Could listening to music during pregnancy be protective against postnatal depression and poor wellbeing post birth? Longitudinal associations from a preliminary prospective cohort study

Abstract

Objectives
This study explored whether listening to music during pregnancy is longitudinally associated with lower symptoms of postnatal depression and higher well-being in mothers post birth.

Design
Prospective cohort study.

Participants
We analysed data from 395 new mothers aged over 18 who provided data in the third trimester of pregnancy and 3 and 6 months later (0–3 and 4–6 months post birth).

Primary and secondary outcome measures
Postnatal depression was measured using the Edinburgh Postnatal Depression Scale, and well-being was measured using the Short Warwick-Edinburgh Mental Well-being Scale. Our exposure was listening to music and was categorised as ‘rarely; a couple of times a week; every day <1 hour; every day 1–2 hours; every day 3–5 hours; every day 5+hrs’. Multivariable linear regression analyses were carried out to explore the effects of listening to music during pregnancy on depression and well-being post birth, adjusted for baseline mental health and potential confounding variables.

Results
Listening during pregnancy is associated with higher levels of well-being (β=0.40, SE=0.15, 95% CI 0.10 to 0.70) and reduced symptoms of postnatal depression (β=−0.39, SE=0.19, 95% CI −0.76 to −0.03) in the first 3 months post birth. However, effects disappear by 4–6 months post birth. These results appear to be particularly found among women with lower levels of well-being and high levels of depression at baseline.

Conclusions
Listening to music could be recommended as a way of supporting mental health and well-being in pregnant women, in particular those who demonstrate low well-being or symptoms of postnatal depression.

Full text

1
FancourtD, PerkinsR. BMJ Open 2018;8:e021251. doi:10.1136/bmjopen-2017-021251
Open access
Could listening to music during
pregnancy be protective against
postnatal depression and poor
wellbeing post birth? Longitudinal
associations from a preliminary
prospective cohort study
Daisy Fancourt,1,2 Rosie Perkins2,3
To cite: FancourtD, PerkinsR.
Could listening to music during
pregnancy be protective
against postnatal depression
and poor wellbeing post birth?
Longitudinal associations from
a preliminary prospective
cohort study. BMJ Open
2018;8:e021251. doi:10.1136/
bmjopen-2017-021251
►Prepublication history and
additional material for this
paper are available online. To
view these les, please visit
the journal online (http:// dx. doi.
org/ 10. 1136/ bmjopen- 2017-
021251).
Received 19 December 2017
Revised 24 May 2018
Accepted 15 June 2018
1Department of Behavioural
Science and Health, University
College London, London, UK
2Faculty of Medicine, Imperial
College London, London, UK
3Centre for Performance
Science, Royal College of Music,
London, UK
Correspondence to
Dr Daisy Fancourt;
d. fancourt@ ucl. ac. uk
Research
© Author(s) (or their
employer(s)) 2018. Re-use
permitted under CC BY.
Published by BMJ.
ABSTRACT
Objectives This study explored whether listening to music
during pregnancy is longitudinally associated with lower
symptoms of postnatal depression and higher well-being
in mothers post birth.
Design Prospective cohort study.
Participants We analysed data from 395 new mothers
aged over 18 who provided data in the third trimester of
pregnancy and 3 and 6 months later (0–3 and 4–6 months
post birth).
Primary and secondary outcome measures Postnatal
depression was measured using the Edinburgh Postnatal
Depression Scale, and well-being was measured using the
Short Warwick-Edinburgh Mental Well-being Scale. Our
exposure was listening to music and was categorised as
‘rarely; a couple of times a week; every day <1 hour; every
day 1–2 hours; every day 3–5 hours; every day 5+hrs’.
Multivariable linear regression analyses were carried out to
explore the effects of listening to music during pregnancy
on depression and well-being post birth, adjusted
for baseline mental health and potential confounding
variables.
Results Listening during pregnancy is associated with
higher levels of well-being (β=0.40, SE=0.15, 95% CI 0.10
to 0.70) and reduced symptoms of postnatal depression
(β=−0.39, SE=0.19, 95% CI −0.76 to −0.03) in the rst
3 months post birth. However, effects disappear by 4–6
months post birth. These results appear to be particularly
found among women with lower levels of well-being and
high levels of depression at baseline.
Conclusions Listening to music could be recommended
as a way of supporting mental health and well-being in
pregnant women, in particular those who demonstrate low
well-being or symptoms of postnatal depression.
INTRODUCTION
Perinatal mental health problems affect
around 20% of women at some point during
the perinatal period.1 In terms of conditions
characterised by negative symptomology,
postnatal depression (PND) is one of the
most common problems, and is a debilitating
condition with symptoms including fatigue,
irritability, insomnia and anhedonia; symp-
toms which in 25% of affected women last
for at least 1 year.2 Over the last two decades,
there has been significant research into the
effects of PND on mother and infant as well
as attention paid to how it can be prevented
or managed.3 4 However, in terms of condi-
tions relating more to the absence of positive
symptomology, such as low hedonic or eude-
monic well-being, there has been much less
research. The few studies that do exist have
found that negative mood, as indicated by
the Edinburgh Postnatal Depression Scale
(EPDS), has a correlation of just −0.46 with
positive experiences of motherhood, as indi-
cated by a principal component analysis of
six positive experiences of motherhood.5
This suggests that, as in the wider popula-
tion, depression and well-being are separate
constructs in the context of perinatal mental
Strengths and limitations of this study
►This preliminary prospective cohort study tracked a
sample of women across the perinatal period pro-
viding data at 12-week intervals.
►The data include a rich set of variables on music
listening behaviours among participants.
►We adjusted for all identied confounding variables
in our analyses and ran sensitivity analyses to test
our assumptions.
►The data are not nationally representative, although
there is a clear spread of participants from varying
socioeconomic backgrounds as well as variations in
the levels of exposure and outcome variables.
►As this is a cohort study and not interventional, it is
not possible to conrm causality.
on 17 July 2018 by guest. Protected by copyright.http://bmjopen.bmj.com/BMJ Open: first published as 10.1136/bmjopen-2017-021251 on 17 July 2018. Downloaded from

