Australian midwives knowledge, attitude and perceived learning needsaround perinatal mental health

Yvonne L. Hauck, BScN, MSc, PhD (Professor of Midwifery)a,b,n,Georgina Kelly, BNsg, PGDipMid, RN, RM (Research Midwife)b,Milan Dragovic, PhD (Senior Scientist – Neuropsychiatry)c, Janice Butt, RN, RM, ADA,PGCEA, MA(Ed), FACM (Co-ordinator Midwifery Education, Associate DirectorMidwifery)a,b, Pamela Whittaker, RMHN, RN (Clinical Nurse Manager)d,Johanna C. Badcock, BA, MA(Clin Psych), PhD (Research Professor, Senior Clinical Psychologist)c,e

a School of Nursing and Midwifery, Curtin University, Western Australia, Australiab Department of Nursing and Midwifery Education and Research, King Edward Memorial Hospital, Subiaco, Western Australia, Australiac Clinical Research Centre, North Metropolitan Health Service – Mental Health, Mount Claremont, Western Australia, Australiad Mother and Baby Unit, King Edward Memorial Hospital, Subiaco, Western Australia, Australiae School of Psychology, University of Western Australia, Western Australia, Australia

a r t i c l e i n f o

Article history:Received 26 February 2014Received in revised form26 August 2014Accepted 4 September 2014

Keywords:Perinatal mental healthKnowledgeLearning needsStigmaAttitudes

a b s t r a c t

Design and aim: a cross sectional survey was undertaken to explore midwives' knowledge of, andattitudes towards, mental health disorders in childbearing women vis-à-vis their perceived mentalhealth learning needs.Setting and participants: a 50.1% response rate included 238 midwives employed in the only publictertiary maternity hospital in Western Australia from March to June 2013.Method and findings: The survey comprised a mixture of custom-designed questions and vignettespresenting various disorders. Only 37.6% of midwives felt well-equipped to support women, whilst 50.2%reported insufficient access to information. Demand was highest for education on: personality disorders(77.8%); the impact of childbearing on mental health disorders (74.2%); and skills for handling stress andaggression (57.8%). Knowledge scores were variable: on average eight out of a maximum 13 questionswere answered correctly, but few (2.7%) answered more than 11 correctly, and 3.7% scored r4 correct.Across disorders, recognition from vignettes was highest for depression (93.9%), and lowest forschizophrenia (65.6%). Surprisingly, there were no associations between general knowledge scores andprevious mental health experience, recent professional development, or access to information aroundmental health. The majority endorsed positive beliefs about midwives' role in mental health assessment,and belief in women's recovery (83.5%), however, cluster analysis of warmth and competence ratingsrevealed negative stereotyping of mental health disorders.Key conclusions: Midwives accept it is their role to assess the mental health status of women but manyfeel ill-equipped to do so and express a strong desire for further knowledge and skills across a range ofperinatal mental health topics. Attitudes to recovery are positive but negative stereotypes exist; thereforeawareness of potential bias is important to negate their influence on care.Implications for practice: Learning needs may change due to trends in clinical practice. Strategies areneeded to recognise negative beliefs and to ensure education is responsive to local contexts.

& 2014 Elsevier Ltd. All rights reserved.

Introduction

In the Australian public health system, midwives play a majorrole in the care of childbearing women, having regular contact andopportunity to support and improve their mental well-being.Midwives are ideally placed to provide identification, referraland support for mental health disorders across the perinatal

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/midw

Midwifery

http://dx.doi.org/10.1016/j.midw.2014.09.0020266-6138/& 2014 Elsevier Ltd. All rights reserved.

n Corresponding author at: School of Nursing and Midwifery, Curtin University,GPO Box U1987, Perth, Western Australia 6845, Australia.

E-mail addresses: y.hauck@curtin.edu.au,yvonne.hauck@health.wa.gov.au (Y.L. Hauck),georgina.kelly@health.wa.gov.au (G. Kelly),milan.dragovic@health.wa.gov.au (M. Dragovic),janice.butt@health.wa.gov.au (J. Butt),pamela.whittaker@health.wa.gov.au (P. Whittaker),johanna.badcock@uwa.edu.au (J.C. Badcock).

Midwifery 31 (2015) 247–255

period. Therefore, to assist in the implementation of this support,an investigation of the perceived perinatal mental health learningneeds of Western Australian midwives was undertaken. To supple-ment this evaluation a brief assessment of midwives currentknowledge of, and attitudes towards, mental health disorderswas undertaken to inform how professional development educa-tion could better address existing needs.

Background

International studies, all from the United Kingdom (UK), havefound that although midwives are willing to learn and progresstheir perinatal mental health roles (Ross-Davie et al., 2006), themajority feel under-skilled to adequately manage mental healthdisorders (Stewart and Henshaw, 2002). In fact, a majority ofBritish midwives (94.3%) felt they had an important role in themanagement of mental health disorders during pregnancy andafter birth (Stewart and Henshaw, 2002). As such, UK midwiveshave expressed considerable concern regarding their confidenceand skills in caring for childbearing women with mental healthproblems. Importantly, these studies emphasise that midwivesoften perceive they have insufficient knowledge about availableresources related to mental illness during pregnancy (Jomeenet al., 2009; Rothera and Oates, 2011).

The need for prevention, early identification and prompt treat-ment is clear given the potential life threatening consequences ofmental health disorders during pregnancy (Sharp, 2009). Three UKEnquiries have highlighted suicide as a leading cause of maternaldeath, with more than half experiencing an underlying history ofmental health disorders (Lewis, 2001, 2004, 2007). The UK reviewof maternal deaths from psychiatric causes (2006–2008) suggeststhat: having a psychiatric disorder is common in pregnancy andafter birth; pregnancy is not protective against these disorders; andcare needs to be taken not to equate suicide risk with socio-economic deprivation. In addition, although previous history ofmental health disorders is important, 10 (34%) of 29 maternalsuicide events had no history and represented a first episode(Centre for Maternal and Child Enquiries, 2011).

