Screening for postpartum anxiety: A quality improvement project to promote the screening of women suffering in silence

Abstract

Background:

Postpartum anxiety is a mental health problem that has largely been ignored by maternity care providers despite an estimated incidence as high as 28.9%. Though postpartum anxiety may or may not be accompanied by depression, and while screening for postpartum depression has become more common place, postpartum anxiety is often not assessed or addressed.

Purpose:

The purpose of this pilot quality improvement project was to implement a screening, treatment and referral program for postpartum anxiety in the birth centre environment.

Procedures:

Midwives from 10 geographically diverse birth centres, and all members of the American Association of Birth Centres, were recruited to participate in the project. An online video was developed which detailed postpartum anxiety, screening through use of the anxiety subscale of the Edinburgh Postnatal Depression Scale and a toolkit for treatment and/or referral for screen positive patients. Participants entered patient scores into the Perinatal Data Registry of the American Association of Birth Centres. Individual interviews of midwives were conducted following the 10-week pilot period.

Main Findings:

There were a total of 387 participants across 9 participating sites. Among all screened participants with follow-up data, (n = 382), 9.69% (n = 37) were lost to follow-up. Among all participants screened with the Edinburgh Postpartum Depression Scale −3A and Edinburgh Postpartum Depression Scale (n = 318), 12.58% (n = 40) had a positive Edinburgh Postpartum Depression Scale −3A score of greater than six. Of all screened participants with an Edinburgh Postpartum Depression Scale score, 15 (6.98%) had a Edinburgh Postpartum Depression Scale score of less than 12 and an Edinburgh Postpartum Depression Scale −3A score greater than six, and would have not received follow up care if only screened for postpartum depression. Midwife participants expressed heightened awareness of the need to screen and felt screening was easy to integrate into clinical practice.

Conclusions:

The Edinburgh Postpartum Depression Scale −3A is a valid, easy-to-use tool which should be considered for use in clinical practice. Modification of the electronic health record can serve as an important impetus triggering screening and treatment. It is important that clinicians are educated on the prevalence of postpartum anxiety, its risk factors, symptoms and implications.

Introduction

Postpartum depression is the primary maternal mental health condition discussed by women’s health providers. It has also received the most attention in the media and academic research. In the shadow of postpartum depression (PPD), the experiences of women with postpartum anxiety (PPA) have been largely ignored. Distinct from other anxiety disorders, PPA occurs within the first 12 months postpartum, causing women to experience worries and fears that are extreme and usually focused on the health and safety of the infant (Postpartum Support International, 2016).

Anxiety and depression may exist as comorbid conditions, so anxiety symptoms in postpartum women tend to be labelled as depression, even though anxiety may present in the absence of depression (Miller et al., 2006). The first postpartum year is a critical time during which many women experience anxiety symptoms that are more than the average parental stress (Giakoumaki et al., 2009). Anxiety symptoms can be significant, requiring intervention in an estimated 20% of postpartum women (Leach et al., 2017), while the prevalence of anxiety in the general population is 14.6% (Baxter et al., 2014). A meta-analysis of studies published between 1950 and 2016 found that 13.9% of women experienced comorbid depression and anxiety between 1 and 24 weeks postpartum (Falah-Hassani et al., 2017).

The Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), the gold standard in psychiatric diagnosis of the American Psychiatric Association (2013), does not provide a distinct diagnosis for postpartum anxiety, so midwives must rely on generalized anxiety symptoms to screen for anxiety to design the woman’s plan of care (Pawluski et al., 2017). The DSM-5 defines generalized anxiety disorder as excessive anxiety, worry and apprehension about a variety of circumstances (American Psychiatric Association, 2013), while the American Congress of Obstetrics and Gynaecology (2015) recommends asking mothers if they are having difficulty sleeping or intrusive thoughts to determine if a woman is experiencing anxiety that may be concerning.

Anxiety symptoms are often dismissed by providers because it is common to be somewhat anxious during the postpartum period and a mother’s heightened level of anxiety serves to protect her infant from harm (Pawluski et al., 2017). Women living with PPA, however, frequently suffer from physiologic symptoms of insomnia, restlessness, racing heart, irritability and panic attacks (Wardrop and Popadiuk, 2013). These symptoms are usually accompanied by intrusive thoughts, worry, sense of dread, fears and phobias (Wardrop and Popadiuk, 2013). There are several risk factors for PPA, including a history of anxiety or depression, low social support, or a history of interpersonal violence (Bayrampour et al., 2016). Table 1 details both risk factors for PPA as well as potential symptoms postpartum women may report.

Table 1.

Risk factors and potential symptoms of PPA.