2FancourtD, PerkinsR. BMJ Open 2018;8:e021251. doi:10.1136/bmjopen-2017-021251
Open access
health and a positive perinatal experience is more than
simply the inversion of negative mood.6 Building on this,
women even without PND have been found to demon-
strate impairments in emotional problems and vitality,
suggesting that even in the absence of depression,
mothers can have impaired well-being.7 In light of this, a
review has argued that psychological well-being defined
as a multidimensional construct should be an integral
part of maternity care,8 and a more recent construct anal-
ysis has highlighted the importance not just of identifying
PND but also of identifying women with suboptimal peri-
natal well-being and supporting them to achieve positive
psychological functioning.9
In seeking to support the perinatal mental health of
women, the pregnancy period has been highlighted as
critical. Prenatal mental health has repeatedly been shown
to be one of the largest predictors of postnatal depres-
sion10–12 and well-being.9 In particular, the third trimester
of pregnancy has been identified as an important tran-
sition period involving adaptation to emotional and
physical changes, leading to feelings of well-being often
less pronounced than in the previous trimesters.13 Early
detection of symptoms of depression and low well-being
during pregnancy and prompt intervention is therefore
important in reducing adverse consequences.
In light of this, there are a number of interventions that
have been developed to try and support mental health in
the prenatal period as a way of reducing postnatal mental
health problems, in particular focusing on the third
trimester as a point of intervention. There have been find-
ings that support the application of cognitive-behavioural
and interpersonal psychotherapy, suggesting that depres-
sion following childbirth could be prevented by brief
interventions in the prenatal period.14 15 In exploring
other interventions, prenatal hypnotherapy has been
found to significantly reduce PND and improve psycho-
logical well-being at 2 weeks and 10 weeks postpartum.16
And a psychosocial intervention involving group meetings
to discuss aspects of parenthood in the final trimester of
pregnancy and first 6 months postpartum has been found
to reduce PND among first-time mothers.14 However, as
many mothers continue to work full time until shortly
before their due dates, in-person interventions may not be
feasible for all mothers and are of course limited by what
is available in different geographical areas. As a result,
there is a need to identify other home-based interven-
tions that could provide similar mental health support.
Over the past two decades, there has been increasing
research showing the effects of listening to music on
mental health. A number of reviews have demonstrated
the effects of regular music listening including in
enhancing mental health in the general population,17
reducing distress in premature infants18 and reducing
stress in adults.19 Specifically in relation to depression,
listening to music has been shown to reduce depression
among adults with chronic pain,20 psychiatric inpatients21
and older adults.17 22 In relation to well-being, music
listening has been shown to be associated with better
well-being not just in controlled interventions but also as
a result of ordinary day-to-day listening. A Swedish study
involving 500 older adults found associations between
music listening and well-being, even when controlling
for potential confounding variables.23 Studies tracking
daily activities have linked music listening with enhanced
well-being both in the workplace and in the wider context
of people’s lives.24 25 Further, music has also been shown
to contribute to creating supportive healthy environ-
ments, connecting individuals with their emotions and
promoting well-being.26 Finally, theoretical studies have
highlighted the role of music listening in enhancing
affect, wellness and resources for recovery and quality
of life.27 28 Consequently, both directed music listening
interventions and routine day-to-day music listening can
affect levels of depression and well-being in a range of
different populations.
Specifically in relation to the perinatal period, a
few studies have suggested that music listening may be
supportive for mental health. Listening to music for just
30 min has been found to reduce cortisol levels and anxiety
in pregnant women, leading to recommendations that
pregnant women might benefit from regular listening to
music as a practice of relaxation (although the effects of
regular listening were not tested).29 A recent study found
that women who listened to recorded music for 20 min a
day for 12 weeks during their pregnancy had significant
improvements in anxiety and depression.30 However, the
study did not track outcomes postnatally and involved a
small sample of women. And a further study has found
cross-sectional associations between listening to music
and depression and well-being among new mothers, with
more frequent listening associated with better mental
health.31 However, this study did not look longitudinally
nor involved pregnant women.
Therefore, to date, despite promising results suggesting
that listening to music can modulate mental health
and well-being during the perinatal period, no studies
have looked specifically at the impact of listening to
music during pregnancy on depression and well-being
post birth. In order to address this research gap, this study
tracked a cohort of mothers across the perinatal period
in order to ascertain whether there was a relationship
between music listening during pregnancy and postnatal
mental health.
METHODS
Participants and procedure
This study used data collected as part of a larger study
exploring the impact of creative interventions on peri-
natal mental health. Women living in England in the last
trimester of pregnancy (28 weeks or more) and the first
9 months postbirth (up to 40 weeks) were recruited from
hospitals, general practices, mother and baby charities
and through social media in England from October 2015
to March 2016, and completed an anonymous cross-sec-
tional online questionnaire. Women in the final trimester
on 17 July 2018 by guest. Protected by copyright.http://bmjopen.bmj.com/BMJ Open: first published as 10.1136/bmjopen-2017-021251 on 17 July 2018. Downloaded from