In Australia, results from a recent National Health Surveyconfirmed that 13.6% of the population reported a mental orbehavioural condition (Australian Bureau of Statistics [ABS],2011). Additionally, the National Survey of Mental Health andWellbeing revealed 20% of Australians (16–85 years) had experi-enced disorders lasting at least 12 months (ABS, 2008). A WesternAustralian (WA) study with data from 1990 to 2005 also confirmedan increase in the prevalence of prior mental health disorders formothers (O’Donnell et al., 2013). This trend represented a 4.7%increase in odds per year of mental health service contacts in theyear preceding birth (O’Donnell et al., 2013).

Mood related conditions such as anxiety and depression areparticularly relevant for maternity clinicians, as women experiencehigher rates of these disorders (ABS, 2008). The prevalence ofantenatal depression in Australia is approximately 9%, whereas therate of depression for women between 1 and 12 months post birth isaround 15% (Beyond Blue, 2013). Specific rates for anxiety in child-bearing women are not currently available; however, many womenexperience anxiety and depression concurrently. Although less pre-valent, puerperal psychosis affects 1 to 2 women per 1000 and womenwith bipolar disorder are known to be at greater risk for psychosis(Beyond Blue, 2013). Similarly, schizophrenia is recognised as a lowprevalence disorder, with approximately 1% of Australians living withthe illness (Schizophrenia.com, 2010). However, in recent years thenumber of women with schizophrenia having babies has increased,leading to a call for specialist maternity services and staff to assist withtheir care (Vidog et al., 2012). Of note, these women also present with

increased rates of obstetric and neonatal complications necessitating acomprehensive approach to care (Nguyen et al., 2012). Some insightinto how to facilitate such care was recently obtained in a qualitativestudy of pregnancy in Australian women with an enduring mentalillness. This study found that these women valued building a relation-ship with a small known team of health professionals who couldprovide respect and understanding without stigma, while offering carethat acknowledged their special needs (Hauck et al., 2012). Despite theimportance of these issues, there is limited evidence on the knowledgeand attitudes of midwives to mental illness.

Australian midwives are involved in providing care to allchildbearing women, therefore it is essential they possess knowl-edge and skills to effectively care for vulnerable women (McCauleyet al., 2011). Australian research on midwives knowledge hasfocussed primarily on perinatal depression, the most prevalentdisorder (Buist et al., 2006; Miles, 2011; Jones et al., 2012). Whilsthaving regular exposure to women with mental health disorders,including depression, anxiety, bipolar, schizophrenia, and person-ality disorders, 93% of midwives in Victoria felt they could bebetter prepared to provide support to these women (McCauleyet al., 2011). Regular updates and study days offer one strategy toimprove midwives' mental health knowledge and history taking(Elliott et al., 2007; Ross-Davie et al., 2007). However, perceptionsof needs must also be assessed to ensure education is appropri-ately focussed.

Although existing evidence predominantly from the UK hasconfirmed that midwives would like more information aroundperinatal mental health (Jomeen et al., 2009; McCauley et al., 2011;Rothera and Oates, 2011; Stewart et al., 2002), further informationis necessary to inform what should be included within continuingprofessional development education in the Australian context.Perinatal depression has been the focus of existing research, dueto its higher prevalence; however, midwives must also supportwomen with low prevalence disorders. Targeted examination ofmidwives' learning needs is important to address this gap inknowledge and guide professional development education formidwives. Obtaining local evidence is essential for health servicesto be able to tailor professional development opportunities to theneeds of their workforce.

Study context

Western Australia is Australia's largest state with a total landarea of 2.5 million km2, a population of approximately 2.4 millionand 30,843 reported births in 2010 (Joyce and Hutchinson, 2012).The majority of the population (two million) reside in the Pertharea, where the study hospital is located. King Edward MemorialHospital (KEMH) is the state's only tertiary maternity hospital andreferral centre for women with highly complex pregnancies.KEMH had 5773 reported births in 2010. The Department ofNursing and Midwifery Education and Research at KEMH haveoffered a full day professional development session on mother-hood and mental health topics since 2004. This session is offeredat no cost to hospital employees. However, in recent years demandto attend this session has increased. Full enrolments and waitinglists are common and repeat sessions have been requested toaddress growing interest.

Method

Design and aim

A cross sectional study design was undertaken to explore theperceived perinatal mental health learning needs of midwivesworking at the only public tertiary maternity hospital setting in

Y.L. Hauck et al. / Midwifery 31 (2015) 247–255248

WA, along with a brief assessment of their current knowledge of,and attitudes towards, mental health disorders in childbearingwomen. This design was chosen to measure key variables in aselective sample at one point in time (Schneider et al., 2013).Convenience sampling was undertaken, whereby all midwivescurrently employed at the study hospital were invited to partici-pate (Liamputtong, 2010). Approval for the study was granted fromthe Women and Newborn's Human Ethics Committee (513QK).

Survey instrument

A customised survey instrument was specially designed for thecurrent study, in consultation with a team of mental healthprofessionals. Prior to commencement, the survey was pilotedwith seven midwives not employed in the study setting, whooffered feedback on the design, clarity, content and format of thequestions. Demographic and employment data was collected fromall participants, including: age, number of years worked as amidwife and current employment service area (i.e. antenatal,intrapartum, postnatal or rotating), along with information aboutprior experience working in a mental health setting, and/or anyrecent training in perinatal mental health.

Professional development

Midwives' professional development learning needs wereascertained from six items. Whether midwives felt well equippedto support women with mental health disorders, and had suffi-cient access to information on pregnancy and mental health wereeach assessed using a single item (yes/no response). However,midwives could then clarify their learning needs by selectingwhich if any particular disorders, topics, and skills they wantedmore education on. The disorders listed included: bipolar disorder,anxiety disorder, depression, schizophrenia, tocophobia, person-ality disorders and substance related disorders. Topics for furtherlearning included: signs and symptoms of mental illness, roles ofother health providers, understanding service options, child safety,and impact of childbearing on mental health. A range of advancedskills development was also offered, including: assessment ofmental health, communication skills, handling stress and aggres-sion, assessment of risk of mental illness and working withfamilies and carers. Finally, information was gathered on thepreferred learning format for professional development. An ‘other’option was provided for each question for further elaboration andinclusion of additional disorders, topics and skills. Alternately, theoption of ‘no additional knowledge or skills required’ could beselected.

