who were bringing their offspring for treatment
12, and the Structured Clinical Interview for Diag-
of depression. We also examined the relationship
nostic and statistical manual of mental disorders, 4th
between improved maternal depression, and sympto-
Edition [DSM-IV; SCID; Spitzer et al., 1992]. The
matic improvement and social functioning in their
other exclusion criteria were also assessed at this time.
offspring. We chose to treat maternal depression with
If after HRSD and SCID evaluation mothers were
interpersonal psychotherapy (IPT) because it is an
eligible to participate in the study, they were entered
empirically tested treatment for depression that also
into the open clinical trial.
addresses interpersonal functioning [Weissman et al.,
Independent evaluations. Upon entry to the study,
2000]. This was a pilot study designed to find a feasible
mothers and their offspring were assessed by an
research strategy for studying the impact of treating
independent clinical evaluator at weeks 0 (baseline), 6
maternal depression on offspring depression.
(interim), and 12 (termination) for depressive sympto-
matology/clinical status and social functioning. Mater-
SUBJECTS AND METHODS nal independent evaluations were carried out in a
blinded manner to those of their offspring and vice
SUBJECTS versa. Independent evaluations were conducted by
Subjects were mothers over 18 years of age who were clinically experienced interviewers who took part in a
ascertained when they sought evaluation or treatment training session with two of the authors (T.F. and H.V.).
for depression in their offspring at the Department of Training involved didactic information, mock evalua-
Child Psychiatry at Columbia Presbyterian Medical tions, and two supervised evaluations.
Treatment. During the open trial, mothers received
Center (CPMC) and themselves met DSM-IV criteria
for a unipolar, nonpsychotic depression. Additional weekly IPT for 12 weeks in 45-min sessions from an
maternal inclusion criteria were: (1) biological mother experienced clinician (e.g., MSW, PhD, or MD)
of depressed offspring, (2) residing with the offspring trained in the delivery of IPT by an experienced IPT
for Z1 year, and (3) fluent in English or Spanish. therapist. IPT is a time-limited, empirically based
Mothers were excluded if they: (1) had a current (past 6 psychotherapy for the treatment of depression [Weiss-
months) substance use disorder, (2) were at acute risk man et al., 2000]. If, according to their clinicians,

3. Brief Report: Depressed Mothers and Children 53
DATA ANALYSIS
mothers did not show improvement by week 6, one of
us (C.B.) undertook a medication evaluation of the
subjects. At the end of treatment, mothers were offered We conducted outcome data analyses using two
a referral for continued treatment in the community if approaches.
Random regression analysis. The change in out-
it was deemed necessary by the clinician. Children
received treatment as usual (psychotherapy and/or come variable that each individual (mother or child)
medication) at a CPMC pediatric psychiatry outpatient experiences over time is first modeled using a
clinic. polynomial regression model with the individual’s
outcome variable at each specific time as the dependent
variable and the number of weeks after baseline
MATERNAL-DEPENDENT MEASURES assessment as the independent variable. Because the
number of observations per individual is r3, we
Depression. We used the 24-item HRSD [Hamil-
ton, 1960] as a clinician-rated measure of depressive employed a linear regression model. The model [Byrk
symptomatology and its change over time. An HRSD and Raudenbush, 1992] posits that the individual
score of 8–17 reflects mild depression, 18–24 moderate response of each subject is a function of the number
depression, and 425 severe depression [Katz et al., of weeks since baseline assessment [a line with inter-
1995]. The Beck Depression Inventory [BDI; Beck cept (baseline response) and slope (improvement rate)].
et al., 1961] was used as a self-report measure of Models were also fitted using a quadratic term, to
depressive symptomatology. A BDI score of Z16 test for deviations from linearity. The model was
suggests a diagnosis of major depression [Katz et al., applied to (a) estimate a mean rate of improvement in
1995]. The Clinical Global Impressions Severity of outcome variable for each of the two groups, mothers
Illness Scale [CGI-SI; McGlashan, 1973] was also used. and children; and (b) examine the extent of variation
The CGI-SI is a brief clinician-rated scale that assesses around each of these means as follows. Initially,
overall severity of mental illness. coef ficients associated with the intercept and slope
Social Functioning. The Social Adjustment Scale- were treated as random coefficients to determine
Self-Report (SAS-SR; Weissman and Bothwell, 1976) whether the rate of improvement varies between
was employed to assess maternal social functioning. subjects. If there is no significant variation, the
The SAS-SR is a brief self-report instrument that is coef ficients are treated as fixed. Using a random
widely used to measure change in social functioning. A regression model to analyze longitudinal data permits
higher score represents greater impairment. Maternal one to use information at all assessment points in the
social functioning was also assessed with the Global study. A further advantage of this approach is that it
Assessment of Function [GAF; APA, 1994]. The GAF allows for missing observations and subjects measured
is a clinician-rated assessment of current (past month) at different times, as well as estimation of random
functional impairment. A higher score represents person-specific effects.
