Article

Onset and Persistence of Postpartum Depression

in an Inner-City Maternal Health Clinic System

Kimberly A. Yonkers, M.D. Objective: Postpartum depressive disor- Results: The estimated rate of major de-
ders lead to maternal disability and dis- pressive disorder during the postpartum
Susan M. Ramin, M.D. turbed mother-infant relationships, but period among women in this setting was
information regarding the rates of major between 6.5% and 8.5%. Only 50% of the
depressive disorder in minority women is depressed women reported onset follow-
A. John Rush, M.D. noticeably lacking. The goal of this study ing birth. Bottle-feeding and not living
was to determine whether the risk factors with one’s spouse or significant other
Carlos A. Navarrete, M.A. for and rate of postpartum major depres- were associated with depression at the
sive disorder in a predominantly African first evaluation; persistent depressive
Thomas Carmody, Ph.D. American and Hispanic clinic population symptoms were linked with the presence
would be similar to those reported for of other young children at home. Greater
Dana March, B.A. Caucasian women. severity of depressive symptoms at first
contact predicted major depressive disor-
Method: Investigators systematically
Stephen F. Heartwell, Dr.P.H. der several weeks later.
screened all women scheduled for their
first postpartum visit on selected days at Conclusions: Rates of postpartum de-
Kenneth J. Leveno, M.D. four publicly funded inner-city commu- pression among Latina and African Ameri-
nity maternal health clinics in Dallas can postpartum women are similar to
County (N=802). A multistage screening epidemiologic rates for Caucasian postpar-
process included the Edinburgh Postnatal tum and nonpostpartum women. As previ-
Depression Scale, the Inventory of De- ously shown for Caucasian women, major
pressive Symptomatology, and the Struc- depressive disorder in many Latina and Af-
tured Clinical Interview for DSM-IV for a rican American postpartum women be-
maximum of three assessments during gins before delivery, revealing the need to
the initial 3–5-week postpartum period. screen pregnant women for depression.

(Am J Psychiatry 2001; 158:1856–1863)

T he risk of depressive disorders occurring during the

puerperium is thought to contribute to the higher rate of
identified and treated (19), which makes detection a major
public health problem. Several factors may identify
mood disorders among women than men in the United women at greater risk for depression during the postpar-
States (1–3). The combined period prevalence of postpar- tum period, including, but not limited to, a history of ma-
tum major and minor depressive disorders is estimated to jor depressive disorder, depressive symptoms during
be between 5% and 25% (2, 4–12). Although the prevalence pregnancy, obstetrical complications, unstable marital (or
rates for postpartum depression have received substantial other significant) relationships, stressful life events, low
attention (13, 14), little research has evaluated the preva- income, low educational attainment, and not breast-feed-
lence and risk in ethnic minority groups or in women of ing (4, 9, 20–27). Many of these factors also contribute to
lower socioeconomic status (1, 4, 7, 15). The only U.S. vulnerability to depression in women at times other than
study of low-income postpartum women found a point the postpartum period (28, 29).
prevalence of 23% for postpartum depression, including It is important to evaluate whether these risk factors for
7% for major depressive disorder and 16% for minor de- major depressive disorder are salient for ethnic minority
pressive disorder. These rates are about twice those pub- women, an inadequately studied group. In a busy, publicly
lished for postpartum women from a middle-class sample funded inner-city clinic, the amount of information pre-
in the United States (4% and 6% for major depressive dis- dictive of risk for mood disorders may be limited, but it
order and minor depressive disorder, respectively) (13), would be helpful to clinicians if commonly available infor-
suggesting that lower income level or minority status may mation could be used to identify postpartum women at
increase the risk of postpartum depression (1). particular risk for major depressive disorder.
An ongoing problem in depressed populations is the This study estimates the rate and persistence of depres-
low rate of detection and treatment (16–18). Less than half sion during the postpartum period in a multiethnic,
of women suffering major depressive disorder will be largely Hispanic population of lower socioeconomic sta-

1856 Am J Psychiatry 158:11, November 2001

 YONKERS, RAMIN, RUSH, ET AL.

tus culled from four publicly funded inner-city maternal participated in the third clinical interview received a $20 gift cer-
health clinics. Given earlier findings, we hypothesized that tificate to a grocery store in the area.
Participants whose scores were below both of the thresholds
the rate of depressive symptoms and the rate of major de-
for depressive symptoms during the initial assessment (time 1)
pressive disorder would be higher among Latina and Afri- were not included in the time 2 or time 3 assessments. Conse-
can American women, who are socioeconomically disad- quently, 463 women who had no substantial symptoms at the
vantaged, than the reported rates for non-Hispanic white, initial assessment concluded their participation after the first as-
middle-class women in the United States. Second, on the sessment. To affirm our assumption that women screening neg-
ative initially at about 3 weeks postpartum would not subse-
basis of commonly available medical and demographic
quently develop depression over the next month, the first 42
information as well as previous work (4, 9, 10, 26), we hy- participants who screened negative for depressive symptoms
pothesized that women living with their extended families during the initial assessment were followed for both a time 2 and
rather than with a spouse or significant other would be at a time 3 assessment. No woman in this group subsequently de-
greater risk for both depressive symptoms and the syn- veloped depressive symptoms (i.e., crossed either of the stated
thresholds).
drome of major depressive disorder. Third, we hypothe-
A concern was that depressed women would be less likely to
sized that women with more severe depressive symptoms keep their postpartum clinic appointments. Therefore, we inves-
immediately after delivery would be more likely to main- tigated a second subgroup of 50 randomly selected patients who
tain these symptoms and have a diagnosis of major de- did not go to the clinics at the appointed times. These women
pressive disorder several weeks later (10, 14, 30–33). were contacted and screened by telephone with the same instru-
ments. Analysis of this subgroup revealed no differences in initial
screening results from the study participants (data available upon
Method request).