3
FancourtD, PerkinsR. BMJ Open 2018;8:e021251. doi:10.1136/bmjopen-2017-021251
Open access
of pregnancy (28 weeks or more) were then invited to
continue their participation in a longitudinal study.
This involved providing another wave of data 3 months,
6 months and 9 months following the first date of data
collection (which equated to providing baseline data T1
during pregnancy, T2 data in the first 3 months postbirth,
T3 data in months 4–6 postbirth and T4 data in months
7–9 postbirth). From an initial sample of 550 mothers who
consented to be involved in this longitudinal study, a total
of 458 mothers provided T2 data (83%), 417 provided
T3 data (75.8%) and 392 (71.3%) provided T4 data. The
study received ethical approval and all participants gave
informed consent prior to involvement in the research.
For this study, and in light of the literature review
presented above, we hypothesised that listening during
pregnancy would support well-being and reduce symp-
toms of PND in the first trimester postbirth. We there-
fore focused on women in the longitudinal study who had
provided complete data on the variables we selected for
analyses at both T1 and T2: 395 women. However, we also
ran some exploratory follow-up analyses with women who
had also provided complete data at T3 (n=299) in order
to explore if effects were maintained.
Patient public involvement
This study was developed as part of a wider grant that
involved mothers, psychiatrists and health workers in the
design of the research questions, the choice of measures
and the recruitment for the study. We also involved these
groups in the dissemination of the results.
Measures
Symptoms of PND were measured using the EPDS, a
10-item scale used extensively both with pregnant women
and new mothers, scored from 0 to 30 with 10+ indicative
of possible symptoms of depression and higher scores of
13+ indicating more severe depression.32
Well-being was measured using the Short Warwick-Edin-
burgh Mental Well-being Scale, a scale that encompasses
both hedonic and eudemonic well-being comprising
seven items scored from 7 to 35 with higher scores repre-
senting higher levels of well-being. The raw scores were
logit transformed prior to analysis.33 The New Economics
Foundation suggests five levels of well-being based
on quintile analyses of data in the UK Understanding
Society Survey, 2009: poor (<22), below average (22–24),
average (25–26), good (27–28) and excellent (29– 35).
In addition, demographic variables assessed the
women’s number of weeks pregnant/postbirth, number
of other children (0, 1, 2, 3 and 4+), household
income (<£16 000, £16 000–£30 000, £31 000–£60 000,
£61 000–£90 000, >£90 000), educational attainment
(school to 16, sixth form/college, undergraduate degree,
postgraduate degree), marital status (married vs not
married), employment status (working vs not working),
partner’s employment status (working vs not working)
and whether the woman had previously been diagnosed
with either anxiety or depression.
Listening to music was categorised as ‘rarely; a couple
of times a week; every day <1 hour; every day 1–2 hours;
every day 3–5 hours; every day 5+hrs’. While these anal-
yses focused on quantity of music listening as a predictor,
we also recorded genre of music listened to.
Statistics
Data were analysed using Stata V.14. Multivariable linear
regression models were used to explore the effects of
listening to music on well-being and PND. Frequency
of listening to music had a normal distribution, so was
treated as a 6-point linear variable, with higher score indi-
cating more frequent listening. For well-being and PND,
we used raw scores. Model 1 was unadjusted, while model
2 adjusted for baseline well-being/depression, mother’s
age, maternal education status, household income and
number of previous children, as well as how many weeks
the baby was post birth, the mother’s marital status at
T2, whether she was working at T1 or T2, whether her
partner was working at T2 and previous histories of both
anxiety and depression.
All models displayed linearity as assessed by
augmented partial residual plots with lowess smoothing;
multicollinearity as assessed by checking variance infla-
tion factors; normality as assessed using kernel density
plots, standardised normal probability (P-P) plots and
Q-Q plots; and there was no evidence of outliers or
undue influence as assessed using added variable plots
regressing each variable against all others, through stem
and leaf plots, and through assessing covariance ratios,
Cook’s distance and leverage. The well-being regression
models demonstrated homoscedasticity as assessed by
plotting the residuals versus fitted (predicted) values
and using the Breusch-Pagan test for heteroskedasticity.
However, the depression regression models showed
signs of heteroskedasticity, so robust standard errors
were calculated.
Planned sensitivity analyses were then conducted in
order to ascertain whether baseline mental health during
pregnancy was a moderator of the association between
listening to music and mental health postbirth. For this,
we included an interaction term of Q1 mental health and
Q1 listening habits in our regression models and then
plotted two-way contour graphs to visualise the interac-
tion (see online supplementary figure 1).
Although there were no significant demographic differ-
ences between those who did and did not provide data at
Q3, we wanted to take account of potential minor demo-
graphic differences between those who provided data at
Q3 and those who failed to. So the propensity score for
non-response was calculated using the indicators listed in
model 2 above (none of which significantly predicted miss-
ingness) and inverse probability weighting was applied to
the T3 regression models. We confirmed goodness of fit
using the Hosmer-Lemeshow test. Weighted analyses did
not differ from unweighted analyses.
on 17 July 2018 by guest. Protected by copyright.http://bmjopen.bmj.com/BMJ Open: first published as 10.1136/bmjopen-2017-021251 on 17 July 2018. Downloaded from