better functioning. Although the preceding analyses will let us deter-
mine whether the mother/child outcome scores im-
prove over the treatment period, it does not allow us to
OFFSPRING-DEPENDENT MEASURES determine whether the changes in children’s outcome
Depression. The HRSD, described above, was also scores (i.e., symptoms and functioning) are associated
employed to assess clinician rated depression in with changes in mother’s symptoms and functioning.
children. We employed the Children’s Depression To determine whether changes in children are related
Inventory [CDI; Kovacs, 1980] as a self-report measure over time to maternal change, we fitted a random
of depression in offspring. A score of Z19 indicates the regression model to child outcomes which included
presence of depression. The CGI-SI was also used to both time-varying and time-invariant covariates as
assess overall severity of mental illness in offspring. independent variables in the model.
Social Functioning. Social functioning of offspring Specifically, the regression model was as follows.
was assessed with the Social Adjustment Inventory for The dependent variable considered was each child’s
Children and Adolescents [SAICA; John et al., 1987]. outcome at each study time point, the time variable
The SAICA is derived from the SAS for use with (i.e., number of weeks from baseline assessment) was
children and adolescents and, like the SAS, a higher the independent variable, and maternal score (HAM-
score indicates greater impairment. Offspring social D, GAS, etc.) at the dif ferent study time points was
functioning was also assessed with the Children’s treated as a nonrandom time dependent covariate. If
Global Assessment Scale [C-GAS; Shaffer et al., the estimate of the coef ficient corresponding to this
1993]. The C-GAS is a clinician-rated assessment of nonrandom time dependent covariate is significantly
current (past month) functional impairment in all areas different from zero, we can infer an association
of a child’s life, with a higher score representing better between change in mother’s score and change in child’s
functioning. A cutoff score of 61 was determined to outcome. All random regression analyses were per-
best discriminate cases of psychopathology in an formed using the Proc Mixed software in the SAS
system (SAS, Cary, NC).
epidemiologic study by Bird and colleagues [1987].

4. 54 Verdeli et al.
Endpoint analysis. The sample consisted of all Nine (75%) of the 12 mothers completed the open
mothers and children who entered treatment. In this trial of IPT. One dropped out at week 3 when her child
analysis, the last score obtained either at interim or was placed in foster care and subsequently hospitalized;
early termination evaluation was used for mothers and another at week 2 owing to lack of time for treatment;
children, respectively, who dropped out of the study. and the third at week 4 stating that treatment had
The end point samples consisted of 12 mothers and 12 helped her and she felt no need to continue. The third
children, respectively. The results of an end point mother and her child completed a termination evalua-
analysis may be considered an estimate of the overall tion and these data are included in the random
performance of a treatment program, including the regression analyses where the sample size is therefore
ability to retain patients in treatment. equal to 10. For the nine mothers who completed the
We assessed dif ferences between baseline scores and open trial, the average number of sessions was 11.1 (sd
values at week 12 on each of the outcome measures 2.7). In addition to IPT, one mother received medica-
using paired t-tests to test whether these differences tion in conjunction with IPT from the study psychia-
were significantly dif ferent from zero. The Bonferroni trist when, following the guidelines of our protocol, she
method was used to control for Type 1 errors owing to did not improve by week 6 of treatment.
multiple comparisons. Both the actual P values and the
Bonferroni P values are reported.