Participants Questionnaires
Potential subjects included all women who came for their ini- Every participant completed a demographic information form
tial postpartum appointments on selected days at four inner-city that included the following: age, ethnicity, language(s), gravidity,
maternal health clinics in Dallas. Of 890 women consecutively ap- parity, date of parturition, current pregnancy status, familial envi-
proached, 802 agreed to participate; 88 (10%) of the patients de- ronment, employment status of participant and partner (if appli-
clined to participate or were unavailable. This constitutes a sys- cable), enrollment in government assistance programs, educa-
tematic sample of about 13% of postpartum patients from these tional level (1=elementary school, 2=some secondary school, 3=
four clinical sites. Subjects were eligible for participation if they high school diploma or General Equivalency Diploma, 4=some
had completed their pregnancies or miscarried. Participants had college, 5=college degree), and breast-feeding status. A brief med-
a mean age of 24.2 years (SD=5.6, range=14–48); most had some ical history form requesting information regarding current medi-
secondary school but had not completed high school; 20 (2%) cal problems, obstetric complications, current medications, and
were white non-Hispanic, 162 (20%) were African American, 604 contraceptive methods was also administered.
(75%) were Hispanic, 5 (1%) were immigrants from Asian coun- Two scales were used to evaluate depressive symptoms. The
tries, and 11 (1%) were of other ethnic backgrounds; 489 (61%) of Edinburgh Postnatal Depression Scale (3, 35) is a 10-item, multi-
the participants spoke Spanish only. Four of the women had ex- ple-choice self-report scale developed specifically for the assess-
perienced spontaneous abortions; the remainder had live births; ment of postnatal depression. This scale was previously trans-
five of the live births were twin pairs. lated into Spanish (39); we field-tested the form and made minor
modifications that would optimize its use with Latina women in
Procedure Dallas County. Scores on this scale range from 0 to 30; higher
The protocol called for assessing the participants three times scores indicate more depressive symptoms. The Inventory of De-
during the puerperium: about 3 weeks (time 1), about 4 weeks pressive Symptomatology (34) is a 30-item scale designed to as-
(time 2), and 4–5 weeks (time 3) postpartum. At the initial post- sess the severity of depression for either inpatients or outpatients.
partum visit (at about 3 weeks), patients were met by a bilingual Possible scores range from 0 to 84, and it requires approximately
(English/Spanish-speaking) master’s-level psychologist (C.A.N.), 15 minutes to complete. This form was previously translated into
who described the project and obtained written informed con- Spanish and used among Spanish-speaking patients in Dallas
sent. At this time the women completed a screening packet con- County.
sisting of demographic and medical history questionnaires and Quality of life and functional impairment were evaluated with
three self-report forms: the 30-item Inventory of Depressive the Quality of Life in Depression Scale (36, 37). Results from these
Symptomatology (34), the Edinburgh Postnatal Depression Scale questionnaires will be presented elsewhere.
(3, 35), and the Quality of Life in Depression Scale (36, 37). The Finally, the SCID (38) was used to assign psychiatric diagnoses
women were also interviewed verbally because some women at the third assessment in women who still had elevated scores on
may have been embarrassed to admit that they were unable to either the Edinburgh Postnatal Depression Scale or the Inventory
read. Each woman was given a free package of diapers for her of Depressive Symptomatology at their second assessment. Ques-
participation. tionnaires and interviews were administered either in English or
Approximately 1 week after the initial assessment (time 2), Spanish by a bilingual master’s-level clinician (C.A.N.). Both were
women whose scores exceeded either the threshold for substan- trained in the administration of the SCID and achieved interrater
tial depressive symptoms (a score of 18 on the Inventory of De- reliability greater than 0.80. These individuals were not blind to
pressive Symptomatology or a score of 12 on the Edinburgh Post- patients’ scores on the Edinburgh Postnatal Depression Scale and
natal Depression Scale) at time 1 were contacted by telephone Inventory of Depressive Symptomatology.
and the questionnaires were repeated. Women who continued to
score above either threshold at time 2 were seen about 1 week Statistical Analysis
later for a third assessment and given the Structured Clinical In- Using previously conducted validity studies, we set the thresh-
terview for DSM-IV, Clinician Version (SCID) (38). Women who old for declaring clinically significant depressive symptoms at

Am J Psychiatry 158:11, November 2001 1857

POSTPARTUM DEPRESSION

TABLE 1. Ethnicity and Other Characteristics of 802 Women at Their First Postpartum Visit to a Community Clinic
Ethnicity
Characteristic Hispanic (N=604) African American (N=162) White (N=20) Other (N=16) Total (N=802)
Mean SD Mean SD Mean SD Mean SD Mean SD

