Journal of Affective Disorders 213 (2017) 131–137

Contents lists available at ScienceDirect

Journal of Affective Disorders

journal homepage: www.elsevier.com/locate/jad

Research paper

Identifying women at risk for sustained postpartum anxiety MARK

a,b,c,⁎ d,e,f a a
Cindy-Lee Dennis , Hilary K. Brown , Kobra Falah-Hassani , Flavia Casasanta Marini ,
Simone N. Vigodb,f
a
Lawrence S. Bloomberg Faculty of Nursing, University of Toronto, Toronto, Ontario, Canada
b
Department of Psychiatry, University of Toronto, Toronto, Ontario, Canada
c
St. Michael's Hospital, Toronto, Ontario, Canada
d
Department of Anthropology (Health Studies Program), University of Toronto Scarborough, Toronto, Ontario, Canada
e
Dalla Lana School of Public Health, University of Toronto, Toronto, Ontario, Canada
f
Women's College Research Institute, Women's College Hospital, Toronto, Ontario, Canada

A R T I C L E I N F O A BS T RAC T

Keywords: Introduction: To describe the prevalence of sustained postpartum anxiety and to develop a multifactorial
Anxiety predictive model to assist in targeted screening procedures.
Postpartum Methods: In a population-based cohort in a health region near Vancouver, Canada, 522 mothers completed a
Risk factors mailed questionnaire at 1, 4, and 8 weeks postpartum measuring socio-demographic, biological, pregnancy-
related, life stressor, social support, obstetric, and maternal adjustment factors. We undertook a sequential
logistic regression analysis to develop a multifactorial predictive model of sustained postpartum anxiety, as
measured by a State Trait Anxiety Inventory (STAI) score > 40 at 1 week and/or 4 weeks, and 8 weeks
postpartum.
Results: The prevalence of sustained postpartum anxiety was 12.6% (95% CI 9.6–16.2). In the multivariable
model, predictors of sustained anxiety in the postpartum period were perceived stress at 1 week (1 SD increase;
aOR 3.74, 95% CI 2.17-6.44) and partner social support at 1 week (1 SD increase; aOR 0.59, 95% CI 0.40–0.85).
Depression symptomatology at 1 week, child care stress, and maternal self-esteem were non-significant.
Limitations: Single women and women from ethnic minority backgrounds were underrepresented in the
sample.
Conclusions: A large proportion of women experience sustained postpartum anxiety. High perceived stress and
low partner social support can be used to facilitate early identification of women likely to experience persistent
anxiety in the postpartum period and suggest the need for urgent access to psychotherapeutic services for these
women. These factors may also be potential targets for individual or couples therapy to treat postpartum
anxiety.

1. Introduction
The prevalence, risk factors for, and sequelae of postpartum depression
are well-established, and recommendations by professional organizations
exist for screening (American College of Obstetricians and Gynecologists,
2010). Less clinical and research attention has been paid to postpartum
anxiety. This is an important omission given that anxiety disorders are
common (Kessler et al., 2010) and often co-occur with other mental health
issues such as depression (Reck et al., 2008). Women with postpartum
anxiety experience persistent and excessive worry, fear, and tension, as well
as difficulty concentrating. Some women experience severe and recurrent
intrusive thoughts or images as well as panic—often about their child being
harmed. There is growing evidence that postpartum anxiety is associated
with negative outcomes for both the mother and her child. Women with
postpartum anxiety often experience low levels of self-confidence (Reck
et al., 2012; Wenzel et al., 2005a, 2005b) and increased fatigue (Taylor and
Johnson, 2013). Excessive anxiety can also lead to impaired maternal-
infant interactions (Arteche et al., 2011; Feldman et al., 2009; Tietz et al.,
2014) which, among infants, is linked with disturbed sleep (Warren et al.,
2006), excessive crying (Petzoldt et al., 2014), poor social engagement
(Feldman et al., 2009), distress to novelty (Reck et al., 2013), internalizing
difficulties (Barker et al., 2011), and poor cognitive (Keim et al., 2011) and
motor development (Pinheiro et al., 2014).
Postpartum anxiety is widespread, with prevalence estimates ran-
ging from 2% to 45% in the first year postpartum, depending on the
study's definition of anxiety, timing of assessment, and sample size

⁎
Corresponding author at: Lawrence S. Bloomberg Faculty of Nursing, University of Toronto, 155 College Street, Toronto, Ontario, Canada M5T 1P8.
E-mail address: cindylee.dennis@utoronto.ca (C.-L. Dennis).

http://dx.doi.org/10.1016/j.jad.2017.02.013
Received 2 November 2016; Received in revised form 19 January 2017; Accepted 13 February 2017
Available online 14 February 2017
0165-0327/ © 2017 Elsevier B.V. All rights reserved.
C.-L. Dennis et al. Journal of Affective Disorders 213 (2017) 131–137

Table 1
Factors assessed for association with anxiety symptomatology at 8 weeks postpartum.