Mental health knowledge

Knowledge around perinatal mental health was collected usingtwo complementary approaches: a general knowledge score and avignette knowledge score. General knowledge was assessed on thebasis of scoring agreement or disagreement with 13 questionsabout mental health disorders drawn from current clinical guide-lines of the study hospital (Women and Newborn Health Service,2011), and credible websites, such as BeyondBlue (http://www.beyondblue.org.au/) and PsychCentral (Psych Central, 2013, http://psychcentral.com). Each correct response was scored one (max-imum score¼13). Seven statements related to risks associatedwith mental illness such as personal and family history, abuse andtrauma, attachment, infant weight, hormonal protection andmaternal age. Two statements specifically addressed the Edin-burgh Post-natal Depression Scale (EPDS). Signs, symptoms andclinical management of mental illness were addressed in the fourremaining statements.

Midwives' knowledge of different mental disorders was alsoderived from their responses to four vignettes describing a child-bearing woman with: an anxiety disorder, depression, a manicepisode and schizophrenia. The vignettes did not represent anyactual person but were a composite of symptoms from theInternational Classification of Diseases (ICD-10) (WHO, 2008).A team comprising a mental health nurse, psychologist, psychia-trist and midwives developed the vignettes. Validation that eachvignette accurately reflected symptoms in ICD-10 was undertakenby a second (independent) psychiatrist. After reading the vignettemidwives were asked: What if anything, is wrong with the personin the vignette? Providing a correct written response was scoredone (i.e. maximum vignette knowledge score¼4).

Attitudes towards mental health

Attitudes towards mental health were also collected in twoseparate survey components. Questions addressed common beliefsand/or misconceptions about a midwife's role in handling perinatalmental health needs, such as whether it is their role to assessmental health status and whether mental health problems shouldalways be referred to a specialist. Other questions addressed beliefsabout womenwith mental health disorders, for example – that theyare likely to be difficult to manage or unlikely to recover, and shouldnot be encouraged to have children. The second approach toassessing midwives attitudes drew on recent studies of publicstigma towards mental health disorders (Sadler et al., 2012). TheStereotype Content Model (SCM) was used to examine how mid-wives judge a range of social groups including women with andwithout a mental health disorder (Fiske, 2012; Fiske et al., 2002,2006). This component was modelled on previous SCM studies andassessed ‘How different groups of childbearing women are per-ceived by midwives in Australia’ in terms of their competence/capability and warmth/friendliness. The wording of the instructionsemphasised ‘We are not asking how you personally view thesegroups, but how you think most midwives view them’ and isdesigned to reduce social desirability response bias. Participantsprovided ratings for nine target groups, including: childbearingwomen who have a household to run; a physical disability; a drugproblem; a business to run; bipolar disorder; anxiety; a housingproblem; depression; and schizophrenia. Ratings were made usinga 5-point scale (1¼not at all; 5¼extremely). An example item was‘how competent and capable are childbearing women who have adrug problem?’

Recruitment and data collection

There were approximately 475 midwives employed at KEMHduring the study period. Raosoft (2004), a survey sample sizecalculator, was used to determine the desired sample size to havesufficient statistical power to detect significant associations/clusters inthe data; 213 participants was recommended, allowing for a 5%margin of error and 95% confidence level. A research midwife visitedeach clinical area, provided general information about the studypurpose and, in consultation with midwifery managers, displayed aposter with study information, distributed surveys and a locked boxfor completed surveys. The survey cover provided information regard-ing study and completion was regarded as implied consent. Data wascollected over four months (March–June 2013).

Data analysis

Descriptive statistics (frequencies and percentages) wereundertaken for demographic data, attitude statements and profes-sional development learning needs. Associations between generalknowledge and vignette knowledge scores with perceptions of

Y.L. Hauck et al. / Midwifery 31 (2015) 247–255 249

being well-equipped to support women, having sufficient access toinformation, mental health experience, recent mental healthtraining, frequency of working with women with mental illnessand years of midwifery experience were examined using Pearsonχ2 with Yates's correction which makes the estimation moreconservative. Potential associations between perceptions of beingwell-equipped and having access to information with demo-graphic variables were also explored using IBM SPSS (v21) (SPSSInc, 2012) or Preacher (2001).

Analysis of attitudes (competence/capability and warmth/friendliness ratings) to each target group was based on thatroutinely employed in SCM studies (Fiske et al., 2002; Sadleret al., 2012). This comprised a two-step cluster analysis, toexamine the differences and commonalities in stereotypes acrossthe social groups tested. In order to determine the number ofclusters that best fit the current data, hierarchical cluster analysiswas used (Ward's method, which minimises within-cluster var-iance). Once the best fitting number of clusters was determined, k-means cluster analyses was conducted to determine which groupsbelong to which cluster (Blashfield and Aldenderfer, 1988).

Findings

Demographic and employment characteristics

The survey was completed by 238 midwives (50.1% response).The distribution of participants across each of nine age categorieswas reasonably broad (Table 1). The majority had been working asa midwife for more than a year, whilst almost 30% had more than

20 years of experience. Most were employed in positions thatrotated across different clinical areas rather than being focused ona single area of clinical practice. Importantly, the vast majority(almost 95%) reported they either sometimes or frequently hadcontact with childbearing women with a mental health disorder.In contrast, less than a quarter had previous experience working ina mental health (psychiatric) setting. Finally, nearly half indicatedthey had attended an education session on perinatal mental healthwithin the past two years.