Risk factors

History of anxiety or depression Low social support Low partner support History of interpersonal violence High perceived stress High perceived childcare need Low self-esteem

Symptoms

Excessive anxiety, worry and apprehension Confusion, indecisiveness Sense of dread Difficulty sleeping Restlessness Irritability Fears, phobias Panic attacks Feelings of heart racing Intrusive thoughts Expressing over- or under-sensitivity to infant cues Feelings which affect activities of daily living and personal relationships Feelings occur with or without apparent triggers Reporting use of alcohol or other substances to self-medicate

Sources: ACOG (2015), APA (2013), Bayrampour et al. (2016), Breitkopf et al. (2006), Dennis et al. (2017), Goodman and Tyer-Viola (2010), Pawluski et al. (2017), Ross and McLean (2006), Seymour et al. (2014), Wardrop and Popadiuk (2013).

The consequences of PPA can be significant as it is associated with poor infant growth, reduced breastfeeding rates (especially for first time mothers), delayed cognitive and social development in the infant, delayed attachment and bonding, reduced touching and communication with the infant, and increased risk of anxiety and other psychiatric disorders in the infant (Lonstein, 2007; Paul et al., 2013). Infants can develop difficult sleeping and eating patterns that can manifest into difficult temperament (Seymour et al., 2014). Maternal anxiety in the first year following birth may be associated with child emotional negativity and behaviour problems at two years of age, qualities that can be predictors of poor emotional development throughout childhood (Prenoveau et al., 2017).

In addition to poor infant outcomes, mothers with anxiety may have difficulty trusting natural maternal cues, hindering the ability to bond with their babies (Seymour et al., 2014). These mothers may overstimulate their infant and may display inconsistent behaviour that confuses the infant (Seymour et al., 2014). The mother’s belief about her ability to parent and her success as a mother can also be negatively affected by PPA (Britton, 2008). Anxiety can drastically interfere with parenting abilities by restricting activities, creating unrealistic fears or leading mothers to self-medicate with alcohol (Breitkopf et al., 2006) or smoking (Park et al., 2009). Mothers battling anxiety are more likely to increase their use of health care resources; burdening time, finances, relationships and the health care system (Yawn et al., 2015).

With relatively high prevalence and potentially significant impact, there is a clear need for routine screening of postpartum women (Ross and McLean, 2006) with prompt intervention for those who screen positive (Bittner et al., 2014). The benefits of intervention are different for each woman, but may include increased bonding with the newborn, increased positivity of relationships, increased ability to form daily functions, and improved quality of life (Breitkopf et al., 2006).

Professional organizations across the globe have expressed the need for increased surveillance of perinatal mood disorders such as PPA. The World Health Organization (2017) recommends integrating maternal mental health into all health services received by the mother, as well as pediatric visits. In 2015, the Association of Women’s Health, Obstetric and Neonatal Nurses released a position statement recommending that all postpartum women be screened for anxiety disorders (Association of Women’s Health, Obstetric and Neonatal Nurses, 2015). The American College of Gynaecologists (2015) recommends that women be screened for both depression and anxiety symptoms at least once during the perinatal period. Currently, the American College of Nurse Midwives (ACNM), the American College of Family Physicians, and the United States Preventive Services Task Force recommend that postpartum women be screened for depression, but the organizations have no position on screening for anxiety; the same is true for the International Confederation of Midwives and the International Federation of Gynaecology and Obstetrics (Toler, 2015).

Women face minimal risk in being screened (Hirst and Moutier, 2010; Breedlove and Fryzelka, 2011; US Preventative Services Task Force, 2016), however, it is important to note that screening by itself is insufficient to improve clinical outcomes and must be coupled with appropriate follow-up and treatment (Gjerdingen and Yawn, 2007). Midwives may be hesitant to screen because there is no standard recommendation for follow up endorsed by a professional body (Toler, 2016). The American Academy of Family Physicians recommends that mild to moderate PPD be treated with either psychotherapy or selective serotonin reuptake inhibitors (SSRIs), while moderate to severe PPD should be treated with both (Langan and Goodbred, 2016). Likewise, the recommended treatment for generalized anxiety disorder is a combination of SSRIs and psychotherapy (Locke et al., 2015). Of the available SSRIs, fluvoxamine, paroxetine, sertraline, and citalopram are first line choices in breastfeeding women because they pass the least amount of medication to the infant through breastmilk (Langan and Goodbred, 2016). Both cognitive behaviour therapy - structured therapy that aims to alter cognition and behaviour, and interpersonal therapy - therapy that addresses relationships with people, loss and change in relationships, are known to be effective at alleviating fear and anxiety in women with PPA (Centre for Perinatal Excellence, 2014). Because PPD and PPA respond similarly to SSRIs (Langan and Goodbred, 2016; Locke et al., 2015), it is reasonable to assume that evidence-based complementary alternative medicine (CAM) that is effective for PPD is also appropriate for PPA. Evidence suggests that medication augmentation with omega-3 fatty acids, exercise, or folate may improve standard treatment outcomes in PPD (Deligiannidis and Freeman, 2014). Bright light therapy, acupuncture and massage may also augment PPD therapy, but should not replace standard treatment (Deligiannidis and Freeman, 2014). See Table 2.

Table 2.

