Journal of Affective Disorders 192 (2016) 56–63

Contents lists available at ScienceDirect

Journal of Affective Disorders

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

Review article

A systematic review and meta-analysis of the association between

unintended pregnancy and perinatal depression
Amanuel Alemu Abajobir a,n, Joemer Calderon Maravilla a, Rosa Alati a,b,
Jackob Moses Najman a,c
a
School of Public Health, The University of Queensland, Australia
b
Centre for Youth Substance Abuse Research, The University of Queensland, Australia
c
School of Social Sciences, The University of Queensland, Australia

art ic l e i nf o a b s t r a c t

Article history: Background: There is a growing interest in exploring maternal mental health effects of unintended
Received 12 August 2015 pregnancies carried to term. However, the evidence base from a small number of available studies is
Received in revised form characterised by considerable variability, inconsistency and inconclusive findings. We present a sys-
3 November 2015
tematic review and meta-analysis of all available studies on unintended pregnancy as these are related to
Accepted 10 December 2015
Available online 17 December 2015
maternal depression.
Methods: Using PRISMA guideline, we systematically reviewed and meta-analysed studies reporting an
Keywords: association between unintended pregnancy and maternal depression from PubMed, EMBASE, PsychINFO
Unintended or unplanned pregnancy and Google Scholar. We used a priori set criteria and included details of quality and magnitude of effect
Childbirth
sizes. Sample sizes, adjusted odds ratios and standard errors were extracted. Random effects were used
Depression
to calculate pooled estimates in Stata 13. Cochran's Q, I2 and meta-bias statistics assessed heterogeneity
Maternal mental health disorders
and publication bias of included studies.
Results: Meta-bias and funnel plot of inverse variance detected no publication bias. Overall prevalence of
maternal depression in unintended pregnancy was 21%. Unintended pregnancy was significantly asso-
ciated with maternal depression. Despite statistically significant heterogeneities of included studies, sub-
group analyses revealed positive and significant associations by types of unintended pregnancies, timing
of measurements with respect to pregnancy and childbirth, study designs and settings.
Conclusions: The prevalence of perinatal depression is two-fold in women with unintended pregnancy.
Perinatal care settings may screen pregnancy intention and depression of women backed by integrating
family planning and mental health services.
& 2015 Elsevier B.V. All rights reserved.

Contents

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
2. Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57
3. Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59
4. Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 61
5. Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Contributors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Funding . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
Conflict of interest. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
Acknowledgement. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63

n
Corresponding author.
E-mail address: amanuel.abajobir@uqconnect.edu.au (A.A. Abajobir).

http://dx.doi.org/10.1016/j.jad.2015.12.008
0165-0327/& 2015 Elsevier B.V. All rights reserved.