Weeks

Domain Risk factor Measure 1 4 8

Socio-demographic Age Maternal age X

Parity Primiparous vs. multiparous ( > 1 child) X
Ethnicity Caucasian vs. other X
Born in Canada Yes vs. no X
Immigrant status Immigrated in past 5 years vs. other X
Marital status Married/common-law vs. single X
Education Postsecondary vs. high school or less X
Income Increments of $20,000 to ≥$80,000 CAD X
Ability to manage on current household Easy, not too bad, sometimes difficult, always difficult X
income
Access to transportation Yes, sometimes, no X
Suitable housing Yes vs. no X
Biological Maternal psychiatric history Yes vs. no: depression, bipolar disorder, anxiety disorders and schizophrenia X
Family psychiatric history Yes vs. no: depression, bipolar disorder, anxiety disorders, and schizophrenia X
History of depression Yes vs. no X
Vulnerable personality Total score: Vulnerable Personality Scale (Boyce et al., 2001) X
Premenstrual symptoms Total number of symptoms: bloating, breast tenderness, irritability, difficulty concentrating, X
anxiety, feeling sad, tiredness
Pregnancy Infertility problems Yes vs. no X
Planned pregnancy Yes, not exactly at this time, no definitely not X
Maternal feelings about pregnancy Very happy, happy, happy in some ways but not in others, unhappy, very unhappy X
Paternal feelings about pregnancy Very happy, happy, happy in some ways but not in others, unhappy, very unhappy X
Coping with pregnancy Yes vs. no X
Exercise in pregnancy Once per week vs. less than once per week X
Pregnancy complications Yes vs. no: threatened miscarriage, preterm labour, excessive nausea, excessive vomiting, X
urinary tract infection, preeclampsia, diabetes
Life stressors Life events Total score: Tennant and Andrews Life Events Scale (Tennant and Andrews, 1976) X
Psychosocial risk score Score on the ALPHA form: substance use/family violence items (Reid et al., 1998) X
Job stress Yes, all of the time, sometimes, no, not at all (women not employed outside the home were X
coded as “not at all”)
Worrying about returning to work Yes, sometimes, no (women not employed outside the home were coded as “no”) X
Satisfaction with job Very satisfied, satisfied, ok, unsatisfied, very unsatisfied X
Childcare stress Total score: Childcare Stress Checklist (Dennis, 2003) X X X
Social support Global support Total score: Social Provisions Scale (Cutrona and Russell, 1987) X X X
Relationship-specific support Total score: Social Provisions Checklist for partner, mother, mother-in-law, women friends X X X
with children (Davis et al., 1998)
Marital status Married vs. divorced/never married X
Relationship with parents Close vs. not close/no relationship: each parent X
Obstetric Induction of labour Yes vs. no X
Mode of delivery Vaginal vs. caesarean section X
Satisfied with pain management Very satisfied, satisfied, okay, unsatisfied, very unsatisfied X
Control during labour Total score: Labour Agentry Scale (Hodnett and Simmons-Tropea, 1987) X
Maternal adjustment Ready for hospital discharge Yes vs. no X
Infant feeding method Exclusive breast-feeding, almost exclusive breast-feeding, high breast-feeding, partial breast- X X X
feeding, token breast-feeding, bottle-feeding
Breastfeeding self-efficacy Total score: Breastfeeding Self-Efficacy Scale (Dennis and Faux, 1999) X
Self-esteem Total score: Rosenberg's Self-esteem Scale (Rosenberg, 1965) X
Anxiety Total score: State Trait Anxiety Inventory (Spielberger, 2010) X X X
Depression at Total score: Edinburgh Postnatal Depression Scale (Cox et al., 1987) X X X

(Enatescu et al., 2014; Martini et al., 2013). In a systematic review and
meta-analysis that included over 100 studies from over 30 different
countries, we reported that the prevalence of postpartum anxiety
decreases from approximately 18% at 1–4 weeks postpartum to 15%
at 5–12 weeks postpartum (Dennis, In press). In most studies, only one
administration of a self-reported or clinical anxiety measure is used to
identify the prevalence of postpartum anxiety. While anxiety disorders
are treatable (Craske and Stein, 2016), many new mothers experience
transient anxiety as a result of the normal stresses of childbirth and the
adjustment to parenthood. Many of these mother-infant dyads are not
at increased risk for negative outcomes and do not require formal
psychotherapeutic or psychopharmacological treatment (Dennis et al.,
2013; Matthey and Ross-Hamid, 2012). Increased knowledge about the
prevalence of, and factors associated with, sustained (i.e., ongoing)
anxiety in the postpartum period will inform treatment and secondary
prevention planning (Britton, 2008).
Risk factors for postpartum anxiety identified at a single time point
include previous diagnosis of anxiety or depression (Britton, 2008;
Engleet et al., 1990; Grant et al., 2008; Reck et al., 2008), reproductive
history (Giannandrea et al., 2013), socio-demographic variables
(Britton, 2008; Martini et al., 2015; Wenzel et al., 2005b), unwanted
pregnancy (Engle et al., 1990), prenatal care dissatisfaction (Barnett
and Parker, 1986), pregnancy and delivery complications (Barnett and
Parker, 1986; Engle et al., 1990), maternity blues (Reck et al., 2009),
poor social support (Martini et al., 2015), stress (Britton, 2008), low
self-efficacy (Martini et al., 2015), and bottle-feeding (Barnett and
Parker, 1986; Wenzel et al., 2005b). To our knowledge, no previous
studies have examined risk factors for sustained postpartum anxiety.
1.1. Aim of the study
In a large cohort sampled from a health region in British Columbia,
Canada, our aims were to identify the prevalence of sustained
postpartum anxiety symptomatology and to develop a multifactorial
predictive model of sustained postpartum anxiety symptomatology.