4FancourtD, PerkinsR. BMJ Open 2018;8:e021251. doi:10.1136/bmjopen-2017-021251
Open access
RESULTS
Demographics
At T1, women had an average age of 31.9 years (SD=4.9,
range 18–47) and an average of 32.9 weeks pregnant
(SD=4.1, range 28–42). Further demographics are
provided in table 1.
The average well-being score at T1 was 24.1 (SD=3.9,
range 11.25–35), at T2 was 23.9 (SD=4.3, range 7–35) and
at T3 was 23.8 (SD=4.1, range 7–35) (table 2). In order to
calculate the change in well-being among these women,
we analysed the difference in scores from T1 to T2 and
T1 to T3. From T1 to T2, 39.2% of mothers experienced
a decrease or at least one point in their well-being, while
30.4% experienced no change and 30.4% experienced an
improvement of at least one point. From T1 to T3, 41.4%
of mothers experienced a decrease or at least one point
in their well-being, while 25.7% experienced no change
and 32.9% experienced an improvement of at least one
point.
As with well-being, we calculated the change in symp-
toms of PND from T1 to T2 and T1 to T3 (table 2). From
T1 to T2, 43.6% of mothers experienced an increase
in the number of symptoms of PND they were experi-
encing, while 11.2% experienced no change and 45.3%
of mothers experienced an improvement in symptoms.
From T1 to T3, 56.4% of mothers experienced an increase
in the number of symptoms of PND they were experi-
encing, while 11.6% experienced no change and 43.6%
experienced an improvement in symptoms. In terms
of the interaction between well-being and symptoms
of PND, there was a large correlation between the two
at T1 (r=−0.67, p<0.001), T2 (r=−0.76, p<0.001) and T3
(r=−0.77, p<0.001), suggesting 45%-59% shared variance.
Regression results
Listening to music while pregnant was associated with
higher raw well-being scores 0–3 months postbirth, even
when accounting for potential confounding variables, with
greater frequency associated with greater effects (B=0.40,
SE=0.15, 95% CI 0.10 to 0.70) (see table 3). However, our
exploratory analyses showed that effects were not evident
4–6 months postbirth. There was also an association
between listening to music while pregnant and raw scores
of symptoms of PND, even when accounting for potential
confounding variables, with more frequent listening to
music during pregnancy associated with lower symptoms
of PND in the first 3 months postbirth (B=−0.39, SE=0.19,
95% CI −0.76 to −0.03). As with well-being, these results
were no longer evident by months 4–6 postbirth.
Further analyses
Sensitivity analyses of the well-being regression models
explored the potential moderating effect of mental
health during pregnancy. Our analyses showed there was
Table 1 Demographic information on participants
n=395
Maternal age, µ (SD) 31.9 (4.9)
Infant age, µ (SD) 32.9 (4.1)
Marital status, %
Married 69.3
Cohabiting 25.9
In a relationship but living separately 3.8
Single 1
Partner working, % 97.0
Educational attainment, %
Education to 16) 13.2
Education to 18 16.5
Undergraduate degree/qualication 41.3
Postgraduate degree/qualication 29.1
Household income, %
< £16 000 6.6
£16000–£30 000 11.1
£31000–£60 000 52.9
£61000–££90 000 17.5
>£91 000 11.9
Frequency of music listening, %
Rarely 5.6
A couple of times a week 17.2
Daily<1 hour 34.4
Daily 1–2 hours 29.9
Daily 3–5 hours 8.6
Daily 5+hours 4.3
Genre of music listened to, %
Jazz 21.0
Pop 93.7
Rock 57.7
Classical 34.2
Folk 22.8
R&B 42.8
Table 2 Levels of well-being and postnatal depression
during pregnancy (T1), 0–3 months postbirth (T2) and 4–6
months postbirth (T3)
T1 T2 T3
Well-being
Poor (<22) 29.4% 31.1% 31.6%
Below average (22–24) 25.6% 24.1% 25.4%
Average (25–26) 23.5% 21.3% 20.9%
Good (27–28) 10.1% 11.7% 13.7%
Excellent (29–35) 11.4% 11.9% 8.5%
Depression
EPDS<10 74.7% 72.2% 73.3%
EPDS>=10 25.3% 27.9% 26.7%
EPDS, Edinburgh Postnatal Depression Scale.
on 17 July 2018 by guest. Protected by copyright.http://bmjopen.bmj.com/BMJ Open: first published as 10.1136/bmjopen-2017-021251 on 17 July 2018. Downloaded from

