�007 vol. 31 no. 1 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH �7
Postnatal depression is a significant
public health issue in Australia as
elsewhere, affecting some one
in seven women.1-4 Untreated postnatal
depression may result in depressed mood
continuing for several years.5 As well as
directly affecting the health and well-being of
many mothers, it may also have implications
for the development of their children.6
Identification of depressed women by
primary health care providers tends to be
poor, with up to 50% of cases not being
recognised.7 There is also evidence that
women affected by postnatal depression
are themselves reluctant to seek help from
health care professionals.8,9 The Edinburgh
Postnatal Depression Scale (EPDS) is a
short (10 item) questionnaire developed
to assist in the early detection of postnatal
depression, given problems with somatic
items on most other depression scales that
complicate assessment of childbearing
women in the postnatal period.10 It is simple
to complete and was originally validated in a
study of women identified by health visitors
as probably depressed.10 Validation studies
have also been carried out in community
samples and in an Australian context.11,12 In
an overview of such studies, Shakespeare
notes that in each case the samples had
been recruited for research purposes and
that considerable variability existed in
the populations studied, the choice and
recruitment procedures for the samples
and the administration of the EPDS.13 She
suggests these factors may account for
the considerable variability in the positive
predictive power of the scale. The question of
how the EPDS performs when administered
Screening for postnatal depression:
not a simple task
Susan Armstrong and Rhonda SmallMother & Child Health Research, La Trobe University, Victoria
Abstract
Objective: To evaluate an established
screening program for postnatal
depression using the Edinburgh Postnatal
Depression Scale (EPDS) in a rural
Victorian shire. By protocol, all women
were screened at three time points post
delivery by maternal and child health
nurses. The efficacy of this approach
in detecting probable depression was
examined and referral pathways analysed.
Methods: Records for a 12-month cohort
of women giving birth in the shire were
audited (n=267). Information collected
included EPDS scores, parity, mother’s
age, reasons for non-completion, referral
details and nurses’ comments. Analysis
was completed using database and SPSS
programs.
Results: The process goal of screening on
all three occasions was rarely achieved – a
goal met for only 15.5% of women; 22%
were never screened at all. The highest
rate of screening was 50.6% at one month,
falling to 38.1% at eight months. Reasons
for non-screening varied, suggesting no
simple remedy. The proportions of women
identified as probably depressed at each
screening point (3.1%, 4.8% and 9.2%)
were considerably lower than statewide
figures for rural women. Referrals of
probably depressed women were mainly to
GPs but the results were unclear.
Conclusions: A well-established program
of universal screening was not effective in
detecting probable depression in women.
There was little evidence of direct feedback
from GPs about women referred as a
result of screening and no collaborative
planning for affected women. Before
universal screening of women for postnatal
depression can be recommended, better
evidence of its feasibility and acceptability
are required, alongside convincing
evidence that screening leads to improved
outcomes for women.
(Aust N Z J Public Health 2007; 31: 57-61)
Correspondence to: Ms Susan Armstrong, Mother & Child Health Research, La Trobe University, 324-328 Little Lonsdale Street, Melbourne, Victoria 3000. Fax: (03) 8341 8555; e-mail: [email protected]
Submitted: September 2006
Revision requested: November 2006
Accepted: December 2006
in routine primary care settings, rather
than in research studies, appears to remain
unanswered.
Statewide surveys in Victoria utilising
the EPDS have found prevalence rates of
depression of 15.4%, 16.9% and 13.9%.2-4
over the period 1989-2000. Yet in a study
published in 1994 from a rural region
of Gippsland, Griepsma et al. reported
that 57.8% of postnatal women scored as
probably depressed.14 This is an unusually
high prevalence and such rates have not
been replicated in other studies. The authors
of the Gippsland study speculated that
the reason for the high reported rate of
probable depression may have been the
high level of unemployment in the area
at the time. They also used a lower than
generally recommended cut-off for probable
depression on the EPDS. However, these
two factors seem unlikely to explain fully
the high proportion of women scoring as
depressed, and the reasons for the finding
remain unclear. Nevertheless, the study had
a significant effect on postnatal care practice
in the region. One of the recommendations
arising from the study was a proposal that
administration of the EPDS be introduced
as routine practice by maternal and child
health nurses in the region, as they were
seen to be in a unique position to screen for
probable depression and to refer women to
available support services. In response, one
local government shire in the Gippsland
region initiated such a screening protocol
for postnatal depression for all women
presenting at maternal and child health
centres. This involved three screening points
for mothers within an eight-month period
Article Health Promotion
�� AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH �007 vol. 31 no. 1
after a child’s birth. Although the program has been in place for
10 years, until now no formal evaluation has been undertaken.