Interventions to consider in treatment of PPA.

Counselling/Psychotherapy	Cognitive behavior therapy (CBT) Structured therapy Interpersonal therapy (IT)
Pharmacotherapy	First-line choices:
	SSRIs Fluvoxamine * Paroxetine Sertraline* Citalopram
	Second-line choices:
	SSRIs Paroxetine* Fluvoxamine
	SNRIs Venlafaxine Duloxetine
	Other Bupropion Mirtazapine
Complementary alternative medicine	Omega-3 fatty acids Exercise Folate supplementation Bright light therapy Acupuncture Massage

Sources: Centre of Perinatal Excellence (2014), Deligiannidis and Freeman (2013), Langan and Goodbred (2016), Locke et al. (2015).

^*Considered safer alternative during lactation.

Key: SSRIs = Selective Serotonin Reuptake Inhibitors; SNRIs = serotonin–norepinephrine reuptake inhibitors

Perhaps because of poorly defined management of PPA, few women who screen positive for PPA are treated or referred – even when they ask for help (Goodman and Tyer-Viola, 2010). Women who receive treatment at the site where pregnancy and postpartum care has been established are more likely to receive follow-up care (Smith et al., 2009), thus making screening and treatment by maternity care providers essential for women with PPA. Women with PPA are likely to interact with the healthcare system frequently both during pregnancy and the postpartum period, providing ample opportunities for health care providers to engage with and screen women (Yawn et al., 2015).

This pilot quality improvement project implemented an evidence-based screening program for women under the care of midwives experiencing postpartum anxiety symptoms. In the United States (U.S.), where this implementation occurred, certified nurse-midwives (CNMs) and certified professional midwives (CPMs) attend 8.3% of the more than 4 million births each year (ACNM, 2015). Midwives in the U.S. may refer to either a CNM who has graduate-level education and may treat women throughout the reproductive lifespan, or a CPM who has training based on an apprenticeship model and whose scope of practice is to care for the woman during the pregnancy and postpartum period (Graduate Nursing EDU, 2017). While all midwives can implement use of a standardized screening for PPA, further evaluation and initiation of treatment may be limited by scope of practice for some types of midwives. Where referral is required for patients to be treated, all types of midwives and other maternity care providers should do so promptly during the early postpartum period. Initiation of treatment before symptoms potentially worsen, holds promise for reducing unnecessary suffering for women and their families (Giakoumaki et al., 2009). This project aimed to increase screening in the selected birth centres by 75%, with successive midwife counselling or referral provided to those women who screened positive.

Materials and methods

This quality improvement project modified existing postpartum care utilizing the following components: 1) development of an educational video for maternity care providers, 2) initiation of screening for PPA, 3) implementation of a counselling tool to guide conversations with postpartum women who screen positive for PPA, and 4) provision of treatment and referral resources for PPA including lists of postpartum mental health specialists specific to their area. These components were based on evidence that suggests maternity care providers are unsure of how to screen for PPA and feel they do not have adequate resources to intervene and refer women who screen positive (Goodman and Tyer-Viola, 2010). The Standards for Quality Improvement Reporting Excellence (SQUIRE) 2.0 guidelines were used as the framework for reporting the outcomes of this quality improvement (QI) project designed to improve healthcare quality (Ogrinc et al., 2015; SQUIRE, 2017).

The Stetler Model (see Fig. 1) was chosen as the framework for this project because it stresses the importance of critical thinking when utilizing research data in the practice setting (Stetler, 2001). Rather than focusing on change at an organizational level, the Stetler Model is oriented around the individual practitioner (Stetler, 2001; White and Dudley-Brown, 2011). This model was an appropriate fit for this project because it has shown validity in similar projects and emphasizes the use of critical thinking and decision making to encourage the application of research (White and Dudley-Brown, 2011).

Fig. 1.

Stetler model of evidence-based practice to promote quality care. (Reprinted from Nursing Outlook, Vol. 49, Stetler, C. B., Updating the Stetler model of research utilization to facilitate evidence-based practice, pg. 272–279, copyright © 2001, with permission from Elsevier).

The first phase of the Stetler Model, Preparation, was implemented when collection and evaluation of data was completed and it was confirmed that the prevalence of pathological PPA ranges from 6.1% to 28.9% in postpartum women (Giakoumaki et al., 2009) and that current midwifery practice does not include screening for PPA despite the 2015 AWHONN and 2015 ACOG position statements that recommend screening. Phase two of the model, Validation, was implemented when the body of research evidence regarding PPA was critiqued. Phase three, Comparative evaluation/decision making occurred when the sites and committee members were selected. The fourth phase, Translation/application, consisted of implementing the program, including the creation of all the resources that were provided to the participating sites (Melnyk-Mazurek and Fineout-Overholt, 2005). The fifth and final phase, Evaluation, occurred following eight weeks of program implementation when data was collected. As suggested by the model, implementing change required consideration of all stakeholders including the project organizers, the site research committee, the participating birth centres and the individual midwives at varying times. Table 3 details the process and phases in implementing this project.