Downloaded for Meera Al Ali (meeali@seha.ae) at Abu Dhabi Health Services Co from ClinicalKey.com by Elsevier on February
27, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
 A.A. Abajobir et al. / Journal of Affective Disorders 192 (2016) 56–63
1. Introduction
Internationally there are concerns associated with the high
prevalence of unintended pregnancies (McIntosh and Finkle, 1995)
carried to term and their impact on maternal mental health in
general and (Brown and Eisenberg, 1995; Gipson et al., 2008; Lo-
gan et al., 2007) perinatal depression in particular (Klier et al.,
2008; Gavin et al., 2005; Kingsbury et al., 2015). Poor maternal
mental health is a public health problem (Balaji et al., 2007) with
implications for familial wellbeing (Burke, 2003), child psycho-
pathology and abnormal development (Goodman and Gotlib,
1999; Cogill et al., 1986; Kurstjens and Wolke, 2001). Moreover,
pregnancy and childbirth related events including unintended
pregnancy (Brown and Eisenberg, 1995; Gipson et al., 2008; Logan
et al., 2007) may make a substantial contribution to maternal
depression (Klier et al., 2008; Gavin et al., 2005; Kingsbury et al.,
2015).
Maternal mental health problems have been purported as both
causes and consequences of unintended pregnancies. For instance,
stress is suggested to increase the risk of poor maternal mental
health (Hall et al., 2014) by reducing coping skills (Kuroki et al.,
2008) in mothers with unintended pregnancies. However, findings
from a randomized clinical trial reveal a negative association be-
tween unintended pregnancy and stress (Kuroki et al., 2008).
Other available data show inconsistent findings (Ikamari et al.,
2013; McCrory and McNally, 2013; Messer et al., 2005; Yanikkerem
et al., 2013; Fellenzer and Cibula, 2014; Iranfar et al., 2005; Lau and
Keung, 2007; Cheng et al., 2009; Najman et al., 1991; Grussu et al.,
2005; Mercier et al., 2013; Christensen et al., 2011). Unintended
pregnancies assessed through retrospective reports were found to
be associated with maternal perinatal mental health including
varying degrees of depressive disorders (Ikamari et al., 2013;
McCrory and McNally, 2013; Messer et al., 2005; Yanikkerem et al.,
2013), moderate to severe prenatal (Yanikkerem et al., 2013;
Fellenzer and Cibula, 2014; Iranfar et al., 2005; Lau and Keung,
2007) and postpartum (Cheng et al., 2009) depression. Findings
from some prospective studies suggest prenatal (Messer et al.,
2005) and postpartum depression (Iranfar et al., 2005; Najman
et al., 1991), and other mood disorders in women (Grussu et al.,
2005). Other longitudinal studies however report statistically
nonsignificant associations between unintended pregnancies and
postpartum depression (Mercier et al., 2013; Christensen et al.,
2011).
Given high rates of unintended pregnancies in the general
population (McIntosh and Finkle, 1995) and its potential maternal
mental health impact (Brown and Eisenberg, 1995; Gipson et al.,
2008; Logan et al., 2007), there is a need to know more about
associations between pregnancy intendedness and maternal de-
pression. A handful of prior studies of the impact of unintended-
ness on maternal depression have had methodological weaknesses
with concerns about the validity of measurement and substantial
inconsistency across findings (Brown and Eisenberg, 1995; Gipson
et al., 2008; Logan et al., 2007). To our knowledge, there has been
no systematic review and meta-analysis of unintendedness of
pregnancy and its relationships with depression. The current
meta-analysis focuses on overall prevalence and association of
unintended pregnancies and maternal depression and provides
sub-group estimates across time of measurement with regard to
pregnancy or childbirth, types of unintendedness, study designs
and settings.
2. Methods
We employed the Preferred Reporting Items for Systematic
Reviews and Meta-Analyses (PRISMA) guideline (Moher et al.,
Downloaded for Meera Al Ali (meeali@seha.ae) at Abu Dhabi Health Services Co from ClinicalKey.com by Elsevier on February
27, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
57
2009) and used two strategies to identify studies—systematic
searches of electronic databases (PubMed, EMBASE, PsycINFO and
Google Scholar) and hand searches of reference lists of included
articles. A Boolean search strategy was established and used. The
key terms were: exposure identifiers—uninten* OR unintended OR
unplanned OR unwanted pregnancy OR childbirth; outcome identi-
fiers—maternal depression OR perinatal depression OR mental health
disorders; and combined exposure and outcome identifiers— un-
inten* OR unintended OR unplanned OR unwanted pregnancy OR
childbirth AND maternal depression OR perinatal depression OR
mental health disorders. Moreover, we used an *operator “uninten*”
to capture possible variations in two terms, “unintentional” and
“unintended”.
We included (McIntosh and Finkle, 1995) all quantitative stu-
dies (Brown and Eisenberg, 1995) which employed robust analyses
controlling for confounders and (Gipson et al., 2008) published
before 2015 reporting unintendedness and maternal depression of
singleton live births. Qualitative (McIntosh and Finkle, 1995) and
quantitative studies of descriptive-only statistics (Brown and Ei-
senberg, 1995) were excluded.
We extracted sample sizes, proportions and adjusted odds ra-
tios with 95% confidence intervals (OR; 95%CI). Numerator and
denominator data and beta coefficients and their standard errors
(if given) were used to estimate ORs where ORs with 95%CI were
not provided. Efforts were made to contact corresponding authors,
whenever there was insufficient information to calculate the
estimates.
Quality was assessed based on study features using an eight-
point checklist (DB, 2014). Criteria for quality assessment include
the validity of study methods, interpretation and applicability of
results. Quality scores were derived for our purposes and did not
necessarily represent the original objective of each included re-
search article. Moreover, we used effect size magnitude definitions
for each included study based on individual OR obtained from
forest plot (small ( o/¼1.44); medium ( 4/¼ 2.47); and large ( 4/
¼4.25)) (Cohen, 1992). The primary author (AAA) conceived the
hypothesis, developed the methodology, identified all potential
studies, wrote and proofread the manuscript. Two authors (AAA
and JCM) reviewed abstracts, extracted data and assessed quality
of included studies independently. Discussions and mutual con-
sensus was sought when possible disagreements were raised be-
tween these two reviewers.
Stata 13 was used for meta-analysis and forest plots that
showed combined estimates with 95%CI. It produced Cochran's Q,
I2 and inverse variances tests of overall and between group het-
erogeneity of estimates (Egger et al., 2008). Cochran's Q set at 0.05
random error and I2 statistic were used to determine hetero-
geneity (Doi and Williams, 2013) suggesting the extent of bias
associated with eligible studies. We reviewed forest plots of
summary estimates of each study to determine whether we could
identify any heterogeneity between studies. We used random ef-
fects (RE) models to account for any remaining heterogeneity in
the estimates across studies because these models account for
both random variability and the variability in effects among the
studies (Olkin, 1999; Higgins et al., 2003). We conducted meta-bias
and funnel plot for publication bias. A series of sub-group analyses
were conducted to estimate effect sizes for maternal depression in
terms of time of outcome measurement (given reference to
pregnancy or childbirth), types of unintendedness, study designs
and settings. The Diagnostic Statistical Manual (DSM-IV) definition
of depression was adapted to assess perinatal depression across
countries and cultures and was used to measure point prevalence
(Gorman et al., 2004).
 58
Table 1
Summary of characteristics of 10 full-text articles with quality scores and effect-size odds ratios magnitude.