The effectiveness of universal screening for postnatal depression
is of particular interest in the light of proposals for the use of the
EPDS to screen all recent mothers in Australia. In particular, as
part of the beyondblue National Depression Initiative it has now
been recommended that a national plan be developed to implement
routine antenatal and postnatal depression screening and that a
national steering committee be convened to oversee this.15 However,
in the United Kingdom, after a review of available evidence, the
National Screening Committee recommended against the use of
the EPDS as a screening tool because of concerns relating to its
performance in routine care.16 The existence of a longstanding
postnatal screening program in Gippsland has provided an
opportunity to examine how screening has operated within
primary care.
ObjectiveThe present study has been undertaken to audit and provide an
initial assessment of the screening process for recent mothers in
the rural shire in question, in order to add to the evidence base
about the use of the EPDS in a primary care setting. A secondary
objective has been to enable practitioners in the area to reflect
upon and improve practice. This study has been carried out by a
health services counsellor with the assistance of the maternal and
child health nurses (MCHNs) responsible for the implementation
of the screening program.
To assess the effectiveness of the screening program, the
following issues were addressed in the audit:
• Was the process objective of universal screening being
reasonably met?
• What proportion of women was actually being screened at each
time point?
• Was the screening actually identifying the number of potentially
depressed women expected?
• What referral and treatment options were offered to women
identified as being probably depressed?
• Were the women so identified referred appropriately?
In summary, has the screening program improved detection
and offered the chance of better outcomes for women affected by
postnatal depression?
MethodThe information collection phase of the study began with a
meeting with all the maternal and child health nurses working in
the shire in order to explain the purposes of the research and seek
their agreement and support. The screening protocol being used
was also clarified at this point, necessary since it had been largely
developed through working discussions over time.
The protocol for screening women was that the EPDS was to be
administered by the maternal and child health nurses to all women
attending MCHN centres with infants at one month, four months
and eight months of age. These screening points coincided with the
Healthy Futures Program designated by the Victorian Department
of Human Services for the monitoring of infant health as part of
the universal, widely accepted MCH service. At each of these
three time points any woman scoring ≥13 on the EPDS was to be
referred to her general practitioner for further assessment.
All available records for the cohort of women giving birth in the
shire from 1 July 2002 to 30 June 2003 were reviewed. Records
relating to mothers moving into the shire during the study period
were not included as they may not have had the opportunity to be
screened at all three time points.
To preserve confidentiality of individual women, the attendant
MCHN at each centre read out the relevant screening data to the
researcher, who recorded the results on a data collection form.
Each of the nurses had been working as the primary MCHN at
that centre during the audit period. They had been trained and were
experienced in the use of the EPDS as well as having undergone
wider training relating to postnatal depression. Ongoing peer
support for nurses was used in implementing the screening
program. All were committed to using the EPDS to identify
women who might be depressed. This commitment was clear in
group meetings with the nurses before and after the audit and
individually during information collection. The data collection
procedure had the added benefit that the nurses were able to
provide additional comments relating to the screening process
on a case-by-case basis. Nurses accessed relevant data from their
manual card index as well as the computer-based records system
in use at the centres.
Information collected included:
• The date of birth of the child (necessary to cross-check the
screening dates).
• The age of the mother.
• Number of children in the family designated by birth order of
the child.
• The woman’s score on the EPDS (if administered) at one, four
and eight months, and details of any referrals made.
• Reasons for screening not being completed, where this was
known.
• Any extra information provided by the nurse about the screening
process on a case-by-case basis at the time of the audit.
Table 1: Women screened at each time point.
Time of screen n % (n=257)One month 130 50.6
Four months 125 48.6
Eight months 98 38.1
Table 2: Adherence to protocol standard.