Professional development

A summary of self-reported perinatal mental health learningneeds is presented in Table 2. Strikingly, only a minority felt wellequipped to support women with perinatal mental health dis-orders, with just over half also indicating they had insufficientaccess to information. Personality disorder was nominated mostfrequently as the mental disorder midwives wanted more educa-tion on. However, the following disorders were also cited by themajority of midwives: bipolar, substance related disorders, toco-phobia, schizophrenia and anxiety disorders. Only four midwivesindicated no further education around disorders was desired.

Table 2 reveals a particularly high level of demand (nearly 75%)for additional education on the impact of childbearing on mentalillness and a better understanding of the service options available;although no topic received fewer than 50% requests. There wassomewhat lower demand for skill development, but the most

Table 1Demographic variables.

Variable n %

Age (n¼235)20–25 16 6.826–30 33 14.031–35 24 10.236–40 28 11.941–45 30 12.846–50 32 13.651–55 38 16.256–60 19 8.160þ 15 6.4

Years worked as a midwife (n¼236)o1 12 5.11-2 22 9.33–5 40 16.96–10 33 14.011–15 30 12.716–20 29 12.321–25 21 8.926–30 19 8.130þ 30 12.7

Area of current work (n¼231)Rotate across different areas 133 57.6Intrapartum 38 16.5Postnatal 36 15.6Antenatal 24 10.4

Frequency of contact with childbearing women with mental illness(n¼236)Frequently 140 59.3Sometimes 84 35.6Rarely/never 12 5.1

Experience in mental health (psychiatric) setting (n¼235) 56 23.8

Attendance at education that has increased mental healthknowledge within past two years (n¼236)

111 47.0

Table 2Perinatal mental health learning needs.

Variable n %

Feels well equipped to support womenwith perinatal mental healthdisorders (n¼229)

86 37.6

Has sufficient access to information on pregnancy and perinatalmental health disorders (n¼225)

112 49.8

Mental health disorders – additional education desired (n¼234)Personality disorders 182 77.8Bipolar 153 65.4Substance related disorders 150 64.1Tocophobia 146 62.4Schizophrenia 145 62.0Anxiety disorders 138 59.0Depression 109 46.6Other 11 4.7

No further education required 4 1.7

Topics – additional education desired (n¼233)Impact of childbearing on mental illness 173 74.2Service options 170 73.0Signs and symptoms 142 60.9Roles of other health care professionals 138 59.2Child safety 128 54.9Other 7 3.0

No further education required 3 1.3

Skills – additional education desired (n¼232)Handling stress and aggression 134 57.8Assessment of mental health 133 57.3Assessment of risk of mental illness 132 56.9Clinical management 130 56.0Discharge planning 109 47.0Communication skills 107 46.1Working with families and carers 97 41.8Support with breast feeding 70 30.2Other 5 2.2

No further education required 1 0.4

Preferred educational format (n¼235)Online packages 143 60.9Seminar with guest speaker 137 58.3Series of seminars/workshops 134 57.0

Y.L. Hauck et al. / Midwifery 31 (2015) 247–255250

frequent requests were for skills in handling stress and aggression(57.8%); assessment of mental health (57.3%); assessment of risk ofmental illness (56.9%) and clinical management (56.0%). Only onemidwife suggested further education targeting skills developmentwas not required. Midwives expressed an interest in having perina-tal education offered in a variety of formats, including: online(60.9%), single seminars with a guest speaker (58.3%) or through aseries of seminar or workshops (57%).

Mental health knowledge

General knowledge statements and vignettes were scored ascorrect for a potential general knowledge score of 13 or amaximum vignette knowledge score of 4. Both the mode andmedian for the knowledge statements was 8 (range 0–13; mean8.27; SD 2.02). There were 3.7% (n¼8) who achieved a score r4,48.2% (n¼105) with a score between 5 and 8 and 2.7% (n¼6) whoachieved Z12 out of a potential 13. In terms of individualquestions, risk related items for mental health such as history ofabuse and trauma, and past personal/family history received thehighest number of correct responses (Table 3). Surprisingly, thetwo items related to knowledge about the EPDS were scoredcorrectly by less than 50% of participants. In addition, only 13.4%correctly identified the EPDS as a screening rather than a diag-nostic tool. Correctly acknowledging that women often need tocontinue psychiatric medication during pregnancy was the highestscored item under signs and symptoms and clinical management.

General knowledge scores (r8 and Z9) were comparedbetween midwives based upon whether they felt equipped tosupport women to those who did not (χ2¼0.026, df¼1, p¼0.871);those with previous mental health experience and those without(χ2¼0.033, df¼1, p¼0.855); those who attended mental healthtraining within the past two years and those who did not(χ2¼1778, df¼1, p¼0.182); those who perceived they had accessto information to those who did not (χ2¼0.7, df¼1, p¼0.402) andyears worked as a midwife (χ2¼0.11, df¼1, p¼0.736). Finally,although 49.2% of midwives who indicated they frequently workwith women with a mental health disorder achieved a score ofr8, this was comparable to the 55.7% who do not (χ2¼0.87, df¼1,p¼0.63).

There were 218 midwives who responded to all vignettesknowledge items (Fig. 1). The mode for correct responses was 3(range 0–4; mean 3.04; SD 1.015) with 74.8% (n¼163) achievinga correct total score of Z3, 16.5% (n¼36) scoring 2 and 8.7%

(n¼19) achieving r1. The depression vignette was most oftencorrectly recognised (93.9%) whilst schizophrenia was least accu-rately identified (65.6%). Vignettes depicting anxiety and a manicepisode or bipolar disorder gained intermediate rates of recogni-tion (74.8% versus 68.3% respectively).

