Data collection procedure

In this paper, symptoms of depression occurring in the antenatal period will be referred to as antenatal depression (AND) and those occurring in the postpartum period will be termed ‘PPD.’ Additionally, since the term perinatal refers to the antenatal and postpartum periods up to and including the infant’s 12-month well-baby visit, the term perinatal will be used interchangeably with the terms antenatal and postpartum.

This research project received Institutional Review Board approval (Protocol #19/09-098) with exempt status. The study employed a quantitative approach, using a convenience sampling method. Participants were invited to partake in an online survey, distributed between October 1^st, 2019, and February 18^th, 2020. Information about the study was disseminated to obstetricians and other primary care providers via the researchers’ professional contacts (phone calls/emails), and Facebook. Online electronic consent was obtained from interested parties indicating that participation was voluntary and could be terminated at any time. Participants indicated their written consent by proceeding with the online survey at the beginning of the survey through an individual link that could only be submitted once. To maximize response rates, a follow-up invitation was sent to those who had not yet completed the survey two weeks after the initial invitation. Participants received a $5 electronic Starbucks gift card sent to their email address upon survey completion. Completion time for the survey was 15–20 minutes. Upon submission, the responses were automatically encrypted and transmitted to a secure database, ensuring the privacy and protection of participant data.

Participant demographics

All participants were required to be over the age of 21, reside in the United States, and be providers of care to women of childbearing age at risk of experiencing AND or PPD. These criteria were established to ensure a level of professional experience and contextual relevance to the healthcare system in the United States. Practice locations within the US were not specified in the survey.

Survey instrument and data management

Two online surveys were created by the researchers to generate data for the study using Qualtrics Core XM software (2019–2020) (Qualtrics, Provo, Utah), an online platform endorsed by the institution for the distribution of survey-based research. The Qualtrics platform uses individual links using email distribution, thus allowing only one opportunity for submission, and preventing multiple submissions of the surveys. Although N = 114 respondents agreed to the electronic consent to participate in the survey, 13 did not respond to any further questions; consequently, the effective sample size is N = 101 (16 obstetricians and 85 other providers). In survey 1, The Role of the Obstetrician in Identifying and Supporting Women’s Mental Health Symptoms, obstetricians (n = 16) were recruited for participation, and in survey 2, The Role of Other Primary Care Providers in Identifying and Supporting Women’s Mental Health Symptoms, pediatricians, nurse practitioners, and physician assistants whose scope of practice included the perinatal period were recruited (n = 85).

Content for the online electronic survey included six demographic questions that noted the participant’s gender, age, ethnicity, highest degree completed, years in the field, and employment status. Thirty questions pertaining to aspects of screening for mental health disorders by physicians and other healthcare providers in antenatal and postpartum women included level of comfort during the screening process, ability to recognize signs and symptoms of mental health disorders, screening tools used to guide the assessment, if scores were reviewed with the patient, if treatment options were offered, if follow-up was done, and if a referral to a psychiatrist and/or therapist was initiated, among others. Several of these questions required the use of one of two Likert response styles: always, most of the time, about half the time, sometimes, and never; or strongly agree, agree, somewhat agree, neither agree nor disagree, somewhat disagree, and disagree. Finally, participants were asked to identify among 19 symptoms, those which they believed were associated with AND and/or PPD in their patients. These symptoms included excessive crying, feeling numb, feeling little or no attachment to the infant, having no interest in things that use to give pleasure, feeling little enjoyment, no energy, feelings of failure, feeling overwhelmed by the smallest of tasks, hopelessness, agitation with self/others/infant, insomnia, anxiety, excessive sleeping, difficulty concentrating, changes in appetite, fear of being alone with the infant, intrusive thoughts of harming the infant or harmful things happening to the infant, difficulty shaking unwanted feelings, and fear the symptoms would last.

Data analysis

Demographic characteristics and survey responses were summarized using descriptive statistics (counts and percentages). Demographic characteristics were compared across participant groups (obstetricians, other providers) using an exact version of Pearson’s Chi-squared test [14].

Survey responses related to AND, PPD, and mental health screening practices were compared across participant groups using the Cochran-Mantel-Haenszel (CMH) test, adjusting for potential imbalance in a key factor: years of practice in the profession. The same CMH approach was used to analyze symptoms considered by the participant to be related to AND, PPD, and anxiety. As this study was exploratory in nature, no adjustment for multiple comparisons was planned or implemented. No imputations of missing data were used; only observed data were analyzed. Statistical tests were 2-sided and p-values < 0.05 were considered statistically significant. All analyses were conducted using R version 4.3.1 [15].