Screening n % (n=257)Screened once 89 34.5
Screened twice 72 28.0
Screened three times 40 15.5
Never screened 56 21.8
Armstrong and Small Article
�007 vol. 31 no. 7 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH ��
Data from the precoded data collection forms were entered
into an Access database and simple descriptive analyses were
undertaken using SPSS.17 Explanatory comments provided by
nurses were also coded and entered.
After the data had been analysed, a meeting was held with the
MCHNs to provide feedback and to give opportunity for discussion
and clarification. With their consent, this meeting was minuted.
ResultsPopulation sample
Data were collected on a total of 257 women of the 298 living
in the shire who had given birth to 304 babies (including six sets
of twins) during the year 1 July 2002 to 30 June 2003. No data
were collected for 41 women, most of whom had left the shire
at some stage after the birth of their babies and their records
transferred elsewhere. Data were obtained from four centres, three
of approximately the same size and one 60% smaller. The median
age of the women screened was 30, the age distribution similar to
statewide figures with two-thirds of childbearing women between
25 and 35 years of age.18
Proportion of completed screeningsTable 1 shows that the highest proportion of women screened
occurred at the four-week time point, when 50.6% of all women
were recorded as having completed an EPDS. This fell to 38% by
the eight-month screening point.
Table 2 indicates level of adherence to the full protocol, showing
that for only 15.5% of women was the screening protocol fulfilled
and 22% of women were never screened at all. Overall, of the total
of 771 planned screening points (three for each woman), screening
was actually carried out on just 46% of the intended occasions.
At the initial meeting when auditing the screening program
was discussed, some nurses believed that they were more likely
to screen first-time mothers. However, parity and age of the
mother were not significantly associated with the likelihood of
being screened. There was also no difference in screening rates
between the centres.
Reasons for screening not occurringReasons for non-completion of screening were available in 61%
of cases. These were obtained from the records and the nurses’
comments (retrospectively) about what had taken place. There
were multiple reasons for the EPDS not being given (see Table 3).
Some were logistical: a relieving maternal and child health nurse
was in attendance and was not familiar with the screening process
at that centre; someone other than the mother brought child to the Table 4: Women identified as probably depressed on screening and compared with a rural sample from a statewide Survey of Recent Mothers (SRM 2000).4
n/N (Number % OR completing (95% CI) EPDS)Score ≥13 at four weeks 4/130 3.1 0.2 (0.05-0.55)
Score ≥13 at four months 6/125 4.8 0.31 (0.11-0.75)
Score ≥13 at eight months 9/98 9.2 0.63 (0.28-1.37)
SRM 2000 – rural sample 62/449 13.8 1.00 Score ≥13
Table 3: Reasons for the EPDS not being completed.
Reason given n %Relieving health nurse did not administer EPDS 61 23.6
MCH nurse did not administer EPDS 58 22.6
Mother did not attend appointment 59 22.9
Mother refused to be screened 51 19.8
Enhanced Home Visiting Nurse allocated 26 10.1
centre; or the Enhanced Home Visiting Service nurse (EHVS) had
been allocated to the mother to provide a home-based service,
thus temporarily replacing the centre-based MCHN service, and
no screening data were available. However, on 42% of occasions
(when explanations were available) either the MCHN or the
mother made a decision not to do the screening. In addition, in
22% of cases where screening had not occurred, the reason was
non-attendance at the scheduled visit.
The nurses who had assisted with the data collection had all
been involved in delivering the screening program and knew many
of the women well. Their comments about individual screening
decisions were recorded, providing insight into further reasons
for non-completion.
“No need, mother was coping well.”
“Mother with a history of depression refused screening. Said she’d rather not know.”
“Mother cruising, back at work. Said she did not feel she needed to be screened.”
“With twin visits there is not enough time.” (Only one out of six mothers with twins completed all three screens)
“Mother always arrives late for appointments and is hard to engage.”