OUTCOME ASSESSMENT
Maternal depression was treated with IPT [Weiss-
RESULTS man et al., 2000]. Our primary hypotheses were that:(a)
Between October 1997 and December 1998, 171 both maternal depressive symptoms and functioning
mothers were approached and asked to be screened for would improve as a result of treatment, (b) offspring
the study. Of these mothers, 168 (98%) agreed to be depressive symptoms and functioning would likewise
screened and 45 of 168 (27%) were eligible for improve, and (c) improvement of maternal depressive
diagnostic evaluation. Thirty-five (78%) of these symptomatology would be associated positively with
eligible mothers completed the study evaluations; 17 offspring improvement in depressive symptoms and
of the 35 (49%) met the study inclusion criteria and 12 functioning.
Effect of treatment on mothers. Results of the
(34%) agreed to participate. Mothers who were eligible
did not enter the study for a variety of reasons random regression analysis showed significant improve-
including offspring refusal of treatment, mothers ment over the 12-week study period in mothers’
wanting treatment closer to her home, and offspring depressive symptoms as measured by both the HRSD
participation in other research that precluded maternal and the BDI (Table 1). There were no significant
participation in a treatment study. differences between mothers in rate of improvement.
Whereas few mothers agreed to participate in the Mother’s HRSD decreased an average of 0.66/week
study, the rate of depressions in the mothers was high: during the study period, whereas mothers’ BDI decreased
15% screened positive for current depression. The an average of 0.83/week over the study period. In
mean age of participating mothers was 42 years [range addition a significant improvement over the study period
32–52, standard deviation (sd) 6.1] and their diagnoses, was observed in mothers’ global functioning measured by
based on the SCID interview, were major depressive the GAF and social functioning as measured by the SAS.
disorder (75%), dysthymia (8%), and concurrent major Results of the end point analysis (not shown) were similar
depressive disorder and dysthymia (17%). Mothers’ to those obtained from the random regression analysis
mean number of children (including full and half and can be obtained upon request.
siblings of the target offspring) was 3.5 (sd 2.1). Nearly Not shown, results of the random regression model
half of the mothers were Spanish speaking (42%) and indicate that baseline severity, as measured by the HRSD,
most were of minority origin (67% Hispanic, 17% were negatively correlated with rate of improvement,
African American, and 17% White). Most mothers implying that more severely depressed mothers showed
were married (33%) or separated/divorced (50%), greater improvement (r ¼ À.4478). These results how-
although 8% were never married and 8% widowed. ever, failed to reach statistical significance (P ¼ .2397).
Change in offspring outcomes over study period.
Most mothers (67%) were high school graduates and
33% had also completed some college. Mothers’ The random regression analysis showed significant
household income fell into the following categories: improvement in offspring’s depressive symptoms over
o$10,000 (60%), $10,000–25,000 (20%), and the study, as measured by the HRSD, with child HRSD
$25,000–60,000 (20%) (household income was missing scores decreasing an average of 0.44/week over the
for two mothers). Most mothers (58%) were not study period (Table 1). However, there was no
employed and 33% were on public assistance. significant difference between subject variation in
The mean age of the offspring was 14.1 years (sd 2.3) these rates of improvement. We observed a trend
and 67% were girls. Diagnoses of the offspring, based (Po.06) in the offspring’s CGI-SI, with offspring
on clinician rating, were major depressive disorder CGI-SI scores decreasing by an average of 0.06/week
(67%) and dysthymia (31%). over the study period. Results of the end point analysis

5. Brief Report: Depressed Mothers and Children 55
TABLE 1. Outcome of mothers and their children over the period of maternal treatment for depressionn
Random regression test of linear trends
(Baseline–Week 12)
Variable Baseline Week 6 Week 12 E4 P (E4)
Depression rating scalesa
7 7 7
HamiltonFmother 10/17.4 8.0 9/14.0 8.7 9/9.1 8.0 .04
À0.66
7 7 7
HamiltonFchild 10/12.7 4.6 8/7.0 3.5 9/7.5 3.6 .007
À0.44
7 7 7
BDIFmother 10/20.0 9.1 9/20.2 15.5 8/9.9 10.1 .03
À0.83
7 7 7
CDIFchild 10/13.1 5.6 8/12.2 5.2 7/10.4 7.5 .30
À0.20
Clinical global impressionsFSIa
10/3.6 7 0.8 9/3.2 7 1.1 9/2.7 7 1.2
Mother .11
À0.07
10/3.4 7 3.4 8/2.9 7 0.8 9/2.7 7 1.0
Child .06
À0.06
Global functioningb
10/56.0 7 5.0 9/59.2 7 9.5 9/69.0 7 9.3
GAFFmother 1.04 .009
10/58.4 7 7.5 8/60.1 7 8.8 9/63.1 7 8.0
C-GASFchild 0.40 .24
Social adjustmenta
10/2.5 7 0.4 8/2.0 7 0.7
SASFmother NA .05
À0.05
10/2.0 7 0.4 8/1.8 7 0.3
SAICAFchild NA .34
À0.01
Values are expressed as mean (sd), unless otherwise indicated.