Sample characteristics

Table 1 lists the demographic characteristics of the whole group (N = 101) and compares obstetricians (n = 16) to other primary care providers (n = 85). Most of the participants in each group were female (93.8% and 88.0%, respectively). Significant differences were found regarding age (p = 0.002) and race/ethnicity of the providers (p < 0.001). Obstetricians in the sample were more likely to be older (45–54 years of age) (56.2%) and Black or African American (43.8%) compared to other primary care providers who were more likely to be younger (25–34 years of age) (61.9%) and White (78.6%). Significant differences regarding level of education (p < 0.001) and how long providers had been practicing (p = 0.001) were noted with obstetricians having a higher level of education (75%) and being in practice for over 20 years (26.7%). Finally, significant differences were evident in the ages (p < 0.001), racial/ethnic groups (p < 0.001), and religious affiliations (p = 0.027) of the providers’ patient populations. Obstetricians were more likely to see patients that were between the ages of 15-to-35 years of age (66.7%), more racially/ethnically diverse (73.3%), and religiously diverse (64.3%).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Demographic characteristics of participants.

Total sample = 101¹.

https://doi.org/10.1371/journal.pone.0306265.t001

Provider perinatal screening practices

Table 2 represents data for participants’ responses to questions related to AND, PPD, and anxiety/mental health symptom screening practices; statistical comparisons across participant groups were adjusted for years of practice in the profession. Statistically significant differences between the two groups of providers were noted with 100% obstetricians addressing mothers that looked “sad, tired, upset, unhappy, or with a flat affect” (p = 0.044) and 93.3% asking mothers about “social support and forms of help” (p = 0.011) most of the time/always compared to 66.2% and 51.9% of other primary care providers, respectively. Regarding offering mothers’ treatment options such as referral to a psychiatrist or therapist (p = 0.005), having PND brochures in the waiting room (p = 0.008), and using the Edinburgh Postnatal Depression Scale (EPDS) [16] (p = 0.027), significant differences were also noted between the groups. For these variables, other primary care providers referred patients to a psychiatrist (35.8%) or therapist (59.3%), provided patients with information on AND and PPD and anxiety in their waiting rooms (77.5%), and used the EPDS (72.8%) more than the obstetrician group.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Perinatal screening practices: Participant responses to questions related to AND, PPD, and anxiety/mental health symptoms.

Total sample = 101¹.

https://doi.org/10.1371/journal.pone.0306265.t002

The following describes results for survey questions that were not statistically significant after adjusting for years of practice, but nonetheless show a tendency for provider differences in perinatal screening practices in their patient populations. For example, for the question, “Do you screen for depression/anxiety/mood disorders of the mother,” 93.3% of obstetricians answered most of the time/always compared to 61% of other primary care providers, and 73% of obstetricians and 70.4% of other providers responded most of the time/always to using a screening tool to identify AND and/or PPD. Seventy-three percent of obstetricians and 60.5% of other providers followed up with mothers about mental/emotional health most of the time/always. Most obstetricians (86.7%) compared to 59.3% of other providers asked about the mother’s connection with her infant during the office visit most of the time/always, and there was a greater tendency for obstetricians to see patients in their office (53.3%) compared to other providers who were more likely to see patients in a hospital (42%) or clinic setting (43.2%). About half (49.4%) of other providers compared to 13.3% of obstetricians responded somewhat agree/agree/strongly agree to the question, “Not sure how to address perinatal mental health,” although almost all providers in both groups offered treatment referral options.

Finally, of the 11 participants in each group that answered, “What tool do you use?” two providers used the Patient Health Questionnaire-9 (PHQ-9) [17] and nine used the EPDS in the obstetrician group, and ten used the PHQ-9 and one used the EPDS in the other provider group. Thus, among those who use a screening tool, there was a greater tendency for other primary care providers to use the PHQ-9.

Identification of mental health symptoms

Table 3 describes survey responses from both groups as participants were asked to identify symptoms [18] they believed were associated with AND, PPD or anxiety in their patients; statistical comparisons across participant groups were adjusted for years of practice in the profession. Significant differences were noted between groups with regard to positively identifying the following symptoms: excessive crying (p < .001) (100% obstetricians), feeling little or no attachment to the infant (p < .001) (100% obstetricians), little feeling of enjoyment (p = .021) (80% obstetricians), feelings of failure (p < .001) (93.3% obstetricians), hopelessness (p < .001) (66.7% obstetricians), agitation with self and infant (p < .001) (86.7% obstetricians), fear of being alone with the infant (p = .011), (73.3% obstetricians), and fear that these symptoms would last (p < .001) (80% obstetricians). A symptom that approached significance in detecting differences between groups was anxiety (p = .054), with most obstetricians (86.7%) identifying it as a symptom associated with AND and/or PPD compared to 43.2% of other primary care providers.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Identification of mental health symptoms.