Age (years) 24.7 5.4 21.9 5.8 23.4 4.4 30.5 4.9 24.2 5.6
Parity 2.3 1.4 2.1 1.5 2.1 1.0 3.0 2.1 2.3 1.4
Time since delivery (weeks) 3.6 3.9 3.2 2.6 4.7 7.5 2.9 0.8 3.5 3.8
Number of children 2.1 1.1 2.0 1.3 2.0 1.0 2.2 1.2 2.1 1.2

N % N % N % N % N %

Breast-feeding 388 64 29 18 2 10 12 75 431 54

Employed 32 5 31 19 6 30 0 0 69 9
Spouse employed 450 75 60 37 11 55 12 75 533 66
Living with spouse 467 77 24 15 11 55 9 56 511 64
Living with family 53 9 18 11 5 25 0 0 76 9

time 1 at 12 or higher on the Edinburgh Postnatal Depression Results

Scale (35, 39) or 18 or higher on the Inventory of Depressive
Symptomatology (34). Major depressive disorder at time 3 was Depressive Symptoms at Time 1 or Time 2
defined as meeting DSM-IV criteria on the SCID. Differences in Table 1 shows the patients’ demographic characteris-
the onset of depressive symptoms among different ethnic groups
tics. Table 2 shows the percentages of women with symp-
were explored with chi-square analysis. The following variables
toms of depression at time 1 and time 2 or the diagnosis of
were used to compare depressed women (i.e., those scoring
above the two thresholds at the initial and second interviews)
major depressive disorder at time 3. Table 2 also shows
with the nondepressed women: age, breast-feeding (yes or no), how many patients met criteria for depressive symptoms
education (level 0 or 1, 2, 3, 4, or 5), race (white, African American, on the Edinburgh Postnatal Depression Scale or the Inven-
Hispanic, other), language (English or Spanish), number of preg- tory of Depressive Symptomatology.
nancies, receiving government assistance (yes or no), number of Only four of the 297 women with depressive symptoms
children, number of babies, employment (yes or no), employ- at the first visit were not available or declined to partici-
ment of spouse (yes, no, or no spouse), living at home with spouse pate in the second evaluation. Eighty-three (28%) of the
(yes or no), living at home with parents (yes or no), living at home 293 depressed women evaluated at time 2 continued to re-
with extended family (yes or no), living at home with significant
port depressive symptoms above threshold on one or both
other (yes or no). All variables that were significant at a 0.05 level
of the instruments (Table 2).
were entered into multiple regression models to test the associa-
tion between each measure and depressive symptoms. Since the At time 3, 67 (81%) of the 83 women who had depressive
two screening instruments may have performed differently, mod- symptoms at time 2 completed the SCID; 52 (78%) of these
els for women who scored higher than the threshold for either 67 women met DSM-IV criteria for major depressive disor-
were also investigated. der. However, only 26 (50%) of the 52 women reported on-
Logistic regression analysis was used to predict risk for depres- set of depressive symptoms during the postpartum pe-
sive symptoms at time 1 or persistent depressive symptoms at riod; 13 (25%) reported onset during pregnancy, and 13
time 2 according to threshold definitions on the Edinburgh Post- (25%) reported onset before pregnancy. In addition, six
natal Depression Scale or Inventory of Depressive Symptomatol- women met criteria for minor depressive disorder, and
ogy. Women whose scores were below threshold at time 1 were one was diagnosed as having bipolar disorder, depressed
designated as euthymic at time 2, since no women in our sub-
phase. Thus, the rate of major depressive disorder in the
group of women scoring below threshold at time 1 subsequently
entire group of women during the first month postpartum
had a score above threshold on either scale at time 2. Demo-
graphic data and scores at time 1 were used to develop a model was from 6.5% (if no woman lost to follow-up was de-
for predicting the full syndrome of major depressive disorder, de- pressed) to 8.5% (if all women lost to follow-up were de-
fined by the SCID at time 3. We repeated this model using time 2 pressed). For postpartum-onset major depressive disorder
scores. A model for predicting the onset of major depressive dis- the rate would be between 3.2% and 5.2%.
order in women with postpartum illness was also investigated. Fi- There were substantial differences between ethnic
nally, receiver operating characteristic curve analysis was used to groups in the rates of depressive symptoms at time 1 and
compute curves that could establish thresholds for the Edinburgh time 2 and in the rates of the syndrome of major depres-
Postnatal Depression Scale and Inventory of Depressive Symp-
sive disorder at time 3. For this analysis, only the African
tomatology for women evaluated at time 2. Again, we assumed
American (N=162) and Hispanic (N=604) groups were
that women who initially scored negative would remain negative
compared because there were too few women in the other
for the syndrome of major depressive disorder. Sensitivity, speci-
ficity, positive predictive value, and negative predictive value ethnic groups. Among African American women, 72 (45%)
were computed for the optimal thresholds. The quality index of and 24 (34%) of those evaluated had clinically significant
efficiency was determined by weighted averages of sensitivity and depressive symptoms at time 1 and time 2, respectively.
specificity estimates (40). Eleven (46%) of the 24 African American women who par-

1858 Am J Psychiatry 158:11, November 2001

 YONKERS, RAMIN, RUSH, ET AL.