132
C.-L. Dennis et al.
2. Methods
2.1. Sample
We undertook a secondary analysis of a prospective cohort study
originally conducted in 2001–2002 in a health region near Vancouver,
British Columbia (Dennis et al., 2004). Eligible women were ≥18 years
of age and could understand English. Study packages, including
informed consent procedures, were provided to participating family
physician, obstetrician, and midwifery offices after receiving study
authorization from the health region and university research ethics
board. Women were recruited at > 32 weeks’ gestation or postnatally
during the usual 48-h post-hospital discharge call provided to new
mothers by public health nurses. They were informed that the
investigators were interested in learning more about the transition to
parenthood in order to better understand how to improve health
services for new mothers. Women recruited during either time period
were mailed questionnaires at 1, 4, and 8 weeks postpartum which
included researcher-addressed, stamped return envelopes. Those who
did not return their questionnaires within 2 weeks of mailing received a
reminder telephone call. All questions, including those about the
partner or family, were completed by the mother.
2.2. Measures
We used the State Trait Anxiety Inventory (STAI) (Spielberger,
2010) to measure anxiety symptomatology at 1, 4, and 8 weeks
postpartum. The STAI is comprised of 20 self-reported questions rated
on a 4-point Likert scale. Scores range from 20 to 80, and higher scores
indicate increased state anxiety. We defined sustained anxiety as
having anxiety symptomatology (i.e., STAI score of > 40) at 1 week
and/or 4 weeks, and 8 weeks postpartum. The STAI has been shown to
perform well in postpartum samples. A score > 40 on the STAI at 1
week postpartum accurately predicts 75% of women who have major
anxiety at 4 and 8 weeks postpartum (Dennis et al., 2013).
To identify risk factors associated with sustained postpartum
anxiety, we conducted a literature search using Medline, PubMed,
CINAHL, PsycINFO, and EMBASE. Potential risk factors were con-
ceptualized in the following domains: socio-demographic, biological,
pregnancy-related, life stressors, social support, obstetric, and mater-
nal adjustment (Table 1). Several of these risk factors used standar-
dized measures. We used the Vulnerable Personality Style
Questionnaire (Boyce et al., 2001) to assess maternal personality traits.
This scale includes 9 items that are rated on a 5-point Likert scale to
generate a total score ranging from 0 to 45; higher scores indicate a
more vulnerable personality style. Although this scale contains two
subscales, we used the total score, which has also been validated and is
correlated with theoretically related constructs as well as postpartum
depression (Dennis and Boyce, 2004). To assess stressors in the past 12
months, we used a 25-item life events checklist (Johnstone et al., 2001)
based on the Tennant and Andrews Life Events Scale (Tennant and
Andrews, 1976). Yes/no responses were summed to produce a total
score ranging from 0 to 25; higher scores indicate a higher number of
stressful events. We used 14 questions related to family violence and
substance use (Wilson et al., 1996) from the Antenatal Psychosocial
Health Assessment (ALPHA) form (Reid et al., 1998) to assess
psychosocial risk. Yes/no responses were summed to produce a total
score ranging from 0 to 14; higher scores indicate higher psychosocial
risk. We measured child care stress using the Childcare Stress Checklist
(Dennis, 2003), a 20-item self-report instrument developed to assess
stress related to the birth of a new baby. Yes/no responses were
summed to produce a total score of 0–20; higher scores indicate higher
levels of childcare stress. To measure maternal perceptions of global
support, we used the Social Provisions Scale (Cutrona and Russell,
1987). This instrument consists of 24 items in which women are asked
to report on their current relationships with their partner, family,
133
Journal of Affective Disorders 213 (2017) 131–137
friends, co-workers, and community members. Items were rated on a
4-point Likert scale to generate a total score of 24–96; higher scores
indicate higher perceived global support. Similarly, we used the 30-
item Social Provisions Checklist to examine relationship-specific social
support (Davis et al., 1998) for the partner, mother, mother-in-law,
and other women with children. Items were rated on a 5-point Likert
scale to generate a total score ranging from 30 to 150; higher scores
indicate higher perceived support. Breastfeeding self-efficacy was
measure using the 33-item Breastfeeding Self-Efficacy Scale (Dennis
and Faux, 1999). Items were rated on a 5-point Likert Scale to produce
a total score of 33–165; higher scores indicate higher breastfeeding
self-efficacy. To assess self-esteem, we used Rosenberg's Self-Esteem
Scale (Rosenberg, 1965). This scale has 10 items which are rated on a
4-point Likert scale to generate a total score ranging from 10 to 40;
higher scores indicate higher levels of self-esteem. Finally, to examine
depression symptomatology, we used the Edinburgh Postnatal
Depression Scale (Cox et al., 1987). This scale consists of 10 items
which are rated on a 4-point Likert scale to produce a total score
ranging from 0 to 30; higher scores indicate lower maternal mood.
2.3. Statistical analysis
Means and standard deviations (SD) as well as frequencies and
percentages were calculated to describe baseline characteristics of the
sample. We undertook a sequential logistic regression analysis using
SPSS to predict sustained anxiety symptomatology as measured by a
STAI score of > 40 at 1 week and/or 4 weeks, and 8 weeks postpartum.
In the univariable analysis, variables were individually tested in the
model; the corresponding odds ratios (OR) and 95% confidence
intervals (CI) were determined. In the multivariable analyses, we tested
statistically significant variables and retained them in the model if the
p-value for the beta-estimate was < 0.05 as calculated by the Wald
statistic. We divided continuous variables by their standard deviation
to obtain an OR for a 1 SD increase in the predictors; this allowed us to
show the relative importance of continuous predictors in a binary
logistic regression. We entered variables into the model in chronolo-
gical order: socio-demographic, biological, pregnancy-related, life
stressors, social support, obstetric, and maternal adjustment. We
created two models: model 1 consisted of statistically significant
predictors as described above, and model 2 consisted of these
predictors plus depression symptomatology at 1 week postpartum.
We evaluated goodness of fit for the models using the Chi-square
statistic determined by the log-likelihood test. We used the formula by
Tabachnick and Fidell (2001) to calculate the needed sample size. The
sample size (N) required to evaluate (m) independent predictors was
N=104+m, assuming a Type I error of 0.05 and a Type II error of 0.20.
With 43 independent predictors in our study, we had more than the
required size (104+43=147). The formula used for the multivariable
analysis was: N=50+8m. Because we did not use more than 11
predictors in a multivariable analysis at one time, we again had a