Vignette knowledge scores (r2 and Z3) were comparedbetween midwives who felt equipped to support women to thosewho did not (χ2¼2.214, df¼1, p¼0.136); those with mental healthexperience and those without (χ2¼1.351, df¼1, p¼0.245); thosewho attended mental health training within the past two yearsand those who did not (χ2¼0.150, df¼1, p¼0.698); and those whoperceived they had access to information to those who did not(χ2¼0.455, df¼1, p¼0.499). In relation to clinical experience,67 out of 93 (72.0%) with 416 years of experience achieved ascore of Z3 and similarly 100 out of 133 (75.2%) who frequentlywork with women with a mental illness achieved a score of Z3.Differences were not statistically significant.

As expected, previous experience in a mental health settingwas associated with perceptions of being well equipped to supportwomen with a mental illness (χ2¼6.86; df¼1; p¼0.009). Feelingwell equipped to support women was also associated withcorrectly identifying the symptoms of bipolar disorder presentedin the vignette (χ2¼5.150, df¼1, p¼0.023). However, this associa-tion was not significant for the depression, schizophrenia oranxiety vignettes. No relationships were found between feelingequipped or having access to information and demographic vari-ables of age, years worked as a midwife, clinical area of work,frequency of contact with women with a mental health disorder,and recent attendance at continuing education.

Attitudes and mental health

The majority of midwives (87.7%) agreed it was within theirrole to assess the mental health status of women in their care, andless than half (42.7%) felt that mental health issues duringpregnancy should always be referred to a specialist. However,only 51.3% agreed that mental health problems were easy todetect. The majority of participants also endorsed the belief thatwomen with mental health disorders are likely to recover (83.5%)with only 19% suggesting that these women are likely to bedifficult to manage. Similarly, only 14% agreed with the statementthat women with a severe mental health disorder should not beencouraged to have children and markedly fewer (3.9%) agreed

Table 3General knowledge statements.

General knowledge statement (n, correct response) Correct response

n %

Related to riskA history of abuse and trauma can increase the risk of mental illness in women (n¼237; true) 213 89.5Past personal history of mental illness is a risk for mental illness during childbearing (n¼236; true) 211 88.7Past family history of mental illness is not a risk for mental illness during childbearing (n¼237; false) 204 85.7Hormones released during pregnancy protect against mental illness (n¼235; false) 196 82.4Women with mental illness may have attachment problems with their baby (n¼236; true) 136 57.1Younger mothers are at greater risk of postnatal depression (n¼235; false) 95 39.9Mental illness during pregnancy can influence birth weight of the baby (n¼236; false ) 49 20.6

Edinburgh Postnatal Depression ScaleThe cut off score for EPDS suggesting further assessment is 9 (n¼227; false) 103 43.3The EPDS is a useful diagnostic tool for depression and anxiety (n¼236; false) 32 13.4

Signs, symptoms and clinical managementWomen should not take psychiatric medication during pregnancy (n¼237; false) 195 81.9Postnatal depression will go away on its own but occasionally requires treatment (n¼237; false) 178 74.8Women must not breast feed if taking medication for mental illness (n¼236; false) 174 73.1Women with postnatal depression are sad and cry constantly (n¼236; false) 162 68.1

Y.L. Hauck et al. / Midwifery 31 (2015) 247–255 251

that they should not be allowed to keep their babies suggesting arange of positive attitudes to mental health disorders.

In contrast, analysis of the warmth/friendliness and compe-tence/capability ratings of the nine target groups indicated perva-sive negative stereotyping. Results from the cluster analysisyielded a clear, two-cluster solution, depicted in Fig. 2. Averageratings of competence/capability and warmth/friendliness withineach cluster were calculated across participants (Table 4). Clusterone represents those social groups that midwives perceived ashighly competent/capable and warm/friendly; these includedwomen with a physical disability and a household and/or businessto run. Cluster two represents all women with a mental healthdisorder plus those with housing and/or drug problems. Fig. 2indicates that attitudes to women with mental health disordersare more negative than to those without (including those with aphysical disability). Importantly, follow-up comparisons betweenclusters, using paired samples t-tests, confirmed that ratings ofwarmth/friendliness and competence/capability were significantlylower in Cluster 2 than Cluster 1.

Discussion

Our results reflect a snapshot of prevailing learning needs,knowledge and attitudes of Perth midwives from the only WAtertiary maternity hospital. Consequently generalisability to othercontexts may be limited. The overall response rate to the surveywas good (�50%), consistent with the evidence that midwivesexpress a growing need for professional development linked tomental health needs of childbearing women. The personalapproach to recruitment by the research midwife probablyassisted in achieving this outcome, as this factor has previouslybeen shown to boost response rates for paper surveys (Nulty,2008). Similarly, participants covered a relatively broad range ofages, experience and service setting; hence results are likely to bereasonably representative. A potential bias around who respondedto this survey may be that midwives who declined to participatehave more negative attitudes, were less willing to divulge theirnegative attitudes or were less confident about their knowledge.Consequently the current findings may represent a somewhatpositively biased picture.

Education must respond to shifting learning needs

The majority of midwives confirmed that it is their role toassess the mental health status of women in their care which mayaccount for strongly expressed needs for professional develop-ment. Only previous experience in a mental health setting wasassociated with perceptions of being well equipped to supportwomen despite nearly half of the participants indicating they hadattended an education session on perinatal mental health withinthe past two years. Recognition of being ill equipped may explainthe high level of motivation to learn and supports UK evidencecalling for more education around maternal mental health issues(Lees et al., 2009). Although evaluation of professional develop-ment programs around perinatal mental health are emerging,there is an obvious need to refine these programs based upon

Depression Vignette Anxiety Vignette Bipolar Vignette SchizophreniaIncorrect 6.1 25.2 31.7 34.3Correct 93.9 74.5 68.3 65.6

Fig. 1. Correct and incorrect responses for vignettes.

2.00

2.50

3.00

3.50

4.00

2.00 2.50 3.00 3.50 4.00

War

mth

Competence

Drug problems

Businessto run

Physical disability

Householdto run

Depression

Bipolar

Schizophrenia

AnxietyHousing problems

Fig. 2. Cluster solution for warmth and competence ratings.