Total sample = 101¹.

https://doi.org/10.1371/journal.pone.0306265.t003

Other modalities to assess for PND

Several studies [7, 9, 10, 13, 21] describe mothers’ preferences for engaging in face-to-face dialogues about PND with their provider, and believed the PND screening process would be more effective if providers reviewed the purpose of the screening tools, explained test scores, and provided timely follow-up referrals. Inquiring about a mother’s mood and her social environment is a valuable screening approach as it provides information on her emotional state and can alert the provider to the risk for PND [22], exceeding the benefits of a checklist alone [23]. Considering the variability in mental health screening practices among perinatal healthcare providers, in addition to using a standardized tool, a comprehensive screening interview that asks about mothers’ social support, connection to the infant, and provides a follow-up after a referral, may improve detection of mental health disorders in perinatal women [9, 24].

Best practice recommendations for provider PND screening

The American Academy of Pediatrics (AAP) and the American College of Obstetricians and Gynecologists (ACOG) both recommend frequent mental health screenings at certain intervals during the perinatal period [25, 26]. The AAP has recommended screening at the 1-, 2-, 4-, and 6-month well-baby visits [25]. Moreover, screening up to the 12-month period following delivery is prudent since symptoms of depression, anxiety, and mental health disorders are commonly detected during this time. A longitudinal study on rates of PPD screening in mothers during well-child visits up to 12 months of the infant’s life found a 15% rate of first-time positive screens in the 6–12-month period, and a 23% rate of mothers screening positive at the 12-month well-baby visit [27]. Yet, screening inconsistencies are common. [28] compared primary care providers’ (pediatricians/nurse practitioners) screening practices during well-child visits: 29% screened according to AAP screening guidelines, 64% screened at least once during the 12-months of the infant’s life according to the National Association of Pediatric Nurse Practitioners (NAPNAP) [29] screening guidelines, and 31% did not meet either AAP or NAPNAP screening guidelines. [30] noted improved PPD screening, from 83% to 88%, by pediatric and family nurse practitioners, when the EPDS was implemented at 1-, 2-, and 6-month well-child visits. Another study using a 20-item survey examined pediatrician attitudes and knowledge with PPD screening [8]. Sixty-six percent (66.3%) stated recognizing signs of PPD in a mother, and of these, only 13% referred the at-risk mothers to a mental health professional. Lack of time (84%) and not being familiar with mental health resources (53.5) were reported as screening barriers. Pediatricians that received training on PPD had improved confidence with assessment (p = .002) and knowledge of PPD (p < .001) and were more likely to screen for PPD (p = .012) in their patients.

Although ACOG [26] has recommended screening for symptoms of depression at the first obstetric visit, at 24–28 weeks of pregnancy, and at the comprehensive 4-week postpartum assessment using a validated tool, PND screening by obstetric healthcare providers in the U.S. remain inconsistent. Screening rates across a large hospital system found that 65.1% of women were screened for depression during pregnancy, and 64.4% were screened in the postpartum period [20]. [19] found screening practices for PND inconsistent and lacking a standardized screening tool in a U.S. community women’s health clinic serving lower income minority women. After implementing an educational in-service on PND screening and electronic health record (EHR) documentation, rates of PND documentation in the patients’ EHR rose from 56% to 92.7% (p < .05). The project’s success was attributed to the educational in-service received by the staff and the addition of a patient checklist in the EHR to properly document EPDS scores [19]. [12] examined the feasibility of universal mandatory PPD screening with the EPDS tool by adding a “hard stop” feature to the electronic health records of a large cohort (2,102) of high-risk women returning for the 6-week post-partum follow-up visit. Results indicated that PPD screening was completed in 99.5% of the women.

Training for providers

The uncertainty in addressing perinatal mental health issues and the variance in symptom recognition as seen in our study and in similar studies may be due to the providers’ lack of knowledge about perinatal mental health symptoms and comfort level with addressing symptoms [8]. Thus, all healthcare providers encountering perinatal women would benefit from in-service training on identifying symptoms associated with PND and addressing perinatal health issues to be better prepared to screen for this condition [8, 31]. Improving physicians and other healthcare providers’ knowledge of PND screening and the available resources for women with positive screens would undoubtedly remove some of the barriers that impede PND detection in the clinical setting and may prove beneficial to maternal mental health outcomes [8, 24, 32]. Our study’s findings, in conjunction with the findings of similar studies, underscore the need to improve provider training about PND as well as utilize a screening tool in maternal and child health primary care to improve rates in detection, referrals to mental health services, and follow-up care.

Limitations

This exploratory study had several limitations. First, the participant sample was small, with more ‘other primary care providers’ than obstetricians, resulting in findings that may limit generalizability to other maternal and child healthcare providers in the U.S. Second, our surveys did not include questions regarding providers’ awareness of the significant risk factors associated with AND and PPD or if the providers screened for these risk factors during their patient contacts. Third, although the surveys were disseminated through the professional networks of the researchers who are located primarily in the northeastern United States, information on actual geographical locations (i.e., city or county) specifying where the providers treated their patients was not collected in the demographic data, thereby limiting generalizability to specific areas in the U.S.