TABLE 2. Depressive Symptoms and Major Depressive Disorder Among 802 Community Clinic Patients at Three Postpar-
tum Assessments
Postpartum Assessment
2: Patients With Depressive 3: Patients With Depressive
Symptoms at Assessment 1 Symptoms at Assessment 2
1: All Patients (N=802) (N=293)a (N=67)b
Measure of Depression N % N % N %
Depressive symptoms
Inventory of Depressive Symptomatologyc 24 3 3 1
Edinburgh Postnatal Depression Scaled 128 16 30 10
Both instruments 144 18 50 17
Either instrument 297 37 83 28
Diagnosis of major depressive disorder by Structured
Clinical Interview for DSM-IV, Clinician Version 52 78
a Four of the 297 women with depressive symptoms declined to participate or were not available.
b Sixteen of the 83 women with depressive symptoms did not complete the SCID because they declined to participate or were not available.
c Score ≥18.
d Score ≥12.

ticipated in SCID interviews at time 3 had major depres- pressive symptoms (odds ratio=0.60, CI=0.44–0.81, p=
sive disorder. Given that nine women were lost to follow- 0.008).
up, the overall rate of major depressive disorder in African The risk of persistent depressive symptoms exceeding
American women was between 6.8% and 12.3%. the threshold on either the Edinburgh Postnatal Depres-
The corresponding rates of depression in the Hispanic sion Scale or the Inventory of Depressive Symptomatology
women differed slightly: 214 (35%) and 55 (26%) had sig- at time 2 was lower among women having some college
nificant depressive symptoms at time 1 and time 2, re- education (odds ratio=0.22, CI=0.06–0.76, p=0.02) and
spectively. Twenty-nine (53%) of the 55 Hispanic women those living with their spouse or significant other (odds ra-
referred for interview at time 3 had major depressive dis- tio=0.49, CI=0.28–0.84, p=0.01). When women were de-
order. Sixteen Hispanic women were lost to follow-up, fined by threshold depression scores on the Edinburgh
which yields an estimate for major depressive disorder be- Postnatal Depression Scale only, women who had addi-
tween 4.8% and 7.4%. The difference between the African tional babies at home had a greater risk of having depres-
American and Hispanic groups in depressive symptoms sive symptoms (odds ratio=1.40, CI=1.06–1.77, p=0.01).
was significant at time 1 (χ2=5.4, df=1, p=0.02) but not at The first model for predicting major depressive disorder
time 2 or time 3. at time 3 used time 1 data. The predictors of major depres-
The capacity to speak English may be a rough measure sive disorder were the time 1 Edinburgh Postnatal Depres-
of acculturation. Therefore, we compared Spanish-only sion Scale score, time 1 Inventory of Depressive Symptom-
speakers (N=489) with those who also spoke English (N= atology score, and living at home with extended family. For
115) among the Hispanic women. The difference was every point increase in the time 1 score on the Edinburgh
significant at time 1 between Spanish-only-speaking and Postnatal Depression Scale, the chance of having the
bilingual women; 162 (33%) of Spanish-only-speaking syndrome of major depressive disorder at time 3 was in-
women had depressive symptoms, but 50 (43%) of the creased by 15% (95% CI=1.06–1.25, p=0.0008), and a 1-
bilingual Hispanic women were depressed (χ2=4.04, df=1, point increase in the time 1 score on the Inventory of De-
p=0.04). No differences in rates of depressive symptoms at pressive Symptomatology increased the chance of major
time 2 or in rate of major depressive disorder at time 3 depressive disorder by 9% (95% CI=1.05–1.14, p=0.0001).
were found between Hispanic women who could or could Also, living at home with extended family (rather than liv-
not speak English. ing with a spouse or significant other) resulted in a three-
fold increase in the chance of major depressive disorder at
Factors Predicting Depressive Symptoms time 3 (95% CI=1.23–7.17, p=0.02).
or Major Depressive Disorder In regard to the prediction of major depressive disorder
For the women who met either threshold definition (i.e., with postpartum onset (N=26), for every point increase in
on the Edinburgh Postnatal Depression Scale or the Inven- the time 1 score on the Edinburgh Postnatal Depression
tory of Depressive Symptomatology) at time 1, the following Scale, the chance of having postpartum-onset major de-
factors significantly predicted depressive symptoms: being pressive disorder at time 3 was increased by 15% (95% CI=
more highly educated (odds ratio=1.65, confidence interval 1.04–1.27, p=0.006). In addition, a 1-point increase in the
[CI]=1.10–2.50, p=0.01), having some secondary school ed- time 1 score on the Inventory of Depressive Symptomatol-
ucation (odds ratio=1.96, CI=1.27–3.02, p=0.002), and hav- ogy increased the chance of postpartum-onset major de-
ing graduated from high school or some college education pressive disorder by 6% (95% CI=1.02–1.11, p=0.006). Also,
(odds ratio=1.80, CI=1.01–3.20, p=0.05). On the other hand, living at home with a spouse or significant other resulted
women who were breast-feeding were less likely to have de- in a 3.3 times greater chance of postpartum-onset major