Table 3.

Phases in project implementation for PPA.

Phase 1 - Foundational preparation	Responsible person(s)	Time frame
Literature Search to Establish Knowledge/Problem-Based Triggers	Project leader	Weeks 1–4
Review/Synthesize Relevant Research	Project leader & team	Weeks 1–4
Phase 2 - Validation of Problem
Identify community problem	Project Leader, Licensed Clinical Social Worker, AABC Research Chair	Weeks 1–4
Determine priority for organization AABC Board of Directors Approval Approval of ethical committees	AABC Research Chair & Executive Board	Weeks 4–5
Select Outcomes to be Achieved	Project leader, Team	Weeks 4–5
Phase 3 - Comparing and Decision Making
Analyse international programs with similar focus	Project leader	Weeks 5–8
Draft program plan and guidelines based on the evidence	Project Leader & Team	Weeks 5–8
Identify appropriate participants	AABC research chair	Weeks 8–10
Communicate with and provide training to participants	Project leader	Weeks 9–11
Phase 4 - Applying the Knowledge
Collect baseline data	Project Leader	Weeks 12–14
Implementation of required variable in online data registry in pilot settings	AABC Research Chair	Weeks 12–20
Phase 5 - Evaluating the Program
Data collection	AABC Research Chair	Week 21
Program evaluation	Project Leader & Team	Weeks 21–28
Program sustainability evaluation	Project Leader, Licensed Clinical Social Worker, AABC Research Chair	In process

Sample

The American Association of Birth Centres (AABC) was selected as the site for this project. The flagship organization for birth centres across the U.S., the AABC promotes national and state regulations and standards and quality improvement programming (American Association of Birth Centers, 2015). There are approximately 334 AABC-accredited birth centres in the U.S. (American Association of Birth Centres, 2017). These birth centres are largely staffed by certified nurse-midwives (CNMs) and certified professional midwives (CPMs) and primarily serve patients who are low-risk but of varied race/ethnicity (Stapleton et al., 2013). In the U.S., 0.3% of total births occur in the birth centre setting each year (Stapleton et al., 2013). The inclusion criteria for the selection of birth centres was AABC membership, participation in the AABC online data registry, AABC Perinatal Data Registry^™, accreditation by the Commission for the Accreditation of Birth Centres (CABC), active and thorough maintenance of records of postpartum care, and agreement to institute a PPA screening and intervention program at the site. Recruitment of birth centres was conducted by passing out flyers at the national AABC conference and through emails sent to 26 birth centres across the U.S. A subset of 11 birth centres from geographically diverse areas of the U.S. were invited to participate in the project. These birth centres were selected based on good histories of data collection including the contribution of data to the AABC Perinatal Data Registry for the past 5–10 years, and had a low rate of inaccurate or incomplete data submission.

The project was granted exempt status by the Baylor University Institutional Review Board (IRBNet #925,155-1) and approved by the AABC Research Committee and Board of Directors prior to implementation.

Screening tool

A subscale of the Edinburgh Postpartum Depression Screen, referred to commonly as the EPDS, was selected for this project. The EPDS has a total of ten items, with three items specific to anxiety. When these three items are used separately from the EPDS to screen for anxiety, the tool is referred to as the EPDS-3A. The EPDS-3A was used for this project for three reasons. First, the tool is public domain, easy to administer, and available online and in a wide number of languages (Massachusetts General Hospital Centre for Women’s Mental Health, MGHCWM, 2015). Second, it is widely used in the U.S. when screening for postpartum depression (Fairbrother et al., 2015) – including at AABC birth centres, thus making it familiar to many maternity care providers and reducing patient burden in completing more than one screening tool. And third, the EPDS has established reliability and validity, including the anxiety subscale, the EPDS-3A (MGHCWM, 2015). A cut-off score of six has been established as more reliable in isolating anxiety (Rowe et al., 2008; Matthey et al., 2003; Musik, 2000; Tendais et al., 2014) and was the cut-off score used in this project.

Pre-implementation

Following recruitment of sites, midwifery staff were asked to watch a video developed specifically for this project. The video was first piloted with a diverse group of midwives (n = 11) who had clinical experience with postpartum depression and anxiety in a variety of practice settings, including birth centres; experience ranged from less than one to thirty years. The video was also reviewed by a licensed clinical social worker with expertise in peripartum mental health. Following review and feedback from the pilot group, revisions were made to the final video (see Table 4). This 2-min video was then emailed to each participating birth centre site, as well as uploaded onto YouTube for viewing. The video reviewed risk factors and symptoms of PPA, as well as how to use the EPDS-3A screening tool.

Table 4.

Characteristics of pilot participants (n = 11).