Authors Study Study Exposure Number and character- Outcome Key finding Confounders adjusted in mul- Significant Response Quality Effect size
and year country design measurement istics of participants measurement tivariable analyses confounders rate (%) score magnitude

Fellenzer USA CS Antepartum 18,059 Women from New Self-reported Mistimed pregnancy was a pre- Maternal socio-demographics, Education, 94 6 Large
and Cibu- depression York State-wide Perinatal prenatal de- dictor for moderate to severe number of termination of a race, marital
la (2014) rated within Data System pression prenatal depression and moder- pregnancy, timing for ANC, ma- status, Medi-
Downloaded for Meera Al Ali (meeali@seha.ae) at Abu Dhabi Health Services Co from ClinicalKey.com by Elsevier on February

72 hours of symptoms ate to severe prenatal depression ternal medical risk, GDM, caid use,
delivery was associated with unintended smoking, amount smoked, drug smoking and
27, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.

pregnancy. and alcohol use drug use

McCrory Ireland CS 9 Months 10,567 Mothers of chil- CESD-8 Possibility of maternal depression Maternal demographics – 64.5 5 Small
and postpartum dren at 9 months was high among women with
McNally unintended pregnancy.
(2013)
Mercier USA Multiphase 3 and 12 688 Women at 3 months EPDS Not statistically significant asso- Socio-demographics and re- – 63% at 6 Large
et al. prospective Months and 550 at 12 months ciation was found between preg- productive history 3 months and
(2013) cohort postpartum nancy intention and postpartum 45 excluded

A.A. Abajobir et al. / Journal of Affective Disorders 192 (2016) 56–63

study depression at 12 month
Abbasi et al. USA Longitudinal 1 Month 2972 Women from hos- EPDS Postpartum depression was Socio-demographics and pre- Race and pre- 98.86 6 Small
(2013) postpartum pital or community higher in unintended pregnancy. pregnancy anxiety and pregnancy
settings depression anxiety/
depression
Dibaba et al. Ethiopia CS Antepartum 622 Pregnant women EPDS Unwanted pregnancy was asso- Socio-demographics, pregnancy Occupation, 99 6 Small
(2013) ciated with depression. status, parity, social support and social support
partner violence and partner
violence
Christensen USA RCT-second- Pregnancy-to 215 Low-income Hispanic BDI-II Unintended pregnancy was not Socio-demographics, lifetime Social sup- 100 [2 ex- 5 Large
et al. ary analyses 12 mo immigrant women associated with the “Pregnancy history of physical or sexual port, health cluded for
(2011) from Mood postpartum High” pattern of PPD but with a abuse, lifetime history of major insurance homo-
and Health marginally significant “Post- depression and social support and history of geneity]
Project partum High” pattern in depres- abuse
sive symptoms.
Cheng et al. USA CS 3 to 9 Months 9048 Women 2 to Not specified 41.4% Pregnancies were reported Maternal socio-demographics Not reported 71 5 Large
(2009) postpartum 9 months postpartum as unintended and was asso- and previous live births
ciated with depression. behaviour
Lau and China CS 2nd Trimester 2178 Pregnant women EPDS 40.7% Mother of unintended Socioeconomic, obstetric and Many were 92.1 6 Small
Keung antepartum visiting 5 hospitals pregnancies had severe family conflict characteristics significant
(2007) depression. correlates of
unintended
pregnancy.
Messer et al. USA Prospective 3 to 12 1908 Women aged 416 CES-D Not intending pregnancy was as- Maternal socio-demographics – 75.3% 6 Medium
(2005) cohort from Months year at 24–29 weeks of sociated with increased risk of
26–29th postpartum GA from Pregnancy, In- low, medium and high levels of
week of GA fection, and Nutrition perceived stress during
to delivery (PIN), a prospective co- pregnancy
hort study
Najman Australia Prospective 6 Months 277 Postpartum women DSSI anxiety Women with unwanted children Socio-demographics and parity Age, marital 93% 6 Small
et al. pre-birth postpartum and had slightly higher rates of anxi- status, parity
(1991) cohort depression ety and depression. and income