Am J Psychiatry 158:11, November 2001 1859

POSTPARTUM DEPRESSION

TABLE 3. Receiver Operating Characteristics of the Edinburgh Postnatal Depression Scale and the Inventory of Depressive
Symptomatology Among Community Clinic Patients at Their Second Postpartum Assessmenta
Threshold Score Positive Negative
for Depressive Quality Index Sensitivity Specificity Predictive Predictive
Instrument Symptoms of Efficiency Correct (%) (%) (%) Value (%) Value (%)
Edinburgh Postnatal Depression Scale 11 0.64 88 78 90 66 94
Inventory of Depressive Symptomatology 18 0.70 89 95 88 65 98
a The percent correct is the probability that the patient would have depressive disorder if the score was above threshold and would not have
depressive disorder if the score was below threshold. Sensitivity is the probability that the score would be above threshold if the patient had
depressive disorder. Specificity is the probability that the score would be below threshold if the patient did not have depressive disorder. Pos-
itive predictive value is the probability that the patient would have depressive disorder if the score was above threshold. Negative predictive
value is the probability that the patient would not have depressive disorder if the score was below threshold.

depressive disorder (95% CI=1.10–10.12, p=0.03). The pre- rates reported in the literature among predominately
dictive value of the Edinburgh Postnatal Depression Scale white, middle-class groups was not supported. The rate
was significant, even when both Edinburgh Postnatal De- for major depressive disorder with postpartum onset in
pression Scale scores and Inventory of Depressive Symp- the United States reported by O’Hara and colleagues (14,
tomatology scores were included in the model. 31) was 4%, which is similar to our rate for postpartum on-
The second model used the time 2 data. The Edinburgh set (3.2%–4.2%). Even though the study group size was
Postnatal Depression Scale was still predictive: a 1-point smaller, the rate found by O’Hara et al. is particularly
increase in score on the Edinburgh Postnatal Depression strong because they followed women longitudinally dur-
Scale increased the odds of major depressive disorder by ing pregnancy and the postpartum period (14). These
14% (95% CI=1.01–1.30, p=0.03). In addition, the Inven- rates are well below the point prevalence reported by Hob-
tory of Depressive Symptomatology was also predictive: foll and colleagues (1), but we do not know the number of
each additional point scored on this measure increased women in that report who had onset of illness after deliv-
the risk by 13% (95% CI=1.06–1.20, p=0.0001). When the ery. It is interesting to note that researchers from the
outcome was major depressive disorder with postpartum United Kingdom found rates as high as 15% for major de-
onset, only the Inventory of Depressive Symptomatology pressive disorder in postpartum women (41). Rates for
was predictive at time 2 (odds ratio=1.09; 95% CI=1.06– major depressive disorder vary among countries, however,
1.13, p=0.0001). making comparisons with other countries and cultures
The receiver operating characteristics of each scale were difficult (42).
investigated by using the time 2 data. As shown in Table 3, The finding that half of the women who were experienc-
the quality index of efficiency was highest for Inventory of ing a major depressive episode postpartum actually had
Depressive Symptomatology scores, suggesting slight su- the onset of the disorder before delivery is important and
periority for this screening scale when used at about 4 confirms work conducted in smaller groups (4, 13, 14, 24,
weeks postpartum. The optimal threshold for screening 31, 32, 43). The rates of major depressive disorder for eth-
with the Edinburgh Postnatal Depression Scale was 11; for nic minority women who are not necessarily postpartum
the Inventory of Depressive Symptomatology the optimal reported in epidemiologic studies (44) are similar to our
threshold was 18. rates for ethnic minority women with major depressive
disorder (between 4.8% and 7.4% in Hispanic women and
6.8% and 12.3% in African American women). This means
Discussion
that either the postpartum period is not a time of substan-
This study reports the point prevalence of postpartum tial increase for risk of depression (19) or that the duration
major depressive disorder among ethnic minority women, of illness episodes is brief. Health care professionals
most of whom were Hispanic, who were receiving care in a should conduct broad screening for major depressive dis-
large publicly funded maternal health system. As such, to order and include pregnant women in screening efforts.
our knowledge, it is the largest study group of women in Ethnic minority groups appear to be at high risk for major
the United States screened from such a setting and repre- depressive disorder, whether or not they are postpartum, a
sents the largest database on the prevalence and risk of fact that has not received adequate attention (44).
postpartum depression in minority women. We found Even though the rates of major depressive disorder and
that the rate of major depressive disorder during the post- minor depressive disorder during the postpartum period
partum period was between 6.5% and 8.5%. Surprisingly, in this study are no higher than for women in general, the
only 50% of the women with major depressive disorder de- rates of clinically significant depressive symptoms are im-
veloped the current episode following delivery; 25% devel- pressive. More than one-third of women endorsed clini-
oped the episode during pregnancy. cally significant mood symptoms at their 3-week postpar-
Our hypothesis that the rate of postpartum major de- tum visit, which decreased but remained high (nearly
pressive disorder would be higher in this multiethnic, so- 10%) during the second postpartum assessment. The res-
cioeconomically disadvantaged group compared with olution of normative postpartum symptoms that overlap

1860 Am J Psychiatry 158:11, November 2001

 YONKERS, RAMIN, RUSH, ET AL.