Participant	Age (years)	Race	Clinical experience (years)	Year CNM Cert	Midwifery education	Likert Response* by Statement**
						S1	S2	S3	S4	S5
1	-	AA/B	30	1992	Certificate	5	5	5	5	5
2	60	C	20	1996	Certificate	5	5	5	5	5
3	32	C	5	2012	DNP	5	5	3	5	5
4	33	C	0	2017	DNP	5	5	5	5	5
5	33	C	5	2012	DNP	5	5	5	5	5
6	31	C	4	2013	MS	5	5	5	5	5
7	35	C	< 1	2016	MS	5	5	5	5	5
8	60	C	21	1995	MS	5	5	5	5	5
9	34	C	3	2013	MS	5	5	5	5	5
10	60	A	30	1980	MS	5	5	5	4	5
11	36	C	1	2014	MS	4	4	4	4	5
					Mean Rating	4.9	4.9	4.7	4.8	5

Key: AA/B = African American/Black; C = Caucasian; A = Asian; CNM = Certified Nurse-Midwife; Cert = certification; DNP = Doctor of Nursing Practice; MS = masters of science

^*Likert scale: 1 = strongly disagree, 2 = somewhat disagree, 3 = neither agree or disagree, 4 = somewhat agree, 5 = strongly agree

^**Statements:

S1: The video was easy to understand.

S2: I understand the definition of postpartum anxiety.

S3: I understand that postpartum anxiety can be a condition that is distinct from postpartum depression.

S4: I understand how to use the Edinburgh Postnatal Depression Scale to screen for postpartum anxiety.

S5: I feel compelled to screen my patients for postpartum anxiety.

Counselling and referral

The project was implemented for a period of 10 weeks during the 4th quarter of 2016 and the 1st quarter of 2017. During this period, midwifery staff were asked to screen all postpartum patients using the EPDS-3A and to record the score in the AABC Perinatal Data Registry^™ (PDR). This online data registry was designed to collect data on the outcomes of midwifery care at AABC-member birth centres across the U.S. and provides valid data for use in research (Stapleton, 2011). A specific mandatory field was added to the PDR database for this pilot project to collect data on whether the screening was performed and if the score was screen-positive.

Where patients screened positive for PPA (EPDS-3A score of ≥ 6), midwives were asked to provide counselling support. To support counselling efforts, all midwives were provided access to a guidance tool, the N* U* R* S* E* Plan for Postpartum Anxiety. This tool was adapted from N* U* R* S* E* Approach for Postpartum Depression, developed by the American College of Nurse-Midwives (2002) (see Table 5). The tool was originally featured in the book Women’s Moods: What Every Woman Must Know About Hormones, the Brain, and Emotional Health, authored jointly by a physician board certified in psychiatry and neurology and a certified clinical nurse specialist who together treat women with postpartum mood disorders (Sichel and Driscoll, 2000).

Table 5.

N*U*R*S*E* plan for postpartum anxiety (adapted from the ACNM, 2002).

Nutrition	Women who are anxious after birth are often overwhelmed by the demands of motherhood and lack of sleep and have no energy to prepare meals. The body needs good food to heal so every effort should be made to eat well. Family and friends can help with food preparation. A multivitamin every day will provide some of the basic requirements for vitamins. Fluids are important for both health and breastfeeding. Drinking 8–10 glasses of water every day will help both mother and baby. Stay away from alcohol because it can add depressive symptoms to PPA. Omega-3 fatty acid is associated with improved outcomes in women with anxiety. Dietary intake is generally low in postpartum women. A daily fish oil supplement of 1–3 gs is encouraged.
Understanding	A woman with PPA feels a constant sense of dread and worry, and might feel unable to care her for baby (Postpartum International, 2016). This is NOT her fault. Understanding and acceptance by family and friends is essential for her to begin to believe in herself again. It is important to get professional help to cope with the anxiety and to begin to recover. Support groups are an excellent idea. The best understanding comes from those who have experienced postpartum mental health disturbances.
Rest and relaxation	Sleep is critical for health and healing. Most women with postpartum anxiety have difficulty sleeping. Try different strategies, such as a warm bath before bedtime, massage, relaxation techniques, or meditation. When women are breastfeeding, they may need assistance with one night-feeding to get some uninterrupted sleep. Call for help if she goes without sleep for more than two days.
Spirituality	It is helpful to draw on what has made her feel uplifted and joyful in the past. Many things, from formal religion to listening to music that helps her find a sense of well-being, will in turn, give her strength to cope and begin to recover. Mindfulness therapy has proven to be helpful at alleviating anxiety symptoms and may be practiced at home, utilizing resources found on the internet like podcasts and videos.
Exercise	Physical exercise improves brain function and a sense of well-being. Set up a program that is realistic, taking small steps to increase her activity. Family and friends can help with short walks, or with offers of childcare while she exercises. Some gyms and yoga studios may offer classes where she can bring her baby. Yoga has been shown to be effective at reducing anxiety symptoms (Li & Goldsmith, 2012).
Psychological and pharmaceutical strategies	Both interpersonal psychotherapy and cognitive behavioral therapy can both have positive effects on anxious behaviors and coping mechanisms (Beck, 2014).
	The SSRI class of antidepressants is the most commonly prescribed medication for women with postpartum anxiety and is safe for breastfeeding. All patients who screen positive for postpartum anxiety should be offered SSRI therapy such as Prozac, Zoloft or Lexapro.