Y.L. Hauck et al. / Midwifery 31 (2015) 247–255252

the perceived needs of midwives working in varying contexts(Stewart and Henshaw, 2002; Elliott et al., 2007; Ross-Davie et al.,2007). Education sessions previously attended by Australian mid-wives in this study may not have been appropriately targeted tothe perceived knowledge and skills desired as needs may changedue to trends in clinical practice. Therefore, education providersmust be responsive to needs within their local context.

A high level of requests for information around personalitydisorders warrants attention. Local mental health educationalsessions have traditionally focused upon high prevalence disorderssuch as depression and anxiety followed by bipolar and schizo-phrenia. The prevalence of personality disorders is around 9% inthe American population and is commonly co-occurring withmajor mental disorder likely to affect functioning (Lenzenwegeret al., 2007). Education sessions therefore need to adapt toemerging learning needs with a broader scope of informationacross diagnostic boundaries. A potential explanation for thisfinding may be that women with these diagnoses present morecomplex needs requiring treatment approaches that are lessfamiliar to midwives. In addition to knowledge about mentalhealth disorders, nearly three quarters of midwives wanted addi-tional education on the impact of childbearing on mental healthand a better understanding of available service options. What ismotivating the desire for information is unclear. Midwives may bepositively responding to the perceived needs of their clients andidentifying gaps in their knowledge. The desire for knowledgecoincides with an openness and flexibility in how the educationmay be offered with online packages, seminars with guest speak-ers and series of seminars or workshops all being positivelyregarded.

Knowledge alone does not reflect preparation to provide care

Midwives correct recognition of mental disorders from vign-ettes was highest for depression and lowest for schizophrenia but,as expected, better than found within the general community. AnAustralian study on mental health literacy for the general com-munity determined knowledge and beliefs about mental disordersand found 75% able to correctly identify depression and one thirdcorrectly identifying schizophrenia (Reavley and Jorm, 2012).Similarly, a Canadian study for depression literacy also found76% of the public could recognise depression from a case vignette(Wang et al., 2007). Unfortunately, 40% of the Australian samplesuggested that weakness of character was a likely cause for mentalillness (Reavley and Jorm, 2012) which was comparable to 43%from the Canadian sample (Wang et al., 2007).

Just below half of the midwives in this study had completedsome recent education on perinatal mental health although thisdid not appear to influence their current knowledge levels. Thisunexpected finding is difficult to explain. Although depression isthe most prevalent perinatal mental illness and midwives knowl-edge was highest in this area, there is room for improvement. Thefinding regarding interpretation of EPDS score as diagnostic ratherthan screening is concerning given midwives administer this toolduring pregnancy in WA public hospitals. In fact, the study settingclinical guidelines (Women and Newborn Health Service, 2011)

recommend an EPDS score of 12, not 9, for referral and furtherassessment. Having knowledge and attending mental healtheducation did not mean these WA midwives felt well equippedto support women. Almost two thirds of midwives in this study donot feel equipped to work with women with a mental healthdisorder which supports findings from a Victorian survey wheremidwives also felt ill equipped and unaware of available resources(McCauley et al., 2011). The challenge of working with womenwith serious mental illness during pregnancy has been recognisedand specialist skills and knowledge are recommended (McCauley-Elsom et al., 2009). In contrast, Jomeen et al. (2009) reported poorlevels of training and confidence among UK midwives but soundlevels of knowledge and illness perception for antenataldepression.

Bandura's theory suggests that self-efficacy (confidence) relatesto a person's perceptions of their ability or capabilities in perform-ing an activity (Bandura, 1986). Midwives' perceptions of not beingwell equipped to provide appropriate care to women with mentalillness may influence motivation, performance and feelings offrustration (Bandura, 1986). Working in a context where continuityof care is not usual practice means midwives are challenged informing a therapeutic relationship with women and may notreceive feedback on the effectiveness of their care. Feedback canoffer the opportunity for reflection including what Bandura (1986)refers to as cognitive simulations where midwives can visualisethemselves skilfully providing care to enhance subsequent perfor-mance. Under existing fragmented models of care, mastery of skillsis challenging and may contribute to a lower sense of self-efficacy(Bandura, 1986).

Midwives and stigma: reflecting on our attitudes

The majority of midwives in this study demonstrated positiveattitudes towards patient recovery, at least when asked directly.The outcome has similarities to community beliefs that fullrecovery is possible for people receiving appropriate professionalhelp for most disorders, aside from schizophrenia (Reavley andJorm, 2012). These positive attitudes toward recovery contrastwith the findings from the current cluster analysis, which showedthat women with a mental health disorder were judged as beinglow in warmth and competence. This pattern of results suggeststhe influence of social desirability bias when attitudes to mentalillness are explicitly assessed. The conflict between explicit andimplicit bias is particularly important as biases are often uncon-scious and occur despite best intentions; however, they caninfluence decisions and interactions associated with discrimina-tion and disparities in health care (Dovidio and Fishe, 2012).

Stigmatising attitudes can differ according to disorders(Reavley and Jorm, 2011). Stigma around mental illness, particu-larly schizophrenia, is pervasive – not only in the general com-munity but amongst [mental] health professionals (Hocking,2003). For example, the medical profession has been noted todemonstrate discriminatory behaviour towards people with men-tal illness suffering from cardiovascular disease. Moreover, thisbehaviour was associated with earlier than expected mortality(Coglan et al., 2001). In addition, nursing evidence suggests that

Table 4Warmth and competence mean scores for two clusters.

Social groups Stereotype dimensions r t p-Value

Warmth Competence

Cluster 1 Disability, household, business 3.48 3.56 0.73 2.45 0.015Cluster 2 Bipolar, anxiety, housing, depressions, schizophrenia drug problems 2.72 2.65 0.87 1.84 0.005

Y.L. Hauck et al. / Midwifery 31 (2015) 247–255 253

lack of knowledge, skills and training with respect to mentalhealth disorders plus workload issues have been associated withnegative attitudes (van der Kluit and Goossens, 2011). Rouch(2012) recommends increasing awareness and knowledge asimportant to addressing biases in our attitudes. A recent Australiansurvey found that characteristics of stigmatising attitudes inhealth professionals were comparable to the general community(Reavley et al., 2013); although there were greater beliefs indangerousness and personal weakness noted by females andgeneral practitioners.