with the symptoms of depression (e.g., fatigue) in the first either view, since half of our patients developed major de-
few weeks may have contributed to the decrease from the pressive disorder before delivery.
first to the second assessment (23, 45). This may also be In the univariate analyses, we saw a protective effect
the reason that the performance on the Inventory of De- against persistent symptoms for women who were living
pressive Symptomatology, which includes questions with a spouse or significant other. Living with extended
about the physical symptoms associated with depression, family may indicate the absence of a confidant or spouse,
improved at the second screening. Alternatively, this which is consistent with findings from others on the risk of
change in rate of depressive symptoms between the first postpartum depression (10, 51). However, other explana-
and second postpartum assessment may be attributable tions may be operative because we did not specifically ask
to a resolution of biological changes that lead to mood our patients about disruptions in marital or other signifi-
symptoms. In this sense, depressive symptoms within the cant relationships; for example, the need to live with other
first few weeks may result from changing hormone levels family members may have indicated more dire financial
during the immediate postpartum period and may be circumstances.
viewed as an extended period of the “maternity blues.” Yet Several methodological limitations in this study are
another possibility is that concern and anxiety over moth- noteworthy. Although women reported an onset of de-
ering tasks contributed to depressive symptoms and that pression before parturition, all of our assessments were
this diminished over the first few weeks postpartum. conducted after delivery and are liable to retrospective re-
Since clinically significant depressive symptoms will porting bias. However, we saw women within weeks after
improve over the course of several weeks postpartum for delivery, thus minimizing the chance of inaccurate re-
many women, it would be useful to determine which ports. Further, given the valence associated with an event
women are likely to stay symptomatic so that early inter- like parturition, it is likely that women would overreport
ventions can be employed and premature interventions rather than underreport the postpartum onset. Second,
can be avoided for those with only transient symptoms. our time frame of 1 month for assessing the point preva-
Our models found that women with more severe depres- lence for depression is much briefer than that in many
sion at 3 weeks were likely to remain persistently de- reports in the literature (3–6 months). However, it is con-
pressed. Receiver operating characteristic curve analysis sistent with the DSM-IV definition and probably encom-
suggests that a score over 11 on the Edinburgh Postnatal passes women in whom the biology of parturition influ-
Depression Scale or over 18 on the Inventory of Depressive enced illness onset. Moreover, our rate is consistent with
Symptomatology at 4 weeks postpartum can be very help- rates found by others in the United States, suggesting that
ful in identifying women likely to have a syndrome of ma- the rate in minority women is reliable. Finally, we did not
jor depressive disorder. These tools can easily be imple- screen all women at these four clinics. However, we sys-
mented in busy clinical settings. tematically screened all women during selected days, and
we investigated a subgroup of women who failed to keep
Depressive symptoms at the first postpartum visit, 3
their appointments. This suggests that our findings are
weeks after delivery, were predicted by higher levels of ed-
representative of the group as a whole.
ucation, which was an unexpected finding but may also
indicate a higher degree of acculturation among Hispanic
women, who were the predominant group in our cohort. Summary
Along these lines, Hispanic women who spoke both En-
The point prevalence of major depressive disorder in
glish and Spanish were at higher risk of depressive symp-
this large group of socioeconomically disadvantaged post-
toms. On the other hand, being in the group of women
partum women was between 6.5% and 8.5%, but only half
who were breast-feeding was associated with a lower like- of the women with major depressive disorder reported an
lihood of symptoms, a finding observed by some (46, 47), onset during the immediate postpartum period. This sup-
but not others (48). There may be protective psychosocial ports the theory that major depressive disorder during
and biological factors that accompany breast-feeding. Al- pregnancy remains underdetected and undertreated (19)
ternatively, positive feelings toward the pregnancy appear and suggests that pregnancy is not protective against the
to function as a buffer against postpartum depression (12, onset or continuation of major depressive disorder.
49), and the interest and motivation to breast-feed may re- Breast-feeding was associated with a lower rate of depres-
flect such a positive attitude. sive symptoms, but this may speak to greater motivation
Having other small children at home increased the like- to breast-feed in nondepressed women. Living with a
lihood that depressive symptoms would persist into the spouse or significant other decreased the likelihood that
second postpartum assessment. Other work (50) suggests depressive symptoms would continue. On the other hand,
that having small children at home increases the risk of having other infants at home increased the risk for persis-
depression in socioeconomically disadvantaged women, tent postpartum depressive symptoms. These factors
but one study (5) failed to show this was predictive of post- should be considered in assessing risk for depression in
partum depression. Our findings are not inconsistent with puerperal women. Simple self-report instruments can