ANC – antenatal care; CS – cross-sectional; GA – gestational age; GDM – gestational diabetes milletus; PPD – postpartum depression; RCT – random clinical trial.
 A.A. Abajobir et al. / Journal of Affective Disorders 192 (2016) 56–63
Identified N=512 studies were identified
duplications.
Screened
n=61 studies were screened.
n=6 full-text articles
from manual search
Eligible
n=19 full-text articles were assessed
for eligibility.
Included
n=10 studies were
systematic review.
Fig. 1. Schematic representation of studies included in the systematic review using PRISMA checklist and flow-diagram.
3. Results
Given the paucity of the existing literature, we included studies
from both developed and developing countries. A total of 512
studies were identified using search engines and strategies. We
screened 61 full-text published articles. Ten full-text studies were
eligible. Half of the studies, 50% (n ¼5), used cross-sectional
methods, 40% (n ¼4) observational longitudinal and 10% (n ¼1)
was a randomized clinical trial design. The majority, 60% (n ¼6),
were conducted in USA. Sample sizes ranged from 215 (Chris-
tensen et al., 2011) to 18,059 (Fellenzer and Cibula, 2014) singleton
mothers. The mean quality score was 5.63 (SD 7 0.48). Based on an
analysis of individual forest plots (Fig. 2), four studies had large
effect size estimates. Table 1 presents detailed summary char-
acteristics of included studies. The PRISMA flow diagram was used
to present the process of screening and selection of eligible studies
(Fig. 1).
We aggregated studies by pregnancy stages, types of unin-
tendedness, designs and settings. Pregnancy intention was mea-
sured by using standard questionnaires as used by National Survey
for Family Growth (NSFG) (Kost and Lindberg, 2015), Pregnancy
Risk Assessment Monitoring System (PRAMS) (Fellenzer and Ci-
bula, 2014) and Demographic and Health Survey (DHS) (Palamu-
leni and Adebowale, 2014). The common survey questions used
were: “At the time you became pregnant, did you want to become
pregnant then, did you want to wait until later, or did you not
want to have any (more) children at all”? Women were also asked
whether the pregnancy was planned or not, intended or not or
wanted or not wanted. The questions: “Are you trying to get
pregnant now or in future? and how important is avoiding a
pregnancy to you?” are used in prospective studies while retro-
spective studies asked women if the pregnancy was intended or
not. Pregnancy intention could be categorized as intended vs.
unintended (Altfeld et al., 1997) or as intended/wanted, mistimed
and unwanted (D'Angelo et al., 2004). The majority (n ¼7) of in-
cluded studies applied the intended vs. unintended scheme.
However, three studies (Fellenzer and Cibula, 2014; Cheng et al.,
Downloaded for Meera Al Ali (meeali@seha.ae) at Abu Dhabi Health Services Co from ClinicalKey.com by Elsevier on February
27, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
59
after avoiding
451
n=187 studies
studiesexcluded
with
(non-relevant
irrelevant topics
topic).
excluded
n=48 studies were
excluded (irrelevant
data of interest).
n=9 full-text articles
were excluded (n=8 did
not adjust for
confounders; and n=1
included in outlier after sensitivity
analysis).
2009; Dibaba et al., 2013) used intended, mistimed and unwanted
categories. We merged mistimed and unwanted into unintended
category for three of these studies and used the intended vs. un-
intended group for overall estimates. One study used the combi-
nation of two methods (Christensen et al., 2011). We used the
crude (intended vs. unintended) and specific classification (mis-
timed vs. unwanted) for overall and sub-group estimates, respec-
tively. We conducted separate sub-group analyses for mistimed vs.
unwanted groups, by measurement time, study designs and set-
tings. Depression was ascertained with both structured and stan-
dardized self-reported scales (Fellenzer and Cibula, 2014) such as
the Delusion-States-Symptoms Inventory anxiety and depression
scale (DSSI) (Najman et al., 1991), the Beck Depression Inventory
(BDI) (Christensen et al., 2011), the Centre for Epidemiological
Studies Depression Scale (CES-D) (McCrory and McNally, 2013;
Messer et al., 2005) and the Edinburgh Postpartum Depression
Scale (EPDS) (Lau and Keung, 2007; Dibaba et al., 2013; Abbasi
et al., 2013; Mercier et al., 2013). Studies with standard tools used
standardised cut-offs to diagnose depression. For depression
measured at different time points (Christensen et al., 2011), we
selected the most conservative effect size for inclusion in the
analysis. One study did not report the measurement of depression
used (Cheng et al., 2009). Three studies measured depression
during pregnancy while seven measured depression postpartum
(Table 1).
The overall point prevalence of maternal depression in women
who were in the unintended pregnancy category and/or childbirth
was 21% (Table 2). The overall risk difference in prevalence of
depression was 12% (95%CI: 5–19%; p o0.001).
Depression was estimated for sub-study characteristics in-
cluding time of measurement, types of unintendedness, study
designs and settings. Corresponding 95%CI were indicated. Sub-
group analysis for types of unintended was restricted for five
studies that had a priori delineation to mistimed and unwanted
classifications. The odds of depression was higher during post-
partum, for unwanted pregnancy/childbirth, in cross-sectional
studies and in women from developed countries. Statistically
60 A.A. Abajobir et al. / Journal of Affective Disorders 192 (2016) 56–63

Fig. 2. Forest plots of unintendedness and maternal depression using log scale.