5
FancourtD, PerkinsR. BMJ Open 2018;8:e021251. doi:10.1136/bmjopen-2017-021251
Open access
a significant interaction between listening and baseline
well-being (B=−0.10 SE=0.04, 95% CI −0.17 to −0.02), with
a two-way contour graph showing that listening to music
particularly seemed to support those with lower well-
being at baseline (see online supplementary figure 1A).
There was also a significant interaction between listening
and baseline depression (B=−0.09, SE=0.04, 95% CI −0.17
to −0.01), with a two-way contour graph showing that
listening to music particularly seemed to support those
who were showing symptoms of PND (evidenced with
a score of 10+ at baseline) (see online supplementary
figure 1B).
Finally, in order to try and ascertain whether listening
to music led to changes in mental health or whether
mental health led to changes in listening habits, we ran
additional analyses reversing the variable order. While
this does not confirm potential causal mechanisms,
it can give an indication as to whether mental health
can predict listening behaviours and therefore support
hypotheses about temporal precedence. There was no
evidence that levels of well-being during pregnancy were
associated with the likelihood of listening to music either
3 or 6 months postbirth. However, there was some indi-
cation that depression symptoms in the final trimester
of pregnancy were associated with listening habits 3
months postbirth (β=−0.03, SE=0.01, 95% CI −0.05 to
−0.01, p=0.003).
DISCUSSION
This study explored associations between listening to
music in the final trimester of pregnancy and mental
health and well-being in mothers postbirth. Listening was
found to be associated with higher levels of well-being
and reduced symptoms of PND in the first 3 months post-
birth, even when adjusting for baseline mental health and
potential confounding variables. These results appear to
be particularly found among women with lower levels
of well-being at baseline. These findings echo the few
existing studies in showing that listening to music is associ-
ated with better mental health in the perinatal period.30 31
However, to the authors’ knowledge, this is the first study
to show that listening to music during pregnancy is longi-
tudinally associated with better mental health postbirth.
Across both symptoms of PND and well-being, however,
associations were only found for the first 3 months post-
birth, and had disappeared by the second quartile post-
birth. The hypnosis study previously described also found
results within the first 3 months postpartum (weeks
2 and 10) but did not measure beyond this, so there is
little data available against which to benchmark these
findings.16 Nevertheless, the immediate period postbirth
has been highlighted as being of particular challenge for
new mothers, with the transition into assuming maternal
tasks and adjusting to the new role lasting until around
the third month postpartum.13 34 It is possible, therefore,
Table 3 Associations between listening to music during pregnancy on well-being and symptoms of postnatal depression
postbirth
Well-being Symptoms of PND
B SE 95% CI Pvalue B SE 95% CI P value
Months 0–3 postbirth (n=395)
Model 1 0.63 0.19 0.26 to 1.00 0.001 −0.52 0.23 −0.98to−0.06 0.028
R2=0.03, F(1,
393)=11.44, p=0.001
R2=0.01,
F(1,393)=4.74, p=0.03
Model 2 0.40 0.15 0.10 to 0.70 0.01 −0.39 0.19 −0.76to−0.03 0.036
R2=0.43,
F(18,376)=15.97,
p<0.001
R2=0.39,
F(18,376)=9.58,
p<0.001
Months 3–6 postbirth (n=299)
Model 1 0.33 0.21 −0.079 to 0.74 0.11 −0.11 0.24 −0.59 to 0.36 0.63
R2=0.01,
F(1,305)=2.17, p=0.14
R2=0.003,
F(1,301)=0.86, p=0.35
Model 2 0.12 0.16 −0.20 to 0.43 0.47 −0.02 0.20 −0.41 to 0.36 0.90
R2=0.44,
F(18,284)=11.22,
p<0.001
R2=0.37,
F(18,280)=7.19,
p<0.001
Model 1: unadjusted; model 2: adjusted for baseline well-being/depression, mother’s age, maternal education status, household income and
number of previous children, as well as how many weeks the baby was postbirth, and the mother’s marital status at T2, whether she was
working at T1 or T2 and whether her partner was working at T2 and previous histories of both anxiety and depression.
*P<0.05.
**P<0.01.
***P<0.001.
PND,postnatal depression.
on 17 July 2018 by guest. Protected by copyright.http://bmjopen.bmj.com/BMJ Open: first published as 10.1136/bmjopen-2017-021251 on 17 July 2018. Downloaded from