Am J Psychiatry 158:11, November 2001 1861

POSTPARTUM DEPRESSION

substantially increase the detection of these illnesses, and 14. O’Hara MW, Neunaber DJ, Zekoski EM: Prospective study of
in our study the best predictor of the persistence of a de- postpartum depression: prevalence, course, and predictive
factors. J Abnorm Psychol 1984; 93:158–171
pressive syndrome was the initial severity score on simple
15. Graham CA, Sherwin BB: A prospective treatment study of pre-
screening questionnaires. Given the high frequency of menstrual symptoms using a triphasic oral contraceptive. J Psy-
mood disorders in women, routine screening can lead to chosom Res 1992; 36:257–266
better recognition and early intervention. 16. Keller MB, Lavori PW, Klerman GL, Andreasen NC, Endicott J,
Coryell W, Fawcett J, Rice JP, Hirschfeld RM: Low levels and lack
Presented in part at the 153rd annual meeting of the American of predictors of somatotherapy and psychotherapy received by
Psychiatric Association, Chicago, May 13–18, 2000, and at the 20th depressed patients. Arch Gen Psychiatry 1986; 43:458–466
annual meeting of the Society of Maternal Fetal Medicine, Miami 17. Spitzer RL, Kroenke K, Williams JB: Validation and utility of a
Beach, Fla., Jan. 31–Feb. 4, 2000. Received Nov. 27, 2000; revision re- self-report version of PRIME-MD: the PHQ Primary Care Study.
ceived May 18, 2001; accepted June 27, 2001. From the Department JAMA 1999; 282:1737–1744
of Psychiatry and the Department of Obstetrics and Gynecology, Uni-
18. Katon W: Chronic fatigue syndrome criteria: a critique of the
versity of Texas, Southwestern Medical Center, Dallas. Address reprint
requests to Dr. Yonkers, 142 Temple St., Suite 301, New Haven, CT
requirement for multiple physical complaints. Arch Intern Med
06510; [email protected] (e-mail). 1992; 152:1604–1609
Supported in part by NIMH grants MH-01908 (Dr. Yonkers) and MH- 19. Depression in Primary Care, vol 1: Detection and Diagnosis:
53799 (Dr. Rush); by Mental Health Connections, a partnership be- Clinical Practice Guideline Number 5. Washington, DC, US De-
tween Dallas County Mental Health and Mental Retardation and the partment of Health and Human Services, Public Health Ser-
Department of Psychiatry of the University of Texas Southwestern vice, Agency for Health Care Policy and Research, 1993
Medical Center, which receives funding from the Texas State Legisla- 20. Swain AM, O’Hara MW, Starr KR, Gorman LL: A prospective
ture and the Dallas County Hospital District (Dr. Rush); and by the Sa-
study of sleep, mood, and cognitive function in postpartum
rah M. and Charles E. Seay Center for Basic and Applied Research in
and nonpostpartum women. Obstet Gynecol 1997; 90:381–
Psychiatry (Dr. Rush).
486
21. O’Hara MW, Schlechte JA, Lewis DA, Varner MW: Controlled pro-
spective study of postpartum mood disorders: psychological,
References environmental, and hormonal variables. J Abnorm Psychol
1991; 100:63–73
1. Hobfoll SE, Ritter C, Lavin J, Hulsizer MR, Cameron RP: Depres-
sion prevalence and incidence among inner-city pregnant and 22. Pitt B: “Atypical” depression following childbirth. Br J Psychia-
postpartum women. J Consult Clin Psychol 1995; 63:445–453 try 1968; 114:1325–1335
2. Llewellyn AM, Stowe ZN, Nemeroff CB: Depression during preg- 23. Campbell SB, Cohn JF: Prevalence and correlates of postpar-
nancy and the puerperium. J Clin Psychiatry 1997; 58:26–32 tum depression in first-time mothers. J Abnorm Psychol 1991;
100:594–599
3. Warner R, Appleby L, Whitton A, Faragher B: Demographic and
24. Watson JP, Elliott SA, Rugg AJ, Brough DI: Psychiatric disorder in
obstetric risk factors for postnatal psychiatric morbidity. Br J
pregnancy and the first postnatal year. Br J Psychiatry 1984;
Psychiatry 1996; 168:607–611
144:453–462
4. Demyttenaere K, Lenaerts H, Nijs P, Van Assche FA: Individual
25. Cox JL, Connor Y, Kendell RE: Prospective study of the psychiat-
coping style and psychological attitudes during pregnancy pre-
ric disorders of childbirth. Br J Psychiatry 1982; 140:111–117
dict depression levels during pregnancy and during postpar-
tum. Acta Psychiatr Scand 1995; 91:95–102 26. Paykel ES, Emms EM, Fletcher J: Life events and social support
in puerperal depression. Br J Psychiatry 1980; 136:339–346
5. Gotlib IH, Whiffen VE, Mount JH, Milne K, Cordy NI: Prevalence
27. Cutrona CE: Social support and stress in the transition to par-
rates and demographic characteristics associated with depres-
enthood. J Abnorm Psychol 1984; 93:378–390
sion in pregnancy and the postpartum. J Consult Clin Psychol
1989; 57:269–274 28. Costello EJ: Married with children: predictors of mental and
physical health in middle-aged women. Psychiatry 1991; 54:
6. Harding JJ: Postpartum psychiatric disorders: a review. Compr
292–305
Psychiatry 1989; 30:109–112
29. Sargeant JK, Bruce ML, Florio LP, Weissman MM: Factors associ-
7. Leopold KA, Zoschnick LB: Postpartum depression. The Female
ated with 1-year outcome of major depression in the commu-
Patient 1997; 22:40–49
nity. Arch Gen Psychiatry 1990; 47:519–526
8. McGill H, Burrows VL, Holland LA, Langer HJ, Sweet MA: Postna-
30. Kendell RE, McGuire RJ, Connor Y, Cox JL: Mood changes in the
tal depression: a Christchurch study. NZ Med J 1995; 108:162–
first three weeks after childbirth. J Affect Disord 1981; 3:317–
165
326
9. O’Hara MW: Social support, life events, and depression during 31. O’Hara MW, Zekoski EM, Philipps LH, Wright EJ: Controlled pro-
pregnancy and the puerperium. Arch Gen Psychiatry 1986; 43: spective study of postpartum mood disorders: comparison of
569–573 childbearing and nonchildbearing women. J Abnorm Psychol
10. O’Hara MW, Swain AM: Rates and risk of postpartum depres- 1990; 99:3–15
sion—a meta-analysis. Int Rev Psychiatry 1996; 8:37–54 32. Cutrona C, Troutman BR: Social support, infant temperament,
11. Pop VJ, Essed GG, de Geus CA, van Son MM, Komproe IH: Prev- and parenting efficacy: a mediational model of postpartum
alence of post partum depression—or is it post-puerperium depression. Child Dev 1986; 57:1507–1518
depression? Acta Obstet Gynecol Scand 1993; 72:354–358 33. Affonso DD, Mayberry LJ, Lovett SM, Paul S: Cognitive adapta-
12. Whiffen VE: Vulnerability of postpartum depression: a prospec- tion to stressful events during pregnancy and postpartum: de-
tive multivariate study. J Abnorm Psychol 1988; 97:467–474 velopment and testing of the CASE instrument. Nurs Res 1994;
13. O’Hara M, Schlechte JA, Lewis DA, Varner MW: Controlled pro- 43:338–343
spective study of postpartum mood disorders: psychological, 34. Rush AJ, Gullion CM, Basco MR, Jarrett RB, Trivedi MH: The In-
environmental, and hormonal variables. J Abnorm Psychol ventory of Depressive Symptomatology (IDS): psychometric
1991; 100:63–73 properties. Psychol Med 1996; 26:477–486