sufficient sample size.
3. Results
3.1. Sample characteristics
We recruited one hundred and sixty-six participants antenatally;
115 (69%) returned the 1-week postpartum questionnaire. We identi-
fied 857 eligible women postnatally; 667 (78%) agreed to participate in
the study and 479 (72%) returned the 1-week postpartum question-
naire. The total sample size at 1 week was 594; 535 (90%) of these
women returned the 4-week and 498 (84%) returned the 8-week
questionnaire (Fig. 1). We included all women who returned the 1-
week questionnaire and who answered questions about anxiety symp-
tomatology (N=522).
Participants had a mean age of 28.6 years (SD=5.2, range 18–44
C.-L. Dennis et al.
Fig. 1. Participant recruitment and response.
years). Most women were married or living common-law (88.9%,
n=464) and were Caucasian (90%, n=472). Fourteen percent (n=73)
of women had very low incomes (less than $20,000 CAD), 23% (n=121)
had low incomes (between $20,000 and $39,999 CAD), and 54.8%
(n=286) did not have low incomes ($40,000 CAD or more). Thirty-five
percent (n=183) reported having a high school diploma or less, 38%
(n=197) had a college diploma, and 21% (n=111) had a university
degree. Forty-four percent (n=230) of the women were primiparous,
and most (74%, n=388) delivered vaginally. The prevalence of sus-
tained postpartum anxiety (i.e., STAI score of > 40 at 1 week and/or 4
weeks and 8 weeks postpartum) was 12.6% (95% CI 9.6–16.2).
3.2. Univariable and multivariable analyses
Variables tested for their association with sustained postpartum
anxiety are listed in Table 1. We performed sequential logistic
regression to determine predictors of sustained postpartum anxiety.
Potential predictors were identified in univariable analyses; statistically
significant factors are shown in Table 2. Among socio-demographic
variables, immigration within the past 5 years, ability to manage on
current household income, and housing were significantly associated
with sustained postpartum anxiety. Biological variables were number
of medical problems, history of psychiatric problems before childbirth,
and vulnerable personality. Pregnancy-related variables were mother's
assessment of her own and her partner's coping with pregnancy. Life
stressors were number of life stressors; conflict with partner, mother,
or other mothers with children; total psychosocial risk score; psycho-
social risk scores for partner or family; boss/work environment support
for pregnancy; worry about returning to work after pregnancy; child
care stress at 4 weeks; and perceived stress at 1 week. Social support
variables were global social support; social support from the partner,
mother-in-law, or other women with children at 1 week; and social
support from the partner at 4 weeks. Maternal adjustment variables
were readiness for discharge from hospital, breastfeeding self-efficacy
at 1 and 4 weeks, and self-esteem at 4 weeks. Depression symptoma-
tology at 1 week was also associated with sustained anxiety.
The final models, consisting of 4 and 5 variables, respectively, are
presented in Table 3. In model 1, perceived stress at 1 week (1 SD
increase; aOR 4.71, 95% CI 2.43–6.84), child care stress at 4 weeks (1
SD increase; aOR 1.60, 95% CI 1.06–2.41), partner social support at 1
week (1 SD increase; aOR 0.59, 95% CI 0.41–0.86), and self-esteem at
134
Journal of Affective Disorders 213 (2017) 131–137
4 weeks (1 SD increase; aOR 0.58, 95% CI 0.35–0.97) were signifi-
cantly associated with sustained postpartum anxiety. These variables
explained 42% of the variance in the outcome. In model 2, depression
symptomatology at 1 week was not an independent predictor of
sustained anxiety (aOR 1.57, 95% CI 0.62–3.94). Significant predictors
were perceived stress at 1 week (1 SD increase; aOR 3.74, 95% CI
2.17–6.44) and partner social support at 1 week (1 SD increase; aOR
0.59, 95% CI 0.40–0.85). These variables explained 45% of the
variance in the outcome. Child care stress and self-esteem at 4 weeks
were no longer statistically significant after adding depression sympto-
matology to the model.
4. Discussion
4.1. Summary of findings
In this large Canadian cohort, we showed that approximately 1 in 8
women experience sustained anxiety during the postpartum period.
Ours is the first study to develop a multifactorial predictive model of
sustained postpartum anxiety. High perceived stress and low partner
social support were significant risk factors, with childcare stress and
maternal self-esteem as other factors to consider when assessing
women. These variables together suggest a subgroup of vulnerable
women and provide clues about the types of interventions that may be
helpful in preventing anxiety.
4.2. Comparison to previous studies
The results of our study add to previous research examining the
persistence of anxiety across the postpartum period. In our recent
meta-analysis, we showed that the prevalence of anxiety symptomatol-
ogy was 17.8% (95% CI 14.2–21.4, 14 studies, N=10,928) at 1–4 weeks
postpartum and decreased to 14.9% (95% CI 12.3–17.5, 22 studies,
N=19,158) at 5–12 weeks postpartum and stabilized over the first year
postpartum (Dennis, In press). These estimates were based on studies
that examined anxiety symptomatology in one but not subsequent
periods. Previous research has shown that among women with anxiety
symptomatology at their first prenatal appointment (12.8%), approxi-
mately half continue to have high anxiety two weeks later (6.7%)
(Matthey and Ross-Hamid, 2012). To our knowledge, no previous
studies have examined the prevalence of sustained anxiety symptoma-
tology in the postpartum period. The significant proportion of women
experiencing ongoing postpartum anxiety in our study demonstrates
the importance of generating a predictive model.
In our study, high perceived stress and low partner social support
were significantly associated with sustained postpartum anxiety. These
findings are highly consistent with previous research examining pre-
dictors of anxiety measured at one time point. For example, in a cohort
of 296 U.S. women recruited from a university medical centre, anxiety
symptomatology at 4 weeks postpartum was predicted by perceived
stress as well as maternal education and history of anxiety and
depression (Britton, 2008). Similarly, in a cohort of 306 women
recruited from gynecologic outpatient settings in Germany, anxiety
“after delivery” was predicted by low social support as well as partner-
ship satisfaction, low education, low self-esteem, and previous diag-
nosis of anxiety or depression (Martini et al., 2015). In our study,
maternal self-esteem was also statistically significant before adding
depressive symptomatology at 1 week to the model. This is consistent
with the German study and with research conducted in the general
population (Liu et al., 2014). In contrast with other studies, prior
history of anxiety or depression, sociodemographic factors such as
young maternal age, unplanned or unwanted pregnancy, and preg-
nancy and delivery complications did not emerge as factors that
predicted sustained postpartum anxiety (Barnett and Parker, 1986;
Britton, 2008; Engle et al., 1990; Martini et al., 2015; Wenzel et al.,
C.-L. Dennis et al. Journal of Affective Disorders 213 (2017) 131–137