6FancourtD, PerkinsR. BMJ Open 2018;8:e021251. doi:10.1136/bmjopen-2017-021251
Open access
that any effects of music listening during the prenatal
period are of most value during this transition period,
but become less significant once mothers and their babies
become more settled.
A key question is how listening to music is associated
with better mental health and well-being in the postnatal
period. There are a number of potential explanations.
First, studies involving psychological tests, neuroimaging,
biomarker analyses and ethnographic observations have
shown that listening to music can have marked effects on
stress and anxiety.25 Specifically in relation to pregnant
women, listening to music for just 30 min can reduce
cortisol levels and anxiety.29 Wider studies involving
listening to music have shown it to be particularly effective
at reducing psychological and physiological responses to
stress, especially when people deliberately listen to music
in order to help them relax.35 36 This effect of music on
stress has in turn been linked specifically through to theo-
ries around well-being,25 37 with a wide literature linking
stress and anxiety with both mental health and well-
being.38 39 It is proposed that high levels of anxiety might
hinder women’s adaptation to motherhood in the initial
postpartum period, with negative effects on well-being.40
Consequently, it is possible that the relaxing effects of
listening to music during the pregnancy period help to
act as a buffer for feelings of stress and anxiety, thereby
supporting mothers in maintaining their adaptation and
leading to enhanced well-being.
Another potential explanation relates to the effects
of music on mood. Mood regulation has been iden-
tified as one of the prime reasons why people listen to
music, with models of mood regulation by music high-
lighting its effects on mood-related subjective experience
(including the intensity and clarity of moods), physiolog-
ical responses (such as energy levels and movement) and
behaviours (such as their ability to express emotions).41
Music listening has been found to modulate depression
and well-being.42 43 Early low mood during the prenatal
period is directly associated with lower well-being and
postnatal depression postbirth,44 leading to propositions
that interventions that deliberately attempt to cultivate
positive emotions, such as relaxation therapies and inter-
ventions focused on finding positive meaning, could
directly optimise health and well-being in this population.
Consequently, it is possible that another route by which
listening to music in the third trimester of pregnancy is
associated with improvements in mental health and well-
being is via enhancing mood.
Finally, a third explanation is that listening to music in
itself did not have an effect postbirth but did enhance
coping skills in women while they were still pregnant, which
in turn led to higher well-being postbirth. Music listening
has been linked with both problem-oriented coping and
emotion-oriented coping, specifically with results showing
that problem-oriented coping by music listening in women
is linked to lower depression levels.45 Life transitions (such as
the perinatal period) depend on both health and well-being
and also on appraisal and coping responses. In the hypnosis
study previously mentioned, the authors proposed that the
intervention during pregnancy helped mothers to maintain
and enhance their well-being while pregnant, which in turn
influenced their appraisal of the perinatal transition period
and supported their coping responses.16 It is possible that a
similar process took place through listening to music, with
listening to music supporting coping in the prenatal period,
which encouraged mothers’ own coping skills, which in
turn led to better coping during and postbirth and conse-
quent higher well-being. Indeed, even just in relation to the
birth, a number of birth preparation courses focus on relax-
ation and mood optimisation, which have been shown to
lead to less negative affect and better coping during labour
and delivery.46 Given that a significant predictor of PND is
the birth experience, enhanced coping prior to the birth,
perhaps through music listening, could be an important
factor in postnatal well-being.47
This study has a number of limitations. First, the study
followed a cohort of women rather than being interven-
tional, so it is not possible to confirm causality. However,
the study had a longitudinal design, there was no evidence
of reverse causality in relation to well-being, there are plau-
sible proposed mechanistic explanations and there is a
strong body of previous literature causally linking music and
mental health in other populations. So this study provides
promising preliminary evidence that remains to be tested in
a future experimental design. A second limitation is that the
population in this study was not nationally representative.
Nevertheless, there was a clear spread of participants from
varying socioeconomic backgrounds as well as variations in
the levels of exposure and outcome variables. So the data set
provides interesting and suitable preliminary data on the
longitudinal associations between music listening and
mental health. Third, this study explored the impact of all
music listening, not specifying particular genres. Previous
research has suggested that certain genres of music (or more
specifically compositional aspects of music such as its valence
and arousal levels) can lead to different responses, such as
variations in relaxation or mood.48 However, most of these
genre-specific effects have been found in tightly controlled
laboratory-based studies, and literature from real-world
interaction with music has suggested that musical prefer-
ence might be more important in determining the effects
of music.49 This study followed these real-world studies in
recording what genres people did listen to but measuring
the quantity of listening based on preference rather than
genre. Future studies could explore the impact of different
genres on mental health in the perinatal period.
In conclusion, this study provides the first preliminary
evidence that listening to music during the third trimester
of pregnancy could be protective against symptoms of
PND and low well-being in the first 3 months postbirth.
Music listening is an attractive intervention in that it is
readily available to people from all echelons of society
regardless of socioeconomic status, educational attain-
ment or cultural background. It can be carried out in a
range of contexts so is not restricted to particular places
or times. It is also inexpensive: indeed the majority of
on 17 July 2018 by guest. Protected by copyright.http://bmjopen.bmj.com/BMJ Open: first published as 10.1136/bmjopen-2017-021251 on 17 July 2018. Downloaded from