n
N ¼ 10 owing to inclusion of data from the termination evaluation completed by the mother–child dyad who dropped out of the study at week 4.
a
A higher score indicates greater impairment.
b
A lower score indicates greater impairment.
NA ¼ variable was not assessed at this time.
were similar to those obtained from the random accepted treatment was comparable with the literature
regression analysis and can be obtained upon request. on clinical trials [Pocock, 1999]. However, recruitment
Relationship of maternal depression with of mothers of offspring attending an urban teaching
offspring outcome. We found no significant associa- hospital for participation in a clinical trial was fraught
tion between change in maternal depression over time with difficulties such as children not residing with their
and change in offspring depression over time for any biological mothers, cancelation of numerous appoint-
measure of depressive symptomatology. When the ments, and family relocation. Whereas the rate of
association between change in maternal depression (as depression was high in the mothers bringing their
measured by the HDRS and BDI) and offspring children to treatment, numerous mothers needed to be
functioning (as measured by the CGI-SI, C-GAS, and screened to find those who would accept treatment for
SAICA) was examined, the change in maternal HDRS themselves.
scores over time was significantly associated with Mothers and offspring in the present study improved
change in offspring C-GAS scores (i.e., decreasing with regard to depressive symptomatology and func-
maternal HRSD scores were associated with increasing tioning over the 12 weeks. Although both mothers and
C-GAS scores in offspring; Table 2). There was a 0.45 offspring improved over the course of the open
increase in offspring C-GAS score associated with a trial, the rate of maternal improvement was approxi-
one-unit decrease in maternal HRSD score (Po.03). mately twice that of offspring improvement and was
We conducted exploratory analyses to determine found in both independent evaluator and self-rated
whether a relationship existed between improvement in measures.
maternal global functioning and offspring outcome. Our finding of maternal improvement is consistent
No significant association was found between maternal with the literature on the treatment of depression with
GAF and offspring CDI, HRSD, CGI, C-GAS, or IPT in adults [Di Mascio et al., 1979; Elkin et al., 1995]
SAICA. However, when offspring outcomes were and whereas the offspring in the present study received
categorized into two groups by the median baseline treatment as usual and did not follow an established
values, there was a significant association (Po.002) treatment protocol, their improvement is suggestive of
between offspring who had a baseline C-GAS score of that found in empirical studies of the treatment of
60 and change in maternal global functioning. These depression in adolescents [Brent et al., 1997; Emslie
offspring had a 0.67 increase in C-GAS for every one- et al., 1997; Mufson et al., 1999]. The lesser rate of
unit increase in maternal GAF. improvement in offspring might be related to several
factors. Offspring treatment was not standardized and
many offspring were already engaged in treatment
DISCUSSION when their mothers entered the study and probably
experienced some symptomatic improvement before
The delivery of IPT was acceptable to mothers in our study began, as reflected in the offspring’s low
baseline HRSD. Finally, time must be considered
that the completion rate of 75% for those mothers who

6. 56 Verdeli et al.
TABLE 2. Regression effects of mothers’ Hamilton and global functioning correlated with their children’s outcome
Child’s outcomen P (E) df
E
Child’s Hamilton 0.06 .61 15
CDI 0.02 .88 13
CGI-SI 0.04 .14 15
C-GAS .03 15
À0.45
SAICA .40 6
À0.009
n
N ¼ 10.