Birth centres were also provided with referral lists of postpartum mental health specialists and support groups specific to their geographic location for referral where needed or indicated. These resource lists were compiled from internet research and information provided by Postpartum Progress International, including support groups, counsellors, psychiatrists, doulas and yoga classes. In addition, a tool created by the Massachusetts Child Psychiatry Access Project, Recommended Steps before Beginning Antidepressant Medication During Pregnancy and Lactation (Byrns et al., 2014), was provided to midwives and included information on how to counsel breastfeeding women on medication use as well as an antidepressant treatment algorithm.

Data analysis

Initial data cleaning and preparation as well as descriptive statistics were produced using R version 3.3.2 (R Core Team, 2016). Prior to analysis, variable names were recoded and labels were applied to the data. Additionally, data for sites with fewer than five participants (n = 2) were removed to ensure a more representative sample for analysis. Removing these sites reduced the total sample size from 391 to 387 participants across nine sites. Counts and proportions were run to examine differences in screening, EPDS-3A and EPDS scores, and follow-up rates by site. Descriptive statistics included a 9-sample test for equality of proportions without continuity correction to determine if the number of midwives differed by site.

Results

Sample characteristics

A total of 11 birth centres were originally recruited for participation. Birth centres were located across the U.S. and represented all geographic regions with exception of the Midwest section of the country. These centres employed an average of 5 midwives (range: 2–8) with a total of 54 midwives across all 11 original sites. Of the 54 midwives, 96% were CNMs and 4% CPMs. Midwives were all currently certified by either the American Midwifery Certification Board (CNMs) or the North American Registry of Midwives (CPMs). Midwives were predominately Caucasian (n = 50, 92%) with the remainder African-American (n = 8, 8%). The average clinical midwifery experience was 10 years (range: 3–15 years). Following final data collection, two sites were removed due to lack of participation (less than five records entered).

Education, treatment and referral intervention

The educational video was viewed by birth centre midwives on YouTube a total of 36 times prior to study initiation. Midwives were asked to view the video prior to the initiation of screening but it was unknown whether midwives watched the video in groups or alone, or if midwives from every centre viewed the video. Similarly, it was unknown the extent to which midwives reviewed the counselling materials or made use of the treatment or referral information sent in advance to all participating sites.

PPA screening

Although referral and treatment are integral to healing for women with PPA, for the purposes of this project, our statistics focused solely on the screening intervention, as it is the critical first step. The project began with 11 sites, but following data collection, data from two sites were eliminated based on the criteria that the site had less than five total participants. A total of 387 patients were eligible to be screened across the final 9 participating sites during the project period with a total of 318 patients screened (89.6%). Of the patients who were screened, birth centre sites 5, 6, and 7 had the greatest number of patient screenings for a combined 232 screenings, accounting for 59.9% of all screenings. Site 1 had the fewest participants (n = 16, 4.1%). A 9-sample test for equality of proportions without continuity correction determined that the proportion of screened participants significantly differed (χ² (8, n = 387) = 146.4, p < 0.0001). Across all sites, 37 participants (9.7%) were lost to follow-up as they did not return for a postpartum visit when screening for PPA would have been conducted. Those lost to follow-up were from five of the nine birth centres. See Table 6.

Table 6.

PPA screenings by birth center site and patients lost to follow-up by site (N = 9).

Birth center site	Screenings (N = 318)		Lost to follow-up (n = 37)
	N	%	N	%
1	15	4.1	0	-
2	35	9.0	0	-
3	32	8.2	6	16.2
4	23	5.9	0	0
5	77	19.9	5	13.5
6	72	18.6	13	35.1
7	83	21.4	10	27.0
8	24	6.2	3	8.1
9	25	6.4	0	-

Note: Percentages may not equal 100 due to rounding.

A portion of women screened positive for PPA; 12.5% (n = 40) had a positive EPDS-3A score greater than 6. A total of 15 (7.0%) screened negative for PPD with a score of less than 12 on the EPDS though had an EPDS-3A score of 6 or greater, indicating a positive anxiety screen. Sites differed by the number of screened participants with an EPDS-3A score greater than 6 with site 6 (n = 13, 32.5%) having the greatest number of participants, and sites 7 (n = 1, 2.5%) and 3 (n = 1, 2.5%) having the fewest (see Fig. 2).

Fig. 2.

Patients who screened EPDS-3A positive by site (N = 318).

Of all screened participants with an EPDS score (n = 215), 14 (6.5%) had a EPDS score of 12 or higher, indicating a positive depression screen. One site (#5) had no participants with valid data for this measure because they mistakenly screened using only the EPDS-3A, but of the remaining 8 sites, site 4 (n = 5, 35.7%) had the greatest number of participants with an EPDS score of 12 or higher. In contrast, there were two sites where no participants had a score of 12 or higher (site 2 and 9).