7
FancourtD, PerkinsR. BMJ Open 2018;8:e021251. doi:10.1136/bmjopen-2017-021251
Open access
women in the Western countries have access to recorded
music already. Finally, there are no obvious side effects
from listening to music. Consequently, listening to music
could be recommended as a way of supporting pregnant
women, in particular those who demonstrate low well-
being or symptoms of PND.
Acknowledgements The study team acknowledge the support of the National
Institute of Health Research Clinical Research Network (NIHR CRN). The authors
would like to thank the hospitals involved as Participant Identication Centres as
well as Miss Sunita Sharma, Prof Aaron Williamon and Sarah Yorke for their support
with the study.
Contributors DF and RP designed the study and collected data. DF ran the
analyses and drafted the paper. Both authors critically revised the manuscript and
approved it for submission.
Funding The study was funded by Arts Council England Research Grants
Fund, grant number 29230014 (Lottery). DF is supported by the Wellcome Trust
[205407/Z/16/Z].
Competing interests None declared.
Patient consent Not required.
Ethics approval UK NHS Research Ethics Committee.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement The data are available from the authors upon request.
Open access This is an open access article distributed in accordance with the
Creative Commons Attribution 4.0 Unported (CC BY 4.0) license, which permits
others to copy, redistribute, remix, transform and build upon this work for any
purpose, provided the original work is properly cited, a link to the licence is given,
and indication of whether changes were made. See: https:// creativecommons. org/
licenses/ by/ 4. 0/.
REFERENCES
1. Khalifeh H, Brauer R, Toulmin H, et al. Perinatal mental health: what
every neonatologist should know. Early Hum Dev 2015;91:649–53.
2. Morrell C, Warner R, Slade P, et al. Psychological interventions
for postnatal depression: cluster randomised trial and
economic evaluation. The PoNDER trial. Health Technol Assess
2009;13:1–153.
3. Cooper PJ, Murray L. Postnatal depression. BMJ 1998;316:1884–6.
4. Junge C, Garthus-Niegel S, Slinning K, et al. The impact of
perinatal depression on children's social-emotional development: a
longitudinal study. Matern Child Health J 2017;21:607–15.
5. Green JM, Kafetsios K. Positive experiences of early motherhood:
predictive variables from a longitudinal study. J Reprod Infant
Psychol 1997;15:141–57.
6. Keyes CL. Mental illness and/or mental health? Investigating axioms
of the complete state model of health. J Consult Clin Psychol
2005;73:539–48.
7. Boyce PM, Johnstone SJ, Hickey AR, et al. Functioning and well-
being at 24 weeks postpartum of women with postnatal depression.
Arch Womens Ment Health 2000;3:91–7.
8. Jomeen J. The importance of assessing psychological status during
pregnancy, childbirth and the postnatal period as a multidimensional
construct: a literature review. Clin Eff Nurs 2004;8:143–55.
9. Allan C, Carrick-Sen D, Martin CR. What is perinatal well-being? A
concept analysis and review of the literature. J Reprod Infant Psychol
2013;31:381–98.
10. Beck CT. Predictors of postpartum depression: an update. Nurs Res
2001;50:275–85.
11. Milgrom J, Gemmill AW, Bilszta JL, et al. Antenatal risk factors for
postnatal depression: a large prospective study. J Affect Disord
2008;108:147–57.
12. Robertson E, Grace S, Wallington T, et al. Antenatal risk factors for
postpartum depression: a synthesis of recent literature. Gen Hosp
Psychiatry 2004;26:289–95.
13. Leifer M. Psychological effects of motherhood : a study of rst
pregnancy / Myra Leifer. New York: Praeger, 1980.
14. Elliott SA, Leverton TJ, Sanjack M, et al. Promoting mental health
after childbirth: a controlled trial of primary prevention of postnatal
depression. Br J Clin Psychol 2000;39(Pt 3):223–41.
15. Spinelli MG. Prevention of postpartum mood disorders. In:
Postpartum mood disorders. Arlington, VA, USA: American
Psychiatric Association, 1999:217–35.
16. Guse T, Wissing M, Hartman W. The effect of a prenatal
hypnotherapeutic programme on postnatal maternal psychological
well‐being. J Reprod Infant Psychol 2006;24:163–77.
17. Chan MF, Wong ZY, Thayala NV. The effectiveness of music listening
in reducing depressive symptoms in adults: a systematic review.
Complement Ther Med 2011;19:332–48.
18. Cassidy JW, Standley JM. The effect of music listening on
physiological responses of premature infants in the NICU. J Music
Ther 1995;32:208–27.
19. Pelletier CL. The effect of music on decreasing arousal due to stress:
a meta-analysis. J Music Ther 2004:41:192–214.
20. Siedliecki SL, Good M. Effect of music on power, pain, depression
and disability. J Adv Nurs 2006;54:553–62.
21. Hsu WC, Lai HL. Effects of music on major depression in psychiatric
inpatients. Arch Psychiatr Nurs 2004;18:193–9.
22. Chan MF, Chan EA, Mok E, et al. Effect of music on depression levels
and physiological responses in community-based older adults. Int J
Ment Health Nurs 2009;18:285–94.
23. Laukka P. Uses of music and psychological well-being among the
elderly. J Happiness Stud 2007;8:215–41.
24. Haake AB. Individual music listening in workplace settings an
exploratory survey of ofces in the UK. Music Sci 2011;15:107–29.
25. Vastfjall D, Juslin PN, Hartig T. Music, subjective wellbeing, and
health: the role of everyday emotions. In: MacDonald R, Kreutz G,
Mitchell L, eds. Music, health, and wellbeing: Oxford University
Press, 2013.
26. McCaffrey R. Music listening: its effects in creating a healing
environment. J Psychosoc Nurs Ment Health Serv 2008;46:39–44.
27. Batt-Rawden KB, DeNora T, Ruud E. Music Listening and
Empowerment in Health Promotion: A Study of the Role and
Signicance of Music in Everyday Life of the Long-term Ill. Nordic
Journal of Music Therapy 2005;14:120–36.
28. Skånland MS. Everyday music listening and affect regulation: the role
of MP3 players. Int J Qual Stud Health Well-being 2013;8:20595.
29. Ventura T, Gomes MC, Carreira T. Cortisol and anxiety response to
a relaxing intervention on pregnant women awaiting amniocentesis.
Psychoneuroendocrinology 2012;37:148–56.
30. Nwebube C, Glover V, Stewart L. Prenatal listening to songs
composed for pregnancy and symptoms of anxiety and depression:
a pilot study. BMC Complement Altern Med 2017;17:256.
31. Fancourt D, Perkins R. Associations between singing to babies
and symptoms of postnatal depression, wellbeing, self-esteem and
mother-infant bond. Public Health 2017;145:149–52.
32. Cox JL, Holden JM, Sagovsky R. Detection of postnatal depression.
Development of the 10-item Edinburgh Postnatal Depression Scale.
Br J Psychiatry 1987;150:782–6.
33. Stewart-Brown S, Tennant A, Tennant R, et al. Internal construct
validity of the Warwick-Edinburgh Mental Well-being Scale
(WEMWBS): a rasch analysis using data from the scottish
health education population survey. Health Qual Life Outcomes
2009;7:15.
34. Pridham KF, Chang AS. Transition to being the mother of a new infant
in the rst 3 months: maternal problem solving and self-appraisals. J
Adv Nurs 1992;17:204–16.
35. Krout RE. The effects of single-session music therapy interventions
on the observed and self-reported levels of pain control, physical
comfort, and relaxation of hospice patients. Am J Hosp Palliat Care
2001;18:383–90.
36. Linnemann A, Strahler J, Nater UM. The stress-reducing effect
of music listening varies depending on the social context.
Psychoneuroendocrinology 2016;72:97–105.
37. Krout RE. Music listening to facilitate relaxation and promote
wellness: Integrated aspects of our neurophysiological responses to
music. Arts Psychother 2007;34:134–41.
38. Iwata M, Ota KT, Duman RS. The inammasome: pathways linking
psychological stress, depression, and systemic illnesses. Brain
Behav Immun 2013;31:105–14.
39. Vedhara K, Miles J, Bennett P, et al. An investigation into the
relationship between salivary cortisol, stress, anxiety and depression.
Biol Psychol 2003;62:89–96.
40. Pond EF, Kemp VH. A comparison between adolescent and adult
women on prenatal anxiety and self-condence. Matern Child Nurs J
1992;20:11–20.
41. Saarikallio S, Erkkilä J. The role of music in adolescents' mood
regulation. Psychol Music 2007;35:88–109.
42. Chin T, Rickard NS. Emotion regulation strategy mediates both
positive and negative relationships between music uses and well-
being. Psychol Music 2014;42:692–713.
on 17 July 2018 by guest. Protected by copyright.http://bmjopen.bmj.com/BMJ Open: first published as 10.1136/bmjopen-2017-021251 on 17 July 2018. Downloaded from