significant overall heterogeneity was recorded (Table 3). The in-
verse of variances of the between group heterogeneity were
computed: for time of measurement (Q¼55.86; po .001), types of
unintendedness (Q¼ 12.32; p o.001), study designs (Q¼26.55;
p o.001) and settings (Q¼59.75; p o.001).
The heterogeneities of included studies were excluded for
antepartum period, longitudinal study designs and developing
countries in sub-group analyses. These findings may mean that
minimal differences in perinatal depression exist during ante-
partum period when longitudinal study design is used. Moreover,
unintended pregnancy may more reliably predict perinatal de-
pression for women in developing regions of the world.
We carried out sensitivity analysis and detected statistically
significant difference for one study and excluded it (Nakku et al.,

Downloaded for Meera Al Ali (meeali@seha.ae) at Abu Dhabi Health Services Co from ClinicalKey.com by Elsevier on February
27, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
 A.A. Abajobir et al. / Journal of Affective Disorders 192 (2016) 56–63 61

Table 2 unwanted pregnancy. This suggests that mistimed and unwanted

Pooled overall and sub-group prevalence of maternal depression (n¼ 10). pregnancies and/or childbirths could result in differential out-
comes (D'Angelo et al., 2004) with higher odds of association for
Sub-group Depression in unin- Depression in in-
tended pregnancy tended pregnancy unwantedness (Brown and Eisenberg, 1995). There were sig-
nificant variations of effects across study designs and settings.
Sample Prevalence Sample Prevalence Longitudinal designs are generally considered more robust and
size (95%CI) size (95%CI) capable of predicting causal pathways and temporal sequences.
Measured at
Longitudinal and cross-sectional studies revealed nearly equiva-
Antepartum 7599 27 (9–47) 13,229 15 (5–27) lent effect sizes for depression. Unintended pregnancies were
Postpartum 9957 19 (6–35) 21,267 8 (3–13) highly associated with depression in mothers from developed
Types of unintendednessa countries.
Mistimed 7861 20 (9–34) 17,921 11 (6–17) In routine prenatal care antenatal charts do not screen for
Unwanted 2794 28 (19–38) whether the pregnancy was intended or not. This may result in
Study design under diagnosis of both unintended pregnancies and associated
Cross-sectional 14,977 28 (14–43) 25,049 13 (7–20) depression. Early prevention and intervention services might be
Longitudinal 2579 15 (0–39) 9447 7 (1–15) possible if routine screening were undertaken. However, this re-
Region of study quires a thoughtful decision as there may be a trade-off between
Developed 16,736 21 (10–35) 32,516 9 (4–16) the cost and potential harm associated with screening and the
Developing 820 20 (11–30) 1980 11 (4–21)
benefits likely to be derived. Given the high rates of unintended
pregnancies (McIntosh and Finkle, 1995), perinatal service settings
Overall pooled estimate 17,556 21 (12–33) 34,496 10 (5–16) are potential areas for interventions of all levels. Moreover, in-
a
tegrating family planning and mental health services in primary
Data extracted from only five studies reporting mistimed and unwanted
pregnancies. The remaining five studies were excluded from analysis of this cate-
healthcare settings may suffice in mitigating unintended preg-
gory (McCrory and McNally, 2013; Messer et al., 2005; Lau and Keung, 2007; nancies and depression.
Najman et al., 1991; Abbasi et al., 2013). Our meta-analysis has the following strengths: it included all
existent well controlled studies that met a priori set criteria from
Table 3
both developed and developing countries. This presumably would
Overall and sub-group estimates of maternal depression of unintended pregnancy/ have unbiased representative estimates across different social at-
childbirth (n¼ 10). tributes. The effect size estimates used robust techniques and sub-
group analyses that aimed to disentangle cause-effect relation-
Sub-group RE OR (95%CI) Heterogeneity p-Value I2
ships and differential interventions of unintendedness and ma-
Measured at ternal depression. Though significant heterogeneities exist across
Antepartum 2.24 (1.88–2.68) 3.84 .15 48.0 included studies, sub-group analyses explored statistically sig-
Postpartum 2.49 (1.50–4.13) 188.87 o.001 97.0 nificant effect sizes. Despite relatively small samples, our findings
Types of unintendednessa have reasonable external validity.
Mistimed 1.92 (1.41–2.61) 37.64 .01 68.0 However, these findings should be interpreted cautiously be-
Unwanted 2.01 (1.33–3.04) 23.94 o.001 83.0 cause of subjective quality assessments and relatively small sam-
Study design ple sizes for sub-group analyses. Pooling estimates may be difficult
Cross-sectional 2.58 (1.71–3.88) 281.31 o.001 99.0 given different modes of measurement and classification (Brown
Longitudinal 2.09 (1.54–2.84) 7.06 .13 43.0
and Eisenberg, 1995). This may limit our predictive ability. More-
Region of study over, half of the studies were cross-sectional and might reflect
Developed 2.50 (1.78–3.52) 285.87 o.001 98.0 recall bias, with a high possibility of ex-post rationalisation of the
Developing 1.98 (1.57–2.50) 0.80 .37 0.0
pregnancy status. For example, it is possible that women with
postpartum depression were more likely to endorse an unin-
Overall pooled 2.36 (1.77–3.16) 289.81 o.001 97.0 tended pregnancy than non-depressive mothers. Moreover, there
estimate
were significant heterogeneities across included studies, limiting
a
Data extracted from only five studies reporting mistimed and unwanted us from firm conclusion though sub-group analyses consistently
pregnancies. The remaining five studies were excluded from analysis of this cate- reveal similar associations. There are problems with the measure
gory (McCrory and McNally, 2013; Messer et al., 2005; Lau and Keung, 2007; of unintended pregnancy from eligible studies included in the
Najman et al., 1991; Abbasi et al., 2013).
meta-analysis. For instance, the measures based on timing and
numbers of children do not count all unintended pregnancies
2007). Meta-bias detected nonsignificant publication bias (bias (those that occurred at the right time but the wrong relationship
coefficient ¼  1.66 (95%CI:  10.79–7.47); p 4.69). Fig. 3 or when the woman had lost a job). These measures could include
mothers who were satisfied with a given number (with right
timing) of children but were pleasantly surprised to find they
4. Discussion would have an extra child (with no right timing). These measures
might not be adequate to identify really problematic pregnancies.
In this study we computed pooled estimates of prevalence and Finally, some depression measurements were not consistent with
OR of unintended pregnancies and maternal depression. We found DSM-IV criteria of psychiatric disorders.
that depression was higher in women who reported an unin- Our analysis was restricted to exploration of prevalence and
tended pregnancy. The findings also point to substantial significant association estimates of unintended pregnancies and perinatal
overall and sub-sample associations of unintendedness and ma- maternal depression. This does not mean an unintended preg-
ternal depression. Sub-group analysis suggests that unintended nancy is the only risk factor for maternal depression. Both organic
pregnancy is associated with maternal depression particularly and inorganic risk factors have been found to predict depression in
during the postpartum period. Mistimed pregnancy and/or child- different stages of perinatal period (Gavin et al., 2005; Lancaster
birth had slightly lower association with depression than does an et al., 2010; O'hara and Swain, 1996). Though perinatal depression