Table 2
Factors associated with sustained postpartum anxiety: univariable analysis.

Domain Risk factor Odds ratio 95% confidence interval p-value

Socio-demographic Immigration within past 5 years 9.18 2.69–31.29 < 0.001

Difficult managing on current household income 1.87 1.04–3.36 0.036
Housing not suitable 2.46 1.09–5.56 0.031
Biological Number of medical problems 1.40 1.08–1.81 0.011
History of psychiatric problems before childbirth 4.11 2.16–7.79 < 0.001
Vulnerable personality 3.34 2.33–4.77 < 0.001
Pregnancy-related Mother not coping well with pregnancy 4.82 2.64–8.82 < 0.001
Partner not coping well with pregnancy 3.42 1.87–6.26 < 0.001
Life stressors Number of life events 1.53 1.19–1.97 0.001
Conflict with partner 0.45 0.33–0.60 < 0.001
Conflict with mother 0.66 0.49–0.87 0.003
Conflict with mother-in-law 0.78 0.58–1.06 0.116
Conflict with other mothers with children 0.66 0.50–0.87 0.003
Psychosocial risk score, total 1.47 1.12–1.92 0.005
Psychosocial risk score, partner 1.53 1.20–1.97 0.001
Psychosocial risk score, family 1.51 1.14–1.98 0.004
Boss/work environment not supportive of pregnancy 3.16 1.50–6.63 0.002
Worried about returning to work after pregnancy
Sometimes 4.85 1.36–17.28 0.015
Yes 12.52 3.14–49.95 < 0.001
No plan to return 4.92 1.17–20.67 0.029
Child care stress at 4 weeks 2.67 1.98–3.60 < 0.001
Perceived stress at 1 week 5.09 3.33–7.77 < 0.001
Social support Global social support at 1 week 0.39 0.29–0.52 < 0.001
Partner social support at 1 week 0.37 0.27–0.49 < 0.001
Mother-in-law social support at 1 week 0.60 0.44–0.83 0.002
Other women with children social support at 1 week 0.62 0.45–0.85 0.003
Partner social support at 4 weeks 0.36 0.26–0.49 < 0.001
Maternal adjustment Ready to discharge from hospital 3.33 1.30–8.55 0.012
Breastfeeding self-efficacy at 1 week 0.54 0.41–0.72 < 0.001
Breastfeeding self-efficacy at 4 weeks 0.49 0.36–0.68 < 0.001
Self-esteem at 4 weeks 0.25 0.17–0.38 < 0.001
Psychiatric Depression symptomatology at 1 week 10.30 5.18–20.51 < 0.001
Anxiety symptomatology at 1 week 40.55 17.90–91.87 < 0.001

Note: Continuous predictors expressed as OR for 1 SD increase in predictor.