CDI, Children’s Depression Inventory; CGI-SI, Clinical Global Impressions Severity of Illness Scale; C-GAS, Children’s Global Assessment
Scale; SAICA, Social Adjustment Inventory for Children and Adolescents.
a factor with regard to improvement of depressive form of psychotherapy were used. The present study is
symptomatology in both mothers and offspring. also limited in that many offspring were already in
Our hypothesis that reduction of maternal depres- treatment when the study began and the average age
sion would be associated with offspring outcome was was in adolescence, a period in which one’s peer group
supported for improvement in offspring functioning is usually the primary interpersonal influence. In
but not for improvement in offspring depressive addition, a longer interval may be required to observe
symptomatology. Furthermore, improvement in ma- the effects of maternal improvement on symptomatic
ternal global functioning was significantly associated outcome in offspring. It must also be acknowledged
with improvement in offspring global functioning that children in treatment for depression would be
when it was stratified on the basis of baseline severity. expected to improve over time even if their mothers
Surprisingly, we found no significant association were not receiving concurrent treatment for depres-
between improvement in maternal and offspring sion. Finally, the present study is generalizable to a self-
depressive symptomatology over the course of the selected group of moderately depressed mothers who
study. This might be explained by offspring treatment are predominantly urban, of low socioeconomic status,
status as discussed above or, alternatively, by the fact Hispanic, and bringing their offspring for evaluation or
that at age 14 the offsprings’ peer group exerted a treatment of depression. In light of these considera-
greater influence on them than did their mothers. It tions, the findings presented here should be considered
might also be that the association between maternal preliminary.
and offspring depressive symptomatology requires a
longer time (e.g., 6 or 12 months) to appear. It remains
CLINICAL AND RESEARCH IMPLICATIONS
unclear whether maternal depression has a stronger
association with offspring functioning than offspring Clinically, the findings alert clinicians to the high
depression or whether changes in global functioning rates of depression in mothers bringing their offspring
precede those of depressive symptomatology. If the for treatment of depression. Whereas many mothers
latter is true, a longer interval might lead to stronger will not accept treatment for themselves, in the context
results in offspring. of bringing their children to treatment, the limited
number who do may benefit. The research implications
regarding feasibility are less positive. Sampling de-
LIMITATIONS pressed mothers from a pediatric clinic where children
The present report is a pilot study and has major are being brought for treatment of depression does not
limitations, the most significant of which are its appear feasible. The mothers were dif ficult to recruit
uncontrolled design and small sample size. As a result, and seemed already overburdened with their children’s
the findings can suggest only associations between schedule. It was difficult to coordinate the timing of
variables and not their direction. This report cannot both treatments.
determine whether the relationship between maternal As a result of this experience, we are now studying
depression and offspring functioning is driven by depressed mothers who have come for their own
change in mother or offspring. In particular, we cannot treatment, and we are evaluating their children. This
illuminate the nature of the relationship between is an ancillary study to the multisite Sequential
change in maternal depression and offspring outcome. Treatment Alternatives to Relieve Depression Study
Furthermore, as an open trial, we did not focus on (STAR*D). STAR*D will compare the effectiveness of
efficacy of treatment but on the existence of a different treatment options for MDD and will enroll
relationship between maternal symptom reduction 4,000 adults across 13 Regional Coordinating Centers
and offspring outcome. To achieve symptom reduction [Rush et al., 2003]. Focusing on the common clinical
in mothers we employed IPT; however, it is possible question of what to do next when patients fail to
that reduction would occur more rapidly or be respond to a standard trial of treatment with an
antidepressant medication, STARnD aims to define
qualitatively different if medication and/or another

7. Brief Report: Depressed Mothers and Children 57
which subsequent treatment strategies are both accep- psychotherapy in acute depression. Arch Gen Psychiatry 36:
1450–1456.
table to patients, and provide the best clinical results.
Elkin I, Gibbons RD, Shea MT, Sotsky SM, Watkins JT, Pilkonis PA,
There are two ongoing studies of the impact of
Hedeker D. 1995. Initial severity and differential treatment
treating depressed mothers on their children: Garber et
outcome in the National Institute of Mental Health Treatment
al. (PI–Vanderbilt University) are conducting a study
of Depression Collaborative Research Program. J Consult Clin
focusing on the effect of treating parental MDD on
Psychol 63:841–847.
children’s (aged 8–16) socioemotional adjustment.