For all screened participants with an EPDS score (n = 215), 12 (5.6%) had a EPDS score of 12 or higher and an EPDS-3A score greater than 6 thus being screen positive for both depression and anxiety. Of the 8 sites with a valid measure (site 5 had no participants with valid data), site 4 (n = 4, 33.3%) had the greatest number of participants who were screen-positive for both depression and anxiety. In contrast, there were two sites where no participants were screen positive for both depression and anxiety (sites 2 and 9).

Midwife feedback

Comments were sought from midwives across sites to better understand issues related to implementation of this change project. Each of the nine birth centres that participated in the project was contacted a minimum of two times for feedback, and six sites responded and answered questions about their experiences with the project via email or phone (54%). Although this quality improvement project is not a qualitative research project, the comments from the midwives were examined to determine if improvements could be made if the project is implemented again in the future.

A central theme that emerged was provider surprise at how few women screened positive using the EPDS-3A. Although each provider said that using the EPDS-3A was easy to integrate into practice, this may be an indication that a more sensitive tool is needed.

In my clinical judgement, we see a lot of women with some degree of anxiety, but so far no one has screened positive using this tool.

[The rate of positive screens] was lower than I expected due to the highly anxious nature of our clientele during pregnancy.

[The rate of positive screens] was lower [than I expected]. And the people we thought would have a difficult time before birth, did have a more difficult transition afterwards.

Participating midwives also commonly felt that the time spent with patients throughout their pregnancies and the subsequent relationships that were built meant that clinical judgment was better than a screening instrument. Once participant commented:

I feel like typically we are quick to identify symptoms because we do a home visit at their three day visit and seeing someone in their home makes a big difference because we are looking for symptoms. We have a really good relationship with our clients and we are the ones they come to if they are feeling anxious or not coping.

All midwives felt that the psychiatric professional resource list was helpful. Some of the midwives used the SSRI algorithm to prescribe medication to patients who screened positive, but some did not prescribe medication at all. These midwives cited patient resistance to medication as the reason that they do not often prescribe medication.

We were not more likely to prescribe medication if someone screened positive because we do a lot of herbals. We have really good herbals we use especially for anxiety and we have a protocol if someone develops symptoms in pregnancy or postpartum.

While most midwives felt that clinical judgment was more valuable than a screening tool, all expressed a need for more focus on anxiety in the clinical setting. Midwives expressed that participating in this project encouraged them to pay more attention to anxiety symptoms during postpartum visits. Several noted that they would continue to use the EPDS-3A subscale as a part of the postpartum visit.

Discussion

This project detailed implementation of a screening program for PPA by midwives working in a birth centre setting. Findings demonstrated that routine screening for anxiety identified 15 (6.5%) women experiencing anxiety symptoms that would have been overlooked if they had been received routine screening for depression alone. Use of a standardized screening tool, the EPDS-3A, demonstrated both usefulness and acceptability, as did the materials for counselling, treatment and referral. Findings from this project suggest important aspects to consider when planning to adopt a screening program for PPA, as well as potential limitations.

Clinical application

The Stetler Model was used to guide this project and provided an important framework for identification of key steps and factors needed for implementation and evaluation of the program. Focus of the model on judgmental activities about the appropriateness, desirability, feasibility, and manner of using the findings in individual or group practice (Stetler, 2001) allowed for an effective project structure. Use of the model would likely be helpful in the successful initiation of screenings in other settings.

A key factor in success of this quality improvement project was support and assistance of the AABC Research Chair (SS) who served as both a key leader and champion for change in practice. In this capacity, she shepherded the project through the AABC approval process, identified and promoted birth centre participation, ensured change in the data collection system, and secured data important in the evaluation of project outcomes. The value of a leader and project champion is consistent with research reporting that clinical leaders and champions are crucial for successful practice change (Soo et al., 2009; Stanley et al., 2008; Stetler et al., 2014).

Feedback from midwives across sites indicated that the project increased awareness of the need for routine screening for PPA, as well as the need for consideration of counselling and treatment, as appropriate. While an overall PPA screening rate of 86.3% demonstrated awareness, the extent of counselling, treatment, and/or referral by midwives was unknown. It would be important to consider further modification of the PDR to include the nature of treatment where patients screen positive. Such data would provide greater knowledge of the impact of findings in the improvement of perinatal outcomes where it is known that as few as a quarter of positively screened postpartum women receive treatment and even fewer are ever referred (Goodman and Tyer-Viola, 2010).