8FancourtD, PerkinsR. BMJ Open 2018;8:e021251. doi:10.1136/bmjopen-2017-021251
Open access
43. Lai YM. Effects of music listening on depressed women in Taiwan.
Issues Ment Health Nurs 1999;20:229–46.
44. Fleming AS, Ruble DN, Flett GL, et al. Postpartum adjustment in
rst-time mothers: Relations between mood, maternal attitudes, and
mother-infant interactions. Dev Psychol 1988;24:71–81.
45. Miranda D, Claes M. Music listening, coping, peer afliation and
depression in adolescence. Psychol Music 2009;37:215–33.
46. Duncan S J, Markman H. Intervention programs for the transition to
parenthood: current status from a prevention perspective, 2017.
47. Green JM. Expectations and experiences of pain in labor: ndings
from a large prospective study. Birth 1993;20:65–72.
48. Labbé E, Schmidt N, Babin J, et al. Coping with stress: the
effectiveness of different types of music. Appl Psychophysiol
Biofeedback 2007;32:163–8.
49. Fancourt D, Ockelford A, Belai A. The psychoneuroimmunological
effects of music: a systematic review and a new model. Brain Behav
Immun 2014;36:15–26.
on 17 July 2018 by guest. Protected by copyright.http://bmjopen.bmj.com/BMJ Open: first published as 10.1136/bmjopen-2017-021251 on 17 July 2018. Downloaded from