Women identified as probably depressedOf the 257 women for whom records were available, 201 were
screened at least once. Of these, 18 scored ≥13 on the EPDS,
indicating as probably depressed, with just one woman scoring
≥13 at more than one time point. Table 4 indicates the number
and proportion of women identified as probably depressed at each
screening point. Comparison with the proportion of rural women
identified as probably depressed on the EPDS at 5-6 months after
birth from a statewide survey is also shown.4 It can be seen that
fewer women were identified as depressed by the shire screening
program than might have been expected from the survey findings,
with lower odds of women identified as depressed at all time
points, and significantly so at the four weeks and four-month
time points.
Referral after screeningFifteen of the 18 women scoring ≥13 at any of the screening
points were referred to a GP, as the protocol required. Of the
others, one was referred to a counsellor, one to a neurologist and
one to a naturopath. Five were also referred to the Enhanced Home
Health Promotion Screening for postnatal depression
�0 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH �007 vol. 31 no. 1
Visiting Service nurse as a result of the screening result. Other
referral outcomes were generally not apparent. Nurses’ comments
indicated that some women scoring <13 were nonetheless
considered to be probably depressed, and were also referred to
their doctor with a view to receiving further assessment.
DiscussionIn the year studied, the screening program for postnatal
depression did not meet its first process objective of screening
all recent mothers three times in the first 12 months after giving
birth. Only 15.5% of women were screened according to the
protocol. Studies suggest that the screening at four weeks may not
be appropriate as women are more likely to be subject to transient
mood changes as they settle into life with a new baby. In this study,
the four-week screening achieved the highest proportion of women
screened but the lowest number of women scoring as probably
depressed. Fewer than half the women were screened at each of
the more critical four and eight-month time points, with only 63%
being screened at either. While screening for depression has been
advocated on the grounds of case finding and early identification of
distress, there is a need for caution when applied on a population
basis. A recent review of the evidence concluded that screening for
depression was not cost efficient and that screening programs have
been implemented without consideration of their effectiveness or
ethical and clinical implications.19 Critical appraisals of the use
of the EPDS for universal screening of mothers differ in their
conclusions. The only available study of an existing, primary care,
universal screening program utilising the EPDS was carried out by
Shakespeare, who conducted an audit of screening in Oxfordshire,
UK. Her finding was that completion rates were low; only 66%
of women were screened at eight weeks post birth and 55% at
eight months post birth. Low numbers of women were detected
as probably depressed – 7% at the first screening point and only
5% at the latter. She concluded that even within a locality with
a well-established strategy, a high level of screening was hard to
achieve.20 In one Australian study, Milgrom et al. concluded that
routine screening with the EPDS did integrate well into a primary
care setting administered by maternal and child health nurses.21
However, in this case screening occurred as part of a research
project and although the number of women participating and
completing the EPDS was estimated at 80%, this was based on
the report of the maternal and child health team leader with no
denominators given. Reporting on the acceptability to women of
screening with the EPDS in the beyondblue National Postnatal
Depression Program, Buist et al. stated that 85% of responders
to the postal survey sent to a random sample of the women who
had participated in screening had no discomfort with screening.22
However, no data are provided on the number of women who
refused the offer of screening and were therefore not subsequently
surveyed. Furthermore, women surveyed were asked to complete
the EPDS again, making it highly likely that the women who
had experienced discomfort on completing the EPDS during the
beyondblue program would be less likely to be among the 59%
of women who responded to the survey. All that can be concluded
with any confidence from this study is that 50% of a random
sample of women who had agreed to the beyondblue screening
process eventually identified themselves as having no discomfort
in completing that screening.
In the current study, the variety of reasons for non-screening
suggests that there is no one solution to low rates of screening.
Examining the nurses’ comments, it would seem that the decision
not to screen was sometimes made on the basis of their own
clinical assessments, which overrode the screening protocol.
This decision making may not always have been explicit but
involved information based on observations at the time, e.g.