The number of patients lost to follow-up, and thus not screened for PPA (9.7%) is an area of concern. Two of the five centres with the greatest loss to follow-up were sites 6 and 7, with 35% and 27% respectively. These sites were known to be larger volume midwifery practices larger numbers of patients with government health insurance. Restrictions in healthcare coverage may have been an important factor in failure to return for postpartum care with PPA screening. This project did not ascertain the nature of the populations served by birth centre midwives and is likely an important variable in larger scale project feasibility. Alternate strategies for screening may need to be considered, such as telephone screening or through use of social media (Baer et al., 1995; Moorhead et al., 2013).

Feedback from participants underscored the appropriateness, desirability, and feasibility of the screening video, as well as treatment and referral intervention in the birth centre setting. However, the manner of using the findings by individuals in a group practice suggested the need to further tailor the intervention with greater detail provided in use of CAM therapies. Overall CAM rates of use among postpartum women in the U.S. is 28%, with higher rates of use among those with depression/anxiety (Birdee et al., 2014), and midwives often suggest the use of CAM during the perinatal period (Hastings-Tolsma and Terada, 2009). The reluctance of patients to use anti-depressants as reported by midwives may have been influenced in part by midwifery education and scope of practice. Birth centre participants included both CNMs and CPMs who have similar but different educational competencies and scope of practice. CNMs have educational competencies related to primary care and the treatment of mood disorders, as well as independent prescriptive authority (American College of Nurse Midwives, 2014). CPMs have a more limited scope of practice which excludes primary care and do not have full prescriptive authority in U.S. states (American Medical Association, 2016). Education, including clinical training in depression/anxiety, and a more restricted scope of practice, are likely influential in a more robust approach to offering counselling, treatment or referral to women who screened positive.

Limitations

Implementation of this quality improvement project had several limitations important to note in considering clinical application. First, the approach to PPA screening and intervention described in this project has not been described for use in other settings and validation is needed for wider use. Second, the PDR did not specifically query the interventions and/or referral or treatment initiated in the presence of PPA; such information is needed to provide a more robust evaluation. When reviewing the PDR, no attempt was made to determine the accuracy of the data that was entered through chart review or otherwise. Additionally, we do not know if the level of clinical experience of the midwife had any influence on whether the patient was treated or referred. Experience may be an important factor in the initiation of therapies such as use of anti-depressants. Third, it is unknown who viewed the video or from which sites, or if it was viewed in its entirely. Knowledge of viewing patterns and how those relate to screening, treatment, and referral practices would be important to ascertain in future research. Fourth, qualitative analysis of comments was conducted by the project leader without determination of trustworthiness (Golafshani, 2003) and more rigorous analysis may have reached different conclusions. Fifth, full implementation and determination of long-term sustainability has yet to be determined. Full, formal adoption of the program is important in the robust evaluation of practice change. Sixth, and finally, the birth centres who participated were a convenience group suggested by the AABC organization and consisted of those who volunteered to participate. More rigorous selection may have led to different outcomes.

Program sustainability

Sustainability of this program is a critical component of increasing screening, awareness, treatment and referral of women with PPA. The midwives and birth centres may choose individually if they will continue to screen for PPA in their practice, thereby sustaining the aim of the project. Approaches to promote uptake of the change are needed for full system adoption and are currently under consideration by the AABC Executive Board. Providing educational webinars for members and revising the AABC Standards for Birth Centres to include PPA screening are potential strategies (AABC, 2016). Audit of the PDR at the 6-month mark is planned for birth sites participating in this project. Maintaining the prompts regarding PPA in the PDR, and policy specific to screening, could be effective in sustaining change as well. These specific interventions have found to be among those strategies which facilitate sustained change in practice (Mitchell et al., 2010).

Implications for practice and/or policy

This quality improvement project demonstrates that a significant portion of women in the birth centre environment experience PPA and midwives are in a prime position to provide initial intervention. PPA has historically been under-identified and under-treated (Goodman and Tyer-Viola 2010). Research on maternal mental health and infant development support the idea that early intervention and treatment of postpartum mood disorders are essential to prevent long-term sequelae (Dennis et al., 2017). It is important that clinicians and childbearing families are educated on the prevalence of PPA, its risk factors, symptoms and implications. Early screening and treatment may lead to improved outcomes for women and their families, and should be recommended by all organizations that seek to improve maternal and infant health. To fulfil its mission of being “with women,” ACNM should consider releasing an official position on PPA and encourage education and advocacy on the issue.

Conclusions

This quality improvement project explored results of initiating a PPA screening protocol within the birth centre environment. The routine screening used in this project identified 15 (6.5%) women experiencing anxiety symptoms that would have been overlooked if they had been screened for depression alone. Findings from this project highlight both the prevalence of PPA and the ability of midwives to screen and provide follow-up. Providing information to midwives likely improved their knowledge on the disorder and empowered them to treat or refer if necessary. Additional tracking of treatment, referral and follow up would provide more information on the potential of this screening project. Screening, counselling and referral for PPA is seriously lacking by maternity care providers (Goodman and Tyer-Viola, 2010) and midwives can play a vital role in supporting women throughout the postpartum year, an action that may lead to improved outcomes for women and their infants.