Downloaded for Meera Al Ali (meeali@seha.ae) at Abu Dhabi Health Services Co from ClinicalKey.com by Elsevier on February
27, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
62 A.A. Abajobir et al. / Journal of Affective Disorders 192 (2016) 56–63

95,000

90,000

85,000

80,000

75,000

70,000

65,000

60,000

55,000

50,000

45,000

40,000

35,000

30,000

25,000

20,000

15,000

10,000

5,000

0
0 1 2

ln OR

Fig. 3. Funnel plot for publication bias.

may be chronic or acute (O'hara and Swain, 1996) we could not
compare these findings with prior large scale systematic review
and meta-analysis (Gavin et al., 2005; Lancaster et al., 2010; O'hara
and Swain, 1996) that focused on depression but had no details of
pregnancy or childbirth intention.
Given extensive evidence on perinatal depression of women
(Gavin et al., 2005; Lancaster et al., 2010; O'hara and Swain, 1996),
this study provides an insight into the contexts of unintendedness
and maternal depression. However, all included studies exclusively
focussed on perinatal depression and little is known about the
trajectory of maternal depression for mothers with unintended
children. Future research should focus on exploring trajectories of
unintended pregnancies and maternal depression beyond the
perinatal period using large-scale longitudinal studies and con-
trolling for potential confounders.
5. Conclusions
mental health services. Moreover, maternal health services including
user-friendly family planning should be strengthened and integrated
in primary healthcare to prevent unintended pregnancy.
Contributors
AAA conceived the hypothesis, developed the methodology,
identified all potential studies, extracted the data, assessed quality,
conducted the analysis, wrote the first draft of the manuscript and
proofread the final version. JCM reviewed abstracts, extracted data
and assessed quality. RA commented on relevance of topic of in-
terest, reviewed and proofread the final version. JMN reviewed
data extraction and synthesis, reviewed the manuscript and
commented and proofread the final version. All authors approved
the final submitted version of the manuscript.

The prevalence and association of perinatal depression is two-fold

in women with unintended pregnancy than intended pregnancy. In Funding
turn, depressed women may also have high rates of unintended
pregnancies. These women would benefit from perinatal screening for No external funding obtained for this systematic review and
unintendedness and depression followed by provision of differential meta-analysis.