A recent investigation into Australian consumer experiencesconfirmed stigmatising attitudes and behaviour from health pro-fessionals and/or mental health service providers – with almost45% indicating their service provider changed their behaviour oncelearning of their mental health disorder (Mental Health Council ofAustralia, 2011). Just under 30% of consumers felt professionalswere not comfortable talking to them, and over half (51%) worriedthat professionals perception of them was unfavourable (MentalHealth Council of Australia, 2011). Nonetheless, 55% reported thattheir health professional was optimistic toward their recovery andencouraging (59%) for the future.

Implications and conclusion

WA midwives confirmed agreement that their role includesassessment of the mental health status of women in their care.Although feeling ill-equipped they demonstrated a strong desirefor information across perinatal mental health topics. The desirefor knowledge coincided with an openness and flexibility in theformat of professional educational opportunities. Perceptions ofother midwives attitudes to womenwith mental illness were morenegative. Awareness of potential bias is important to negate anydetrimental influence on the care provided to childbearingwomen. Finally, we must acknowledge that learning needs mayvary due to exposure to mental illness and emotional challengesmore broadly, as well as changing trends in clinical practice andtherefore ongoing strategies must ensure that continuing profes-sional education is responsive to the local context.

Conflict of interest statement

None of the authors had any conflicts of interest.

Acknowledgements

Funding support was received from King Edward MemorialHospital and the Clinical Research Centre in North MetropolitanHealth Service – Mental Health.

References

Australian Bureau of Statistics [ABS], 2011. Australian Health Survey: First Results2011–12. 4364.0.55.001. Commonwealth of Australia, Canberra, ACT.

Australian Bureau of Statistics [ABS], 2008. National Survey of Mental Health andWellbeing: Summary of Results, 2007. 4326.0. Commonwealth of Australia,Canberra, ACT.

Bandura, A., 1986. Social Foundations of Thought and Action: A Social CognitiveTheory. Prentice-Hall, Englewood Cliffs, NJ.

Beyond Blue, 2013. Beyond Blue – Depression. Anxiety. Accessed December 2013from ⟨http://www.beyondblue.org.au/⟩.

Blashfield, R., Aldenderfer, M., 1988. The methods and problems of cluster analysis.In: Nesselroade, J., Cattell, R. (Eds.), Handbook of Multivariate ExperimentalPsychology, 2nd edn. Plenum Press, New York, pp. 447–473.

Buist, A., Bilszta, J., Milgrom, J., Barnett, B., Hayes, B., Austin, M., 2006. Healthprofessional's knowledge and awareness of perinatal depression: results of anational survey. Women Birth 19, 11–16.

Coglan, R., Lawrence, D., Jablensky, A., 2001. Duty to Care: Physical Illness in Peoplewith Mental Illness. University of Western Australia, Perth, WA.

Dovidio, J., Fishe, S., 2012. Under the radar: how unexamined biases in decision-making processes in clinical interactions can contribute to health care dispa-rities. Am. J. Public Health 102, 945–952.

Elliott, S., Ross-Davie, M., Sarkar, A., Green, L., 2007. Detection and initial assess-ment of mental disorder: the midwife's role. Br. J. Midwifery 15, 759–764.

Fiske, S., 2012. Warmth and competence: stereotype content issues for cliniciansand researchers. Can. Psychol. 53, 14–20.

Fiske, S., Cuddy, A., Glick, P., 2006. Universal dimensions of social perception:warmth and competence. Trends Cogn. Sci. 11, 77–83.

Fiske, S., Cuddy, A., Glick, P., Xu, J., 2002. A model of (often mixed) stereotypecontent: competence and warmth respectively follow from perceived statusand competition. J. Personal. Soc. Psychol. 82, 878–902.

Hauck, Y., Allen, S., Ronchi, F., Faulkner, D., Frayne, J., Nguyen, T., 2012. Pregnancyexperiences of Western Australian women attending a Specialist Childbirth andMental Illness Antenatal Clinic. Health Care Women Int. 34, 380–394.

Hocking, B., 2003. Reducing mental illness stigma and discrimination: everybody'sbusiness. Schizophrenia 178, S47–S48.

Jomeen, J., Glover, L., Davies, S., 2009. Midwives' illness perceptions of antenataldepression. Br. J. Midwifery 17, 296–303.

Joyce, A., Hutchinson, M., 2012. Western Australia's Mothers and Babies, 2010:Twenty-Eighth Annual Report of the Western Australian Midwives' NotificationSystem. Department of Health, Perth, WA.

Jones, C., Creedy, D., Gamble, J., 2012. Australian midwives' awareness and manage-ment of antenatal and postpartum depression. Women Birth 25, 23–28.

Lewis, G. (Ed.), 2001. Why Mothers Die 1997–1999. The Fifth Report of theConfidential Enquiries into Maternal Deaths in the United Kingdom. RCOGPress, London.

Lewis, G. (Ed.), 2004. Why Mothers Die 2000–2002. The Confidential Enquiry intoMaternal and Child Health. Report on Confidential Enquiries into MaternalDeaths in the United Kingdom. RCOG Press, London.

Lewis, G. (Ed.), 2007. The Confidential Enquiry into Maternal and Child Health(CEMACH). Saving Mothers' Lives: Reviewing Maternal Deaths to MakeMotherhood Safer – 2003–2005. The Seventh Report on Confidential Enquiriesinto Maternal Deaths in the United Kingdom. CEMACH, London.

Centre for Maternal and Child Enquiries (CMACE), 2011. Saving mothers' lives:reviewing maternal deaths to make motherhood safer: 2006–08. The eighthreport on confidential enquiries into maternal deaths in the United Kingdom.Br. J. Obstet. Gynaecol. 118, 1–203.