“mother cruising”, “mother told me she was well”. It is possible
that better outcomes were achieved by such choices to override
the screening protocol for these women. It is also possible that
in these instances women who were depressed might have been
missed. In other cases, women who scored below the cut-off for
referral were still assessed as depressed and referred. Here the
nurses’ comments suggest that the completion of the EPDS may
have acted to facilitate discussion between the mother and the
nurse, and as a gateway to services. One in five women refused
the nurse’s offer of screening. Follow-up research will explore
why women make this choice. A qualitative research study that
sought women’s views on the acceptability of the EPDS concluded
that fewer than half of those women surveyed found the EPDS
acceptable, with most women preferring to talk about how they
felt rather than filling out a questionnaire.23
The low rates of screening and the low numbers of probably
depressed women detected and referred bring into question the
value of relying on a simple pen-and-paper test as the strategy
to identify and then assist women suffering from postnatal
depression. In real-life primary care settings, a range of factors
may undermine the integrity of universal screening programs. It
would seem that for a significant proportion of women, they, the
nurse, or both consider that screening is inappropriate. Whether for
this reason – or others such as non-attendance at screening visits
– it seems likely that a considerable number of depressed women
are not being identified by the screening program in this rural shire.
The strong suspicion arises that it is not the assessment tool itself
that is of most significance – the EPDS has been found to be very
acceptable to women in research studies – but rather the context in
which it is used and the salience that each party places upon the
interaction. If this is so, then other forms of clinical assessment,
such as more systematic case finding, may prove to be more
effective in detection and the provision of support and treatment
services, and ultimately in better outcomes for women.
The referral component of the audited screening program
operated much as had been intended, with 83% of women
identified as probably depressed referred to a GP and 28% to the
Enhanced Home Visiting Service. The low number of referrals
to other services is likely to reflect the relative deficiencies
of specialist services within this rural environment. However,
during data collection and in meetings the MCHNs indicated
dissatisfaction with the almost complete lack of feedback from
GPs resulting from referrals, which meant that the outcome of
Armstrong and Small Article
�007 vol. 31 no. 7 AUSTRALIAN AND NEW ZEALAND JOURNAL OF PUBLIC HEALTH �1
referral was often not known.
Difficulties with administration, completion rates and low
numbers of depressed women being detected may therefore not
be the only concerns in primary care screening. Screening for
postnatal depression is unlikely to lead to a successful referral
in the absence of acceptable services or a trusting relationship
between the woman and her care providers. In the beyondblue
National Postnatal Depression Program, more than 40,000 women
were screened with the EPDS antenatally and 12,000 women
screened postnatally, with more than 30% of those identified as
probably depressed ignoring the advice to seek help.15 In a New
Zealand study of antenatal screening using the EPDS with a
sample of 400 women (completion rate 92.5%, with 13.2% scoring
as probably depressed), “severe losses occurred subsequently”
when follow-up support was offered, and eventually only one
woman received treatment that she rated as valuable. The authors
concluded that “while the vast majority of pregnant women were
willing to complete a depression screening questionnaire, most
did not agree to additional contact for assessment, and either were
not offered treatment or did not accept treatment”.24 Further, a
systematic review of interventions to reduce depression after
birth concluded that none of the instruments available to screen
women antenatally performed sufficiently well for their use to be
recommended for identifying those women at risk of subsequent
postnatal depression.25
This well-established program of universal screening in one
rural shire was not effective in detecting probable depression in
women. There was little evidence of direct feedback from GPs
of women referred as a result of screening and no collaborative
planning for affected women. Before universal screening of women
for postnatal depression can be recommended, better evidence of
its feasibility and acceptability are required, alongside convincing
evidence that screening leads to improved outcomes for women.
Reasons underlying the problems identified with the screening
program reported in this paper are the subject of ongoing study.
Interviews with participating maternal and child health nurses and
with GPs have been conducted to investigate their views about
the screening program and the referral process, and women’s own
attitudes to screening are soon to be explored in a postal survey
and follow-up interviews. Some changes to the way the shire
screening program operates have also been made since the audit
of screening was completed. These include a common method
for recording women’s EPDS scores, integration of the generic
MCH and EHVS records and a change in the screening protocol
from three screening time points to two, completed with women
at two months and eight months after birth. The new screening
protocol has been in operation since the beginning of 2005 and a
second audit will shortly be undertaken for all women who gave
birth in that year.
AcknowledgementsThank you to all the maternal and child health nurses
who participated in the research for their open and willing
contributions.
References1. O’Hara M, Swain A. Rates and risk of postpartum depression — A meta-
analysis. Int Rev Psychiatry. 1996;8:37-54.2. Astbury J, Brown S, Lumley J, Small R. Birth events, birth experiences and
social differences in postnatal depression. Aust J Public Health. 1994;18(2):176-84.