Downloaded for Meera Al Ali (meeali@seha.ae) at Abu Dhabi Health Services Co from ClinicalKey.com by Elsevier on February
27, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
 A.A. Abajobir et al. / Journal of Affective Disorders 192 (2016) 56–63
Conflict of interest
No conflict of interest declared.
Acknowledgement
We are grateful for the University of Queensland library for providing us with a
wide range of available online databases.
References
Altfeld, S., Handler, A., Burton, D., Berman, L., 1997. Wantedness of pregnancy and
prenatal health behaviors. Women Health 26 (4), 29–43.
Abbasi, S., Chuang, C.H., Dagher, R., Zhu, J., Kjerulff, K., 2013. Unintended pregnancy
and postpartum depression among first-time mothers. J. Women's Health 22
(5), 412–416.
Brown, S.S., Eisenberg, L., 1995. The Best Intentions: Unintended Pregnancy and the
Well-being of Children and Families. National Academies Press.
Balaji, A.B., Claussen, A.H., Smith, D.C., Visser, S.N., Morales, M.J., Perou, R., 2007.
Social support networks and maternal mental health and well-being. J. Wo-
men's Health 16 (10), 1386–1396.
Burke, L., 2003. The impact of maternal depression on familial relationships. Int.
Rev. Psychiatry 15 (3), 243–255.
Cogill, S., Caplan, H., Alexandra, H., Robson, K.M., Kumar, R., 1986. . Impact of ma-
ternal postnatal depression on cognitive development of young children. Br.
Med. J. 292 (6529), 1165–1167.
Cheng, D., Schwarz, E.B., Douglas, E., Horon, I., 2009. Unintended pregnancy and
associated maternal preconception, prenatal and postpartum behaviors. Con-
traception 79 (3), 194–198.
Christensen, A.L., Stuart, E.A., Perry, D.F., Le, H.-N., 2011. Unintended pregnancy and
perinatal depression trajectories in low-income, high-risk Hispanic immigrants.
Prev. Sci. 12 (3), 289–299.
Cohen, J., 1992. A power primer. Psychol. Bull. 112 (1), 155.
DB, W., 2014. Practical Meta-analysis Effect Size Calculator. The Campbell Colla-
boration, Oslo.
Doi, S.A., Williams, G.M., 2013. Methods of Clinical Epidemiology. Springer.
D'Angelo, D.V., Gilbert, B.C., Rochat, R.W., Santelli, J.S., Herold, J.M., 2004. Differ-
ences between mistimed and unwanted pregnancies among women who have
live births. Perspect. Sex. Reprod. Health 36 (5), 192–197.
Dibaba, Y., Fantahun, M., Hindin, M.J., 2013. The association of unwanted pregnancy
and social support with depressive symptoms in pregnancy: evidence from
rural Southwestern Ethiopia. BMC Pregnancy Childbirth 13 (1), 135.
Egger, M., Smith, G.D., Altman, D., 2008. Systematic Reviews in Health Care: Meta-
analysis in Context. John Wiley & Sons.
Fellenzer, J.L., Cibula, D.A., 2014. Intendedness of pregnancy and other predictive
factors for symptoms of prenatal depression in a population-based study.
Matern. Child. health J. 18 (10), 2426–2436.
Gipson, J.D., Koenig, M.A., Hindin, M.J., 2008. The effects of unintended pregnancy
on infant, child, and parental health: a review of the literature. Stud. Fam. Plan.,
18–38.
Gavin, N.I., Gaynes, B.N., Lohr, K.N., Meltzer-Brody, S., Gartlehner, G., Swinson, T.,
2005. Perinatal depression: a systematic review of prevalence and incidence.
Obstet. Gynecol. 106 (5 Pt 1), 1071–1083.
Goodman, S.H., Gotlib, I.H., 1999. Risk for psychopathology in the children of de-
pressed mothers: a developmental model for understanding mechanisms of
transmission. Psychol. Rev. 106 (3), 458.
Grussu, P., Quatraro, R.M., Nasta, M.T., 2005. Profile of Mood States and parental
attitudes in motherhood: comparing women with planned and unplanned
pregnancies. Birth (Berkeley, Calif.) 32 (2), 107–114.
Gorman, L.L., O'Hara, M.W., Figueiredo, B., Hayes, S., Jacquemain, F., Kammerer, M.
H., et al., 2004. Adaptation of the structured clinical interview for DSM-IV
disorders for assessing depression in women during pregnancy and post-par-
tum across countries and cultures. Br. J. Psychiatry 184 (46), s17–s23.
Hall, K.S., Kusunoki, Y., Gatny, H., Barber, J., 2014. The risk of unintended pregnancy
Downloaded for Meera Al Ali (meeali@seha.ae) at Abu Dhabi Health Services Co from ClinicalKey.com by Elsevier on February
27, 2023. For personal use only. No other uses without permission. Copyright ©2023. Elsevier Inc. All rights reserved.
63