Lees, S., Mills, N., McCalmont, C., 2009. Professionals' knowledge of perinatalmental health care. Ment. Health Pract. 13, 24–27.

Lenzenweger, M., Lane, M., Loranger, A., Kessler, R., 2007. DSM-IV personalitydisorders in the national comorbidity survey replication. Biol. Psychiatry 62,553–564.

Research Methods in Health: Foundations for Evidence-Based Practice. In: Liam-puttong, P. (Ed.), 2010. Oxford University Press, South Melbourne, VIC.

McCauley, K., Elsom, S., Muir-Cochrane, E., Lyneham, J., 2011. Midwives andassessment of perinatal mental health. J. Psychiatr. Ment. Health Nurs. 18,786–795.

McCauley-Elsom, K., Cross, W., Kulkarni, J., 2009. Best practice when working withwomen with serious mental illness in pregnancy. Ment. Health Learn. Disabil.Res. Pract. 6, 185–203.

Mental Health Council of Australia (MHCA), 2011. Consumer and Carer Experiencesof Stigma from Mental Health and Other Health Professionals. MHCA, Canberra,ACT.

Miles, S., 2011. Winning the battle: a review of postnatal depression. Br. J.Midwifery 19, 22–27.

Nguyen, T., Faulkner, D., Frayne, J., et al., 2012. The obstetric and neonatal outcomesof pregnant women with serious mental illness attending a specialist antenatalclinic. Med. J. Aust. 1, 26–29.

Nulty, D., 2008. The adequacy of response rates to online and paper surveys: whatcan be done? Assess. Eval. High. Educ. 33, 301–314.

O’Donnell, M., Anderson, D., Morgan, V., Nassar, N., Leonard, H., Stanley, F., 2013.Trends in pre-existing mental health disorders among parents of infants bornin Western Australia from 1990 to 2005. Med. J. Aust. 198, 485–488.

Preacher, K., 2001. Calculation for the Chi-Test: An Interactive Calculate Tool for Chi-Square Tests of Goodness of Fit and Independence [Computer Software].Accessed November 2013 from ⟨http://www.quantpsy.org/chisq/chisq.htm⟩.

Psych Central, 2013. Psych Central Learn Share Grow. Accessed from ⟨http://psychcentral.com/⟩.

Raosoft Inc., 2004. Sample Size Calculator. Accessed March 2013 ⟨http://www.raosoft.com/samplesize.html⟩.

Reavley, N., Jorm, A., 2012. Mental Health Literacy in Australia: Is It Getting Better?The 2011 National Survey of Mental Health Literacy and Stigma. University ofMelbourne, Melbourne, VIC.

Reavley, N., Jorm, A., 2011. Young people's stigmatizing attitudes towards peoplewith mental disorders: findings from an Australian national survey. Aust. N. Z. J.Psychiatry 45, 947–956.

Reavley, N., Mackinnon, A., Morgan, A., Jorm, A., 2013. Stigmatizing attitudestowards people with mental disorders: a comparison of Australian healthprofessionals with the general community. Aust. N. Z. J. Psychiatry 48, 433–441,http://dx.doi.org/10.1177/0004867413500351.

Ross-Davie, M., Elliott, S., Green, L., 2007. Planning and implementing mentalhealth training. Br. J. Midwifery 15, 199–203.

Ross-Davie, M., Elliott, S., Sarkar, A., Green, L., 2006. A public health role in perinatalmental health: are midwives ready? Br. J. Midwifery 14, 330–334.

Y.L. Hauck et al. / Midwifery 31 (2015) 247–255254

Rothera, I., Oates, M., 2011. Managing perinatal mental health: a survey ofpractitioners' views. Br. J. Midwifery 19, 34–42.

Rouch, K., 2012. Examining our biases about mental illness. Am. J. Nurs. 112, 7.Sadler, M., Meagor, E., Kaye, K., 2012. Stereotypes of mental disorders differ in

competence and warmth. Soc. Sci. Med. 74, 915–922.Schizophrenia.com, 2010. Schizophrenia Facts and Statistics. The Internet Mental

Health Initiative. Accessed December 2013 from ⟨www.schizophrenia.com⟩.Schneider, Z., Whitehead, D., LoBiondo-Wood, G., Haber, J., 2013. Nursing and

Midwifery Research: Methods and Appraisal for Evidence-Based Practice, 4thedn. Mosby, Sydney, NSW.

Sharp, D., 2009. Perinatal mental health. In: Gask, L., Lester, H., T Kendrick, T., RPeveler, R. (Eds.), Primary Care Mental Health. Royal College of Psychiatrists,London, London, pp. 198–210.

SPSS Inc. Released 2012. IBM SPSS Statistics for Windows, Version 21.0. IBM Corp,Armonk, NY.

Stewart, C., Henshaw, C., 2002. Midwives and perinatal mental health. Br. J.Midwifery 10, 117–121.

van der Kluit, M., Goossens, P., 2011. Factors influencing attitudes of nurses ingeneral health care toward patients with comorbid mental illness: an inte-grative literature review. Issues Ment. Health Nurs. 32, 519–527.

Vidog, S., Seeman, M., Ray, J., et al., 2012. Temporal trends in general and age-specific fertility rates among women with schizophrenia (1996–2009): apopulation-based study in Ontario, Canada. Schizophr. Res. 139, 169–175.

Wang, J., Adair, C., Fick, G., et al., 2007. Depression literacy in Alberta: findings froma general population sample. Can. J. Psychiatry 52, 442–449.

Women and Newborn Health Service, 2011. King Edward Memorial HospitalClinical Guidelines Antepartum Visits. WNHS, Perth, WA.

World Health Organization [WHO], 2008. ICD-10: International Statistical Classifi-cation of Diseases and Related Health Problems, 10th rev. edn. WHO, New York.

Y.L. Hauck et al. / Midwifery 31 (2015) 247–255 255