2005a, 2005b). This may be a function of the fact that our study is the
first to investigate risk factors for anxiety at multiple instead of only at
one potentially transient timepoint.
4.3. Explanations for findings
It is notable that the factors independently associated with sus-
tained postpartum anxiety in our model were psychosocial/psycholo-
gical factors. Women with greater levels of perceived stress were more
likely to have sustained anxiety, while women reporting greater social
support from their partners were less likely to have sustained anxiety.
It is likely that these two constructs are inter-related. For example, a
woman with high stress levels because of negative life events (including
poor partner support) may experience more anxiety. However, a
woman who is anxious may perceive life events to be more stressful
and may have difficulty effectively communicating with her partner
(Endler, 1997). Emotional social support enhances the ability to cope
with stressful life events, and structural social support in the form of
help with household tasks and infant care may remove some of the
stressors experienced by postpartum women (Leahy-Warren et al.,
2011). The impact of maternal self-esteem is worth exploring in greater
detail. Individuals with high self-esteem do not experience negative
emotions as often as those with low self-esteem and develop better
coping strategies (Liu et al., 2014). Given the inherent psychological
vulnerability of the transition to parenthood, a woman with low self-
esteem may become anxious. It is notable that these psychological
factors overpowered socio-demographic variables like maternal age,
income level, and parity as well as specific stressors like child care
stress in our multivariable models. This is important for targeted
identification of high risk women and also for development of
psychosocial/psychological interventions that are likely to be effective
at preventing and/or treating sustained anxiety.

Table 3
Predictors of sustained postpartum anxiety.

Domain Risk factor Model 1 Model 2

Odds ratio 95% confidence interval p-value Odds ratio 95% confidence interval p-value

Life stressors Perceived stress at 1 week 4.71 2.43–6.84 < 0.001 3.74 2.17–6.44 < 0.001
Child care stress at 4 weeks 1.60 1.06–2.41 0.02 1.52 1.00–2.31 0.05
Social support Partner social support at 1 week 0.59 0.41–0.86 0.01 0.59 0.40–0.85 0.01
Maternal adjustment Self-esteem at 4 weeks 0.58 0.35–0.97 0.04 0.61 0.47–1.01 0.06
Psychiatric Depression symptomatology at 1 week – – – 1.57 0.62–3.94 0.34

Note: Continuous predictors expressed as OR for 1 SD increase in predictor.