among young women with mental health symptoms. Soc. Sci. Med. (1982) 100,
62–71.
Higgins, J.P., Thompson, S.G., Deeks, J.J., Altman, D.G., 2003. Measuring incon-
sistency in meta-analyses. Br. Med. J. 327 (7414), 557.
Ikamari, L., Izugbara, C., Ochako, R., 2013. Prevalence and determinants of unin-
tended pregnancy among women in Nairobi, Kenya. BMC Pregnancy Childbirth
13, 69.
Iranfar, S., Shakeri, J., Ranjbar, M., NazhadJafar, P., Razaie, M., 2005. Is unintended
pregnancy a risk factor for depression in Iranian women? East. Mediterr. Health
J. ¼ La. Rev. De. sante De. La. Mediterr. Orient. ¼ al-Majallah al-sihhiyah li-sharq
al-mutawassit 11 (4), 618–624.
Klier, C.M., Rosenblum, K.L., Zeller, M., Steinhardt, K., Bergemann, N., Muzik, M.,
2008. A multirisk approach to predicting chronicity of postpartum depression
symptoms. Depression anxiety. 25 (8), 718.
Kingsbury, A.M., Hayatbakhsh, R., Mamun, A.M., Clavarino, A.M., Williams, G.,
Najman, J.M., 2015. Trajectories and predictors of women's depression follow-
ing the birth of an infant to 21 years: a longitudinal study. Matern. Child Health
J. 19 (4), 877–888.
Kurstjens, S., Wolke, D., 2001. Effects of maternal depression on cognitive devel-
opment of children over the first 7 years of life. J. Child Psychol. Psychiatry 42
(05), 623–636.
Kuroki, L.M., Allsworth, J.E., Redding, C.A., Blume, J.D., Peipert, J.F., 2008. Is a pre-
vious unplanned pregnancy a risk factor for a subsequent unplanned preg-
nancy? Am. J. Obstet. Gynecol. 199 (5), 517.e1-7.
Kost, K., Lindberg, L., 2015. Pregnancy intentions, maternal behaviors, and infant
health: investigating relationships with new measures and propensity score
analysis. Demography 52 (1), 83–111.
Logan, C., Holcombe, E., Manlove, J., Ryan, S., 2007. The consequences of unintended
childbearing. Child Trends Natl. Campaign Prev. Teen Pregnancy 28, 142–151.
Lau, Y., Keung, D.W.F., 2007. Correlates of depressive symptomatology during the
second trimester of pregnancy among Hong Kong Chinese. Soc. Sci. Med. 64 (9),
1802–1811.
Lancaster, C.A., Gold, K.J., Flynn, H.A., Yoo, H., Marcus, S.M., Davis, M.M., 2010. Risk
factors for depressive symptoms during pregnancy: a systematic review. Am. J.
Obstet. Gynecol. 202 (1), 5–14.
McIntosh, C.A., Finkle, J.L., 1995. The Cairo conference on population and devel-
opment: a new paradigm? Popul. Dev. Rev., 223–260.
McCrory, C., McNally, S., 2013. The effect of pregnancy intention on maternal pre-
natal behaviours and parent and child health: results of an irish cohort study.
Paediatr. Perinat. Epidemiol. 27 (2), 208–215.
Messer, L.C., Dole, N., Kaufman, J.S., Savitz, D.A., 2005. Pregnancy intendedness,
maternal psychosocial factors and preterm birth. Matern. Child. Health J. 9 (4),
403–412.
Mercier, R.J., Garrett, J., Thorp, J., Siega-Riz, A.M., 2013. Pregnancy intention and
postpartum depression: secondary data analysis from a prospective cohort.
BJOG: Int. J. Obstet. Gynaecol. 120 (9), 1116–1122.
Moher, D., Liberati, A., Tetzlaff, J., Altman, D.G., 2009. Preferred reporting items for
systematic reviews and meta-analyses: the PRISMA statement. Ann. Intern.
Med. 151 (4), 264–269.
Mercier, R.J., Garrett, J., Thorp, J., Siega‐Riz, A.M., 2013. Pregnancy intention and
postpartum depression: secondary data analysis from a prospective cohort.
BJOG: Int. J. Obstet. Gynaecol. 120 (9), 1116–1122.
Najman, J.M., Morrison, J., Williams, G., Andersen, M., Keeping, J.D., 1991. The
mental health of women 6 months after they give birth to an unwanted baby: a
longitudinal study. Soc. Sci. Med. 32 (3), 241–247.
Nakku, J., Nakasi, G., Mirembe, F., 2007. Postpartum major depression at six weeks
in primary health care: prevalence and associated factors. Afr. Health Sci. 6, 4.
Olkin, I., 1999. Diagnostic statistical procedures in medical meta‐analyses. Stat.
Med. 18 (17–18), 2331–2341.
O'hara, M.W., Swain, A.M., 1996. Rates and risk of postpartum depression-a meta-
analysis. Int. Rev. Psychiatry 8 (1), 37–54.
Palamuleni, M.E., Adebowale, A., 2014. Prevalence and determinants of unintended
pregnancies in Malawi. Afr. Popul. Stud. 28 (1), 551–563.
Yanikkerem, E., Ay, S., Piro, N., 2013. Planned and unplanned pregnancy: effects on
health practice and depression during pregnancy. J. Obstet. Gynaecol. Res. 39
(1), 180–187.