Emslie GJ, Rush AJ, Weinberg WA, Kowatch RA, Hughes CW,
Their study includes three groups receiving either 16 Carmody T, Rintelmann J. 1997. A double-blind, randomized,
weekly sessions of cognitive-behavioral therapy (CBT), placebo-controlled trial of fluoxetine in children and adolescents
pharmacotherapy, or placebo, with approximately 60 with depression. Arch Gen Psychiatry 54:1031–1037.
patients/group, as well as a comparison group of 90 Ferro T, Verdeli H, Pierre F, Weissman MM. 2000. Screening for
nondepressed mothers (patients are recruited from depression in mothers bringing their offspring for evaluation or
psychiatric settings). Riley (PI–Johns Hopkins School treatment of depression. Am J Psychiatry 157:375–379.
of Public Health; 5RO1MH5834-02) targeted 150 Hamilton M. 1960. A rating scale for depression. J Neurology
depressed women and their children [75 mothers Neurosurg Psychiatry 12:56–62.
receiving 16 weekly group CBT sessions, and 75 Hammen C, Burge D, Burney E, Adrian C. 1990. Longitudinal study
of diagnoses in children of women with unipolar and bipolar
receiving medication (Paroxetine)] and a comparison
affective disorder. Arch Gen Psychiatry 47:1112–1117.
group of 50 nondepressed mothers and their children,
John K, Gammon GD, Prusoff BA, Warner V. 1987. The social
recruited from a family-planning clinic.
adjustment scale inventory for children and adolescent (SAICA):
All three proposals using different treatments are
Testing of a new semistructured inerview. J Am Acad Child Adolesc
top-down studies sampling from the depressed
Psychiatry 26:898–911.
mothers. All have the goal of determining the effect
Johnson J, Weissman MM, Klerman GL. 1992. Service utilization
of depressed mothers’ symptom reduction on their and social morbidity associated with depressive symptoms in the
children. community. JAMA 267:1478–1483.
Katz R, Shaw BF, Vallis TM, Kaiser AS. 1995. The assessment of
Acknowledgments. This study was supported by severity and symptom patterns in depression. In: Beckham EE,
the National Institute of Mental Health (5T32MH- Leber WR, editors. Handbook of depression. Second edition.
16434 to T.F., 5P30MH43878 to M.M.W., 5T32MH1- New York: Guilford Press, p 61–85.
Kessler RC. 1994. The National Comorbidity Survey of the United
5144 to C.B.); a Samuel Priest Rose Research Award
States. Int Rev Psychiatry 6:365–376.
(to M.M.W.); the Klingenstein Third Generation
Kessler RC, McGonagle KA, Swartz M, Blazer DG, Nelson CB.
Foundation (to M.M.W.); and a NARSAD Young
1993. Sex and depression in the National Comorbidity Survey, I:
Investigator’s Award (to C.B.). We thank Florence
Lifetime prevalence, chronicity and recurrence. J Affect Disord
Pierre, M.A., Talia Zaider, and Wanda Liriano,
29:85–96.
Amarilis Lendof, C.S.W., and Kristen Kenefick,
Kovacs M. 1980. Rating scales to assess depression in school-aged
Psy.D., for help in dif ferent phases of this study. children. Acta Paedopsychiatry 46:305–315.
Leaf PJ, Livingston MM, Tischler GL, Weissman MM, Holzer CE
III, Myers JK. 1985. Contact with health professionals for the
treatment of psychiatric and emotional problems. Med Care
REFERENCES 23:1322–1337.
McGlashan T. 1973. The documentation of clinical psychotropic
American Psychiatric Association. 1994. Diagnostic and statistical
drug trials. Rockville, MD: National Institute of Mental Health.
manual of mental disorders. Fourth edition. Washington, DC:
Mufson L, Weissman MM, Moreau D, Garfinkel R. 1999. Efficacy
American Psychiatric Association.
of interpersonal psychotherapy for depressed adolescents. Arch
Beck AT, Ward CH, Mendelson M, Mock J, Erbaugh J. 1961. An
Gen Psych 56:573–579.
inventory for measuring depression. Arch Gen Psychiatry 4:5
Murray CJ, Lopez AD. 1996. Evidence-based health policyF
61–571.
lessons from the Global Burden of Disease Study. Science
Bird HR, Canino G, Rubio-Stipec M, Ribera JC. 1987. Further
274:1593–1594.
measures of the psychometric properties of the Children’s Global
Pocock SJ. 1999. Clinical trials: A practical approach. New York:
Assessment Scale. Arch Gen Psychiatry 44:821–824.