135
C.-L. Dennis et al.
4.4. Limitations
Strengths of our study include its prospective design, large sample
size, and socioeconomically diverse sample of women from a large
health region near Vancouver, Canada. We also used a validated
measure of self-reported anxiety that performs well in perinatal
populations (Dennis et al., 2013). However, we did not undertake
any psychiatric interviews and were therefore not able to differentiate
specific types of anxiety disorders (e.g., generalized anxiety disorder,
obsessive compulsive disorder, panic disorder, post-traumatic stress
disorder). Further, sustained anxiety was defined as anxiety at 1 week
and/or 4 weeks, and 8 weeks postpartum; the persistence of anxiety
beyond this 4-8 week period is unknown. Although we aimed to
generate a population-based sample by recruiting women from both
clinicians (e.g., family physicians, obstetricians) and public health
nurses providing care to women in the community following delivery,
our sampling methods would have excluded women with no contact
with the health care system. Such women likely experience additional
stressors which impact their risk for postpartum mood disorders.
Women with ethnic minority backgrounds and single mothers were
slightly underrepresented in our sample; our results may have limited
generalizability to these populations. Similar to other longitudinal
epidemiologic studies (Fewtrell et al., 2008), 81% of eligible women
agreed to participate in our study, and 60% of these responded to the 8-
week questionnaire. Low social support emerged as an important
predictor of postpartum anxiety; however, we were unable to deter-
mine whether anxiety was a result of frustration due to the partner's
lack of involvement or an appropriate response to high levels of
conflict. Similarly, although we measured stressful life events and
psychosocial risk resulting from a number of issues, including family
violence (Reid et al., 1998), we did not examine intimate partner
violence in detail; future studies should examine this issue more
closely. The analysis examining immigration within the past 5 years
may have been underpowered, given the wide confidence intervals;
results should be interpreted with caution. Finally, our study was
originally conducted in 2001–2002. Although the healthcare landscape
may have changed subsequently, impacting the generalizability of
prevalence estimates, we expect the multivariable associations to
remain valid (Rothman et al., 2013).
4.5. Implications
Persistent mental health problems are a major health care challenge
and are associated with significant morbidity and social burden (Young
et al., 2008). Research in the general population has shown that, like
depression, anxiety disorders can be persistent; such disorders are
associated with substantial morbidity (Schweizer and Rickels, 1995).
Our study demonstrates that a significant proportion of women
experience sustained anxiety in the postpartum period specifically.
This finding has important implications for early identification and
treatment. Screening of women with multiple administrations of an
anxiety scale across the span of several weeks will identify women with
sustained postpartum anxiety (Matthey and Ross-Hamid, 2012).
However, universal screening of new mothers is often difficult because
of inadequate financial and administrative resources. Therefore, it may
be useful to target screening among women who are at high risk for
sustained postpartum anxiety. Our findings suggest that these high risk
women may be those who experience high levels of perceived stress and
poor social support. They may also experience high levels of perceived
childcare stress and have low self-esteem. These factors are not
traditionally measured in postpartum psychosocial risk scales; this is
perhaps a gap that should be addressed to identify the women who are
most in need of treatment.
The lack of correlation between the depression scores at 1 week
postpartum and sustained postpartum anxiety suggests that both
conditions may be screened separately in the postpartum period.
136
Journal of Affective Disorders 213 (2017) 131–137
Since the Edinburgh Postnatal Depression Scale is already included
in many postpartum assessments, a screening tool for anxiety could be
added to provide a more comprehensive clinical assessment. We are
currently examining sustained anxiety up to two years postpartum in
another cohort to understand the long-term relationship between
depression and anxiety.
Our findings also have implications for treatment and prevention of
postpartum anxiety. Given the importance of perceived stress and
partner social support in the prediction of sustained postpartum
anxiety, health care providers may want to consider providing antici-
patory guidance related to co-parenting to new parents. Research has
shown that co-parenting, which involves cooperation and collaboration
within the parent dyad in postpartum tasks such as breastfeeding and
caring for the newborn, is associated with positive postpartum out-
comes (Abbass-Dick and Dennis, 2017). It is possible that such
approaches could also prevent poor mental health outcomes, such as
postpartum anxiety.
5. Conclusion
Given the association between postpartum anxiety and adverse
maternal and child outcomes, it is important that researchers and
clinicians understand who experiences sustained postpartum anxiety.
Our findings suggest several psychosocial factors that may be used to
identify women with ongoing anxiety during the postpartum period.
Given that this study is one of the first to examine risk factors for
sustained postpartum anxiety, our findings should be replicated and
validated in different samples. Early identification and treatment of
women with sustained postpartum anxiety may improve outcomes for
both mother and child.
Funding
This work was supported by the Canadian Institutes of Health
Research, Canada Research Chair Program and a Tom Kierans
International Postdoctoral Fellowship.
Acknowledgments
We thank the Lawrence S. Bloomberg Faculty of Nursing of
University of Toronto for providing a Tom Kierans International
Postdoctoral Fellowship to KFH.
References
Abbass-Dick, J., Dennis, C.-L., 2017. Breast-feeding coparenting framework: a new
framework to improve breast-feeding duration and exclusivity. Fam. Comm. Health
1, 28–31.
American College of Obstetricians & Gynecologists, 2010. Committee opinion no. 453:
screening for depression during and after pregnancy. Obstet. Gynecol. 115, 394.
Arteche, A., et al., 2011. The effects of postnatal maternal depression and anxiety on the
processing of infant faces. J. Affect. Disord. 133, 197–203.
Barker, E.D., et al., 2011. The contribution of prenatal and postnatal maternal anxiety
and depression to child maladjustment. Depress Anxiety 28, 696–702.
Barnett, B., Parker, G., 1986. Possible determinants, correlates and consequences of high
levels of anxiety in primiparous mothers. Psychol. Med. 16, 177–185.
Boyce, P., et al., 2001. The development of a brief personality scale to measure
vulnerability to postnatal depression. Arch. Women’s Ment. Health 3, 147–153.