3. Brown S, Lumley J. Physical health problems after childbirth and maternal depression at six to seven months postpartum. Br J Obstet Gynaecol. 2000;107(10):1194-201.
4. Brown S, Darcy M-A, Bruinsma F. Victorian Survey of Recent Mothers. Melbourne (AUST): Centre for the Study of Mothers’ and Children’s Health; 2001.
5. Williams H, Carmichael A. Depression in mothers in a multi-ethnic urban industrial municipality in Melbourne. Aetiological factors and effects on infants and preschool children. J Child Psychol Psychiatry. 1985;26(2):277-88.
6. Murray L. The impact of postnatal depression on infant development. J Child Psychol Psychiatry. 1992;33(3):543-61.
7. Hearn G, Iliff A, Jones I, Kirby A, Ormiston P, Parr P, et al. Postnatal depression in the community. Br J Gen Pract. 1998;48(428):1064-6.
8. Small R, Brown S, Lumley J, Astbury J. Missing voices: what women say and do about depression after childbirth. J Reproduction and Infant Psychology. 1994;12:89-103.
9. Whitton A, Warner R, Appleby L. The pathway to care in post-natal depression: women’s attitudes to post-natal depression and its treatment. Br J Gen Pract. 1996;46(408):427-8.
10. 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.
11. Carothers AD, Murray L. The validation of the Edinburgh Postnatal Depression Scale on a community sample. Br J Psychiatry. 1990;157:288-290.
12. Boyce P, Stubbs J, Todd A. The Edinburgh Postnatal Depression Scale: validation for an Australian sample. Aust N Z J Psychiatry. 1993;27(3):472-6.
13. Shakespeare J. Evaluation of Screening for Postnatal Depression against the NSC Handbook Criteria. London (UK): National Screening Committee; 2001 August 1.
14. Griepsma J, Marcollo J, Casey C, Cherry F, Vary E, Walton V. The incidence of postnatal depression in a rural area and the needs of affected women. Aust J Adv Nurs. 1994;11(4):19-23.
15. Buist A, Bilszta J. National Screening Program. In: The beyondblue National Depression Program. Prevention and Early Intervention 2001-2005 Final Report. Vol 1. Melbourne (AUST): beyondblue – The Depression Initiative; 2006.
16. Screening Specialist Library. National Screening Committee Policy Position – July 2006. London (UK): National Library for Health; 2006.
17. SPSS: statistical package for the social sciences [computer program]. Version 14.0. Chicago (IL): SPSS; 2005.
18. Victorian Perinatal Data Collection Unit. Births in Victoria 2003, 2004. Melbourne (AUST): Public Health, Department of Human Services, Government of Victoria; 2005 December.
19. Gilbody S, Sheldon T, Wessely S. Should we screen for depression? Br Med J. 2006;332(7548):1027-30.
20. Shakespeare J. Health visitor screening for postnatal depression using the Edinburgh Postnatal Depression Score: a process study. Community Pract. 2002;75:381-4.
21. Milgrom J, Ericksen J, Negri L, Gemmill AW. Screening for postnatal depression in routine primary care: properties of the Edinburgh Postnatal Depression Scale in an Australian sample. Aust N Z J Psychiatry. 2005;39(9):833-9.
22. Buist A, Condon J, Brooks J, Speelman C, Milgrom J, Hayes B, et al. Acceptability of routine screening for perinatal depression. J Affect Disord. 2006;93 233-7.
23. Shakespeare J, Blake F, Garcia J. A qualitative study of the acceptability of routine screening of postnatal women using the Edinburgh Postnatal Depression Scale. Br J Gen Pract. 2003;53(493):614-9.
24. Carter FA, Carter JD, Luty SE, Wilson DA, Frampton CM, Joyce PR. Screening and treatment for depression during pregnancy: a cautionary note. Aust N Z J Psychiatry. 2005;39(4):255-61.
25. Lumley J, Austin MP, Mitchell C. Intervening to reduce depression after birth: a systematic review of the randomized trials. Int J Technol Assess Health Care. 2004;20(2):128-44.
Health Promotion Screening for postnatal depression