Wiley.
Brent DA, Holder D, Kolko D, Birmaher B, Baugher M, Roth C,
Regier DA, Burke JD Jr. 1987. Psychiatric disorders in the
Iyengar S, Johnson BA. 1997.
community: the ECA study. In: Hales RE, Frances AJ, editors.
A clinical psychotherapy trial for adolescent depression comparing
Psychiatry update: American Psychiatric Association Annual Re-
cognitive, family, and supportive therapy. Arch Gen Psychiatry
view, vol. 6. Washington, DC: American Psychiatric Press,
54:877–885.
p 610–624.
Broadhead WE, Blazer DG, George LK, Tse CK. 1990. Depression,
Robins LN, Helzer JE, Weissman MM, Orvaschel H, Gruenberg E,
disability days, and days lost from work in a prospective
Burke JD Jr, Regier DA. 1984. Lifetime prevalence of specific
epidemiologic survey. JAMA 264:2524–2528.
psychiatric disorders in three sites. Arch Gen Psychiatry 41:
Bryk AS, Raudenbush SW. 1992. Hierarchical linear models.
949–958.
Newbury Park, CA: Sage.
Rush AJ, Trivedi M. 2003. STAR*D Treatment Trial for Depression.
DiMascio A, Weissman MM, Prusoff BA, Neu C, Zwilling M,
Am J Psychiatry 160:2.
Klerman GL. 1979. Differential symptom reduction by drugs and

8. 58 Verdeli et al.
Shaffer D, Gould MS, Brasic J, Ambrosini P, Fisher P, Bird H, Weissman MM, Leaf PJ, Tischler GL, Blazer DG, Karno M,
Livingston B, Florio LP. 1988. Affective disorders in five United
Aluwahlia S. 1983. A children’s global assessment scale (CGAS).
States communities. Psychol Med 18:141–153.
Arch Gen Psychiatry 40:1228–1231.
Weissman MM, Bland RC, Canino GJ, Faravelli C, Greenwald S,
Spitzer RL, Williams JBW, Gibbon M, First MB. 1992.
Hwu HG, Joyce PR, Karam EG, Lee CK, Lellouch J, Lepine JP,
The Structured Clinical Interview for DSM-III-R (SCID), I:
Newman SC, Oakley-Browne MA, Rubio-Stipec M, Wells JE,
History, rationale, and description. Arch General Psychiatry Wickramaratne PJ, Wittchen HU, Yeh EK. 1996. Cross national
49:624–629. epidemiology of major depression and bipolar disorders. JAMA
Spitzer RL, Williams JBW, Kroenke K, Linzer M, de Gruy FV III, 246:293–299.
Hahn S, Brody D, Johnson JG. 1994. Utility of a new procedure Weissman MM, Warner V, Wickramaratne PJ, Moreau D, Olfson M.
1997. Offspring of depressed parents: Ten years later. Arch Gen
for diagnosing mental disorders in primary care: The PRIME-MD
Psychiatry 54:932–940.
1000 Study. JAMA 272:1749–1756.
Weissman MM, Markowitz JC, Klerman GL. 2000. Comprehensive
Spitzer RL. 1997. PRIME-MD Patient Problem Questionnaire. New
guide to interpersonal psychotherapy. New York: Basic Books.
York: New York State Psychiatric Institute, Biometrics Unit.
Wells KB, Stewart A, Hays RD, Burnam MA, Rogers W, Daniels M,
Weissman MM, Bothwell S. 1976. The assessment of social adjustment
Berry S, Greenfield S, Ware J. 1989. The functioning and well-
by patient self-report. Arch Gen Psychiatry 33:1111–1115.
being of depressed patients: Results from the Medical Outcomes
Weissman MM, Klerman GL, Prusoff BA, Sholomskas D, Padian N.
Study. JAMA 262:914–919.
1981. Depressed outpatients: Results one year after treatment with Young AS, Klap R, Sherbourne CD, Wells KB.2001. The quality of
drugs and/or interpersonal psychotherapy. Arch Gen Psychiatry care for depressive and anxiety disorders in the United States. Arch
30:51–55. Gen Psychiatry 58:55–61.