Britton, J.R., 2008. Maternal anxiety: course and antecedents during the early
postpartum period. Depress Anxiety 25, 793–800.
Cox, J.L., et al., 1987. Detection of postnatal depression: development of the 10-item
Edinburgh Postnatal Depression Scale. Br. J. Psychitry 150, 782–786.
Craske, M.G., Stein, M.B., 2016. Anxiety Lancet. http://dx.doi.org/10.1016/S0140-
6736(16)30381-6.
Cutrona, C.E., Russell, D.W., 1987. The provisions of social relationships and adaptation
to stress. Adv. Pers. Rel. 1, 37–67.
Davis, M.H., et al., 1998. Relationship-specific and global perceptions of social support:
associations with well-being and attachment. J. Pers. Soc. Psychol. 74, 468.
Dennis, C.L., et al., 2004. Identifying women at‐risk for postpartum depression in the
immediate postpartum period. Acta Psych. Scand. 110, 338–346.
Dennis, C.L., et al., 2013. Can we identify mothers at-risk for postpartum anxiety in the
immediate postpartum period using the state-trait anxiety inventory? J. Affect.
C.-L. Dennis et al.
Disord. 25, 1217–1220.
Dennis, C.L., 2017. The prevalence of antenatal and postnatal anxiety: asystematic
review and meta-analysis. Brit. J. Psychiatry, (In press).
Dennis, C.-L., 2003. The effect of peer support on postpartum depression: a pilot
randomized controlled trial. Can. J. Psychiatry 48, 115–124.
Dennis, C.-L., Faux, S., 1999. Development and psychometric testing of the breastfeeding
self-efficacy scale. Res. Nurs. Health 22, 399–409.
Dennis, C.-L., Boyce, P., 2004. Further psychometric teseting of a brief personality scale
to measure vulnerability to postpartum depression. J. Psychosom. Obstet. Gynecol.
25, 305–311.
Enatescu, V.-R., et al., 2014. State and trait anxiety as a psychopathological phenomenon
correlated with postpartum depression in a Romanian sample: a pilot study. J.
Psychosom. Obstet. Gynecol. 35, 55–61.
Endler, N.S., 1997. Stress, anxiety and coping: the multidimensional interaction model.
Can. Psychol. 38, 136.
Engle, P.L., et al., 1990. Prenatal and postnatal anxiety in Mexican women giving birth in
Los Angeles. Health Psychol. 9, 285.
Feldman, R., et al., 2009. Maternal depression and anxiety across the postpartum year
and infant social engagement, fear regulation, and stress reactivity. J. Am. Acad.
Child Adolesc. Psychiatry 48, 919–927.
Fewtrell, M.S., et al., 2008. How much loss to follow-up is acceptable in long-term
randomised trials and prospective studies? Arch. Dis. Child. 93, 458–461.
Giannandrea, S.A., et al., 2013. Increased risk for postpartum psychiatric disorders
among women with past pregnancy loss. J. Women’s Health 22, 760–768.
Grant, K.-A., et al., 2008. Maternal anxiety during the transition to parenthood: a
prospective study. J. Affect. Disord. 108, 101–111.
Hodnett, E.D., Simmons-Tropea, D.A., 1987. The Labour Agentry Scale: psychometric
properties of an instrument measuring control during childbirth. Res. Nurs. Health
10, 301–310.
Johnstone, S.J., et al., 2001. Obstetric risk factors for postnatal depression in urban and
rural community samples. Aust. N. Z. J. Psychiatry 35, 69–74.
Keim, S.A., et al., 2011. A prospective study of maternal anxiety, perceived stress, and
depressive symptoms in relation to infant cognitive development. Early Hum. Dev.
87, 373–380.
Kessler, R.C., et al., 2010. Epidemiology of anxiety disorders. Curr. Top. Behav. Neurosci.
2, 21–35.
Leahy-Warren, P., et al., 2011. Postnatal depression in first-time mothers: prevalence
and relationships between functional and structural social support at 6 and 12 weeks
postpartum. Arch. Psychiatr. Nurs. 25, 174–184.
Liu, Y., et al., 2014. Affect and self-esteem as mediators between trait resilience and
psychological adjustment. Pers. Individ. Dif. 66, 92–97.
Martini, J., et al., 2013. Maternal anxiety disorders prior to conception, psychopathology
during pregnancy and early infants' development: a prospective-longitudinal study.
Arch. Women’s Ment. Health 16, 549–560.
Martini, J., et al., 2015. Risk factors and course patterns of anxiety and depressive
disorders during pregnancy and after delivery: a prospective-longitudinal study. J.
Affect. Disord. 175, 385–395.
137
Journal of Affective Disorders 213 (2017) 131–137
Matthey, S., Ross-Hamid, C., 2012. Repeat testing on the Edinburgh Depression Scale
and the HADS-A in pregnancy: differentiating between transient and enduring
distress. J. Affect. Disord. 141, 213–221.
Petzoldt, J., et al., 2014. Maternal anxiety disorders predict excessive infant crying: a
prospective longitudinal study. Arch. Dis. Child. 9, 800–806.
Pinheiro, K.A., et al., 2014. Serum NGF, BDNF and IL-6 levels in postpartum mothers as
predictors of infant development: the influence of affective disorders. PloS One,
e94581.
Reck, C., et al., 2008. Prevalence, onset and comorbidity of postpartum anxiety and
depressive disorders. Acta Psych. Scand. 118, 459–468.
Reck, C., et al., 2009. Maternity blues as a predictor of DSM-IV depression and anxiety
disorders in the first three months postpartum. J. Affect. Disord. 113, 77–87.
Reck, C., et al., 2012. Effects of postpartum anxiety disorders and depression on maternal
self-confidence. Infant Behav. Dev. 35, 264–272.
Reck, C., et al., 2013. Infant distress to novelty is associated with maternal anxiety
disorder and especially with maternal avoidance behavior. J. Anxiety Disord. 27,
404–412.
Reid, A.J., et al., 1998. Using the alpha form in practice to assess antenatal psychosocial
health. Can. Med. J. Assoc. 159, 677–684.
Rosenberg, M., 1965. Society and the Adolescent Self-image. Princeton University Press,
Princeton, NJ.
Rothman, K.J., Gallacher, J.E.J., Hatch, E., 2013. Why representativeness should be
avoided. Int. J. Epidemiol. 42, 1012–1014.
Schweizer, E., Rickels, K., 1995. The long-term management of generalized anxiety
disorder: issues and dilemmas. J. Clin. Psychiatry 57, 9–12.
Spielberger, C.D., 2010. State‐Trait Anxiety Inventory. Wiley Online Library, Palo Alta,
CA.
Tabachnick, B.G., Fidell, L.S., 2001. Using Multivariate Statistics. Allyn & Bacon,
Boston, MA.
Taylor, J., Johnson, M., 2013. The role of anxiety and other factors in predicting
postnatal fatigue: from birth to 6 months. Midwifery 29, 526–534.
Tennant, C., Andrews, G., 1976. A scale to measure the stress of life events. Aust. N. Z. J.
Psychiatry 10, 27–32.
Tietz, A., et al., 2014. Maternal bonding in mothers with postpartum anxiety disorder:
the crucial role of subclinical depressive symptoms and maternal avoidance
behaviour. Arch. Women’s Ment. Health 17, 433–442.
Warren, S.L., et al., 2006. Maternal panic disorder: infant prematurity and low birth
weight. J. Anxiety Disord. 20, 342–352.
Wenzel, A., et al., 2005a. Sexual adjustment in postpartum women with generalized
anxiety disorder. J. Reprod. Infant. Psychol. 23, 365–366.
Wenzel, A., et al., 2005b. Anxiety symptoms and disorders at eight weeks postpartum. J.
Anxiety Disord. 19, 295–311.
Wilson, L.M., et al., 1996. Antenatal psychosocial risk factors associated with adverse
postpartum family outcomes. Can. Med. J. Assoc. 154, 785.
Young, A.S., et al., 2008. Persistent depression and anxiety in the United States:
prevalence and quality of care. Psychiatr. Serv. 59 (12), 1391–1398.