1862 Am J Psychiatry 158:11, November 2001

 YONKERS, RAMIN, RUSH, ET AL.

35. Cox JL, Holden JM, Sagovsky R: Detection of postnatal depres- 43. Gotlib IH, Whiffen VE, Wallace PM, Mount JH: Prospective inves-
sion: development of the 10-item Edinburgh Postnatal Depres- tigation of postpartum depression: factors involved in onset
sion Scale. Br J Psychiatry 1987; 150:782–786 and recovery. J Abnorm Psychol 1991; 100:122–132
36. Hunt SM, McKenna S: The QLDS: a scale for the measurement 44. Blazer DG, Kessler RC, McGonagle KA, Swartz MS: The preva-
of quality of life in depression. Health Policy 1992; 22:307–319 lence and distribution of major depression in a national com-
37. McKenna SP, Hunt SM: A new measure of quality of life in de- munity sample: the National Comorbidity Survey. Am J Psychi-
pression: testing the reliability and construct validity of the atry 1994; 151:979–986
QLDS. Health Policy 1992; 22:321–330 45. Klein MH, Essex MJ: Pregnant or depressed? the effect of over-
38. First MB, Spitzer RL, Gibbon M, Williams JBW: Structured Clini- lap between symptoms of depression and somatic complaints
cal Interview for DSM-IV Axis I Disorders (SCID), Clinician Ver- of pregnancy on rates of major depression in the second tri-
sion. Washington, DC, American Psychiatric Press, 1996 mester. Depression 1995; 2:308–314
39. Cox J, Holden J (eds): Perinatal Psychiatry: Use and Misuse of 46. Hannah P, Adams D, Lee A, Glover V, Sandler M: Links between
the Edinburgh Postnatal Depression Scale. London, Gaskell early post-partum mood and post-natal depression. Br J Psy-
(Royal College of Psychiatrists), 1994 chiatry 1992; 160:777–780
40. Kraemer HC: Reporting the size of effects in research studies to 47. Lane A, Keville R, Morris M, Kinsella A, Turner M, Barry S: Post-
facilitate assessment of practical or clinical significance. Psy- natal depression and elation among mothers and their part-
choneuroendocrinology 1992; 17:527–536 ners: prevalence and predictors. Br J Psychiatry 1997; 171:
41. Cooper PJ, Murray L: The impact of psychological treatments of 550–555
postpartum depression on maternal mood and infant devel- 48. O’Neill T, Murphy P, Green VT: Postnatal depression—aetiologi-
opment, in Postpartum Depression and Child Development. cal factors. Irish Med J 1990; 83:17–18
Edited by Murray L, Cooper PJ. New York, Guilford, 1997, pp 49. Kumar R, Robson KM: A prospective study of emotional disor-
201–220 ders in childbearing women. Br J Psychiatry 1984; 144:35–47
42. Weissman MM, Bland RC, Canino GJ, Faravelli C, Greenwald S, 50. Brown GW, Adler Z, Bifulco A: Life events, difficulties and recov-
Hwu H-G, Joyce PR, Karam EG, Lee C-K, Lellouch J, Lepine J-P, ery from chronic depression. Br J Psychiatry 1988; 152:487–
Newman SC, Rubio-Stipec M, Wells JE, Wickramaratne PJ, Wit- 498
tchen H-U, Yeh E-K: Cross-national epidemiology of major de- 51. Beck CT: A meta-analysis of predictors of postpartum depres-
pression and bipolar disorder. JAMA 1996; 276:293–299 sion. Nursing Res 1996; 45:297–303

Am J Psychiatry 158:11, November 2001 1863