Patient outcomes following positive depression screening in a postpartum clinic: A single center retrospective cohort study Authors: Aleesha Jasmin Jethwa, Nan Guo, Natasha Harrison, Dipro Chakraborty, Kimberly Mendoza, Natsumi Kii, Brendan Carvalho, Danielle Mari Panelli, Pervez Sultan Published: 2026 ABSTRACT Background:Depression screening using the Edinburgh Postnatal Depression Scale is commonly performed at the 6-week postpartum visit, yet subsequent referral patterns and clinical outcomes are incompletely characterized.Methods:A single center cohort study was conducted using electronic health records of patients with a completed depression screen at the 6-week postpartum clinic between 2014–2023. Exposure was categorized as screen negative (Edinburgh Postnatal Depression Scale 0–9) or screen positive (Edinburgh Postnatal Depression Scale ≥10) for depressive symptoms. Primary outcome: the proportion of new mental health referrals made (to social work and/or psychiatry services) following the 6-week postpartum depression screen up to 1 year after childbirth. Secondary outcomes: the proportion of patients referred to each specialty after 6 weeks and during the peripartum period (antenatal to 1 year postpartum); patient variables associated with specialty referral from 6 weeks to 1 year postpartum using a multivariable logistic regression model; timing of postpartum referrals after 6-week depression screening; and psychiatric (medications, hospitalizations, mental health diagnoses) and infant associated outcomes within 1 year of childbirth.Results:Of 41,790 deliveries, 8,155 patients met inclusion criteria. 12.6% received a new mental health referral following 6-week postpartum depression screening up to 1 year after childbirth. After screening, social work and psychiatry referrals occurred in 4.7% and 3.7% of negative-screened patients and 37.1% and 19.5% of positive-screened patients, respectively. Across the peripartum period, social work and psychiatry referrals occurred in 8.6% and 17.5% of negative-screened patients, and 42.0% and 42.5% of positive-screened patients, respectively. Variables associated with referrals (6 weeks to 1 year postpartum) included divorced, separated and widowed marital status, ASA grade ≥3, delivery of a non-viable fetus, neonatal intensive care hospitalization, psychiatric diagnosis, and medication use. Referrals within 2 months postpartum were more frequent following positive screening compared to negative screening; Social Work (84% vs 49%); Psychiatry (53% vs 32%). Psychiatric medications, hospitalizations, mental health diagnoses and infant associated outcomes were more frequent following positive screening.Conclusion:Positive postpartum depression screening was associated with increased referrals to social work and psychiatry services; however, a critical gap persists between positive screening and specialty referral, diagnosis, and initiation of treatment. Findings support longitudinal mental health screening across the peripartum period. Introduction Postpartum depression (PPD) occurs in approximately 15% of individuals following childbirth in the US, and is recognized as a major public health issue associated with increased maternal and neonatal morbidity and mortality.1–5 PPD impairs maternal infant bonding, is associated with maternal suicide and also has negative effects on infant development, such as lower intelligence quotient, delayed language development, and behavioral problems.6 Given the individual, societal and economic burden associated with PPD, professional guidance from the American College of Obstetricians and Gynecologists (ACOG) endorses routine screening with a validated tool, such as the Edinburgh Postnatal Depression Scale (EPDS).7–12 Screening, when coupled with follow-up care, reduces both depressive symptoms and overall prevalence of PPD.13 However maternal and neonatal outcomes, particularly regarding specialty referrals, commencement of pharmacotherapy and subsequent neonatal diagnoses following positive PPD screening are under explored.13,14 Internationally, a 26,000 patient audit from Australia linked positive antenatal EPDS to adverse obstetric and neonatal events, but provided no information on mental-health referrals, medication use, or subsequent associated International Classification of Diseases, Tenth Revision (ICD-10) diagnoses.15 The largest US single-site study of 10,000 postpartum women revealed higher rates of bipolar-spectrum illness and suicidal ideation among screen positive patients but did not report downstream referrals, treatment or follow-up care.16 Addressing this knowledge gap would be invaluable for staff and patient education and could help improve screening program compliance and implementation.13,17 Understanding clinical outcomes following positive PPD screening could also help primary care service planning, resource and staffing allocation, and help generate actionable targets for quality-improvement initiatives and guideline refinement.13 The primary aim of this study was to determine the proportion of new mental health referrals made (to social work and/or psychiatry services) from the 6-week postpartum EPDS screen up to 1 year after childbirth. Secondary aims were to determine the proportion of patients referred to each specialty after 6 weeks and during the peripartum period (antenatal to 1 year postpartum); patient variables associated with specialty referral from 6 weeks to 1 year postpartum using a multivariable logistic regression model; timing of postpartum referrals after 6-week EPDS screening; and psychiatric (medications, hospitalizations, ICD-10 mental-health diagnosis codes) and infant associated outcomes within 1 year of childbirth. Methods Using STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guidelines, a retrospective cohort study was conducted following Institutional Review Board (IRB) approval.18 People delivering at a US academic healthcare system between 2014 and 2023 who completed EPDS screening during the postpartum clinic appointment (typically done 4–8 weeks postpartum, hereafter “6 weeks postpartum”), were included in the study. The annual rate of delivery at this center is 5,200 and EPDS screening at the postpartum clinic has been conducted since 2014. Inclusion criteria for participants included the following: age ≥ 18 years at the time of childbirth; delivery of a live or non-viable fetuses after gestational age ≥ 20 weeks, via any delivery mode at the hospital system; completion of all items of the postpartum EPDS screening tool (completed between 4 weeks 0 days and 8 weeks 0 days postpartum) and postpartum care delivered within the healthcare system. Exclusion criteria for participants were: no or incomplete documentation of the postpartum EPDS in the electronic health record (EHR); EPDS completed outside the specified time window (<4 weeks 0 days or >8 weeks 0 days postpartum); and EPDS score ≥10 while inpatient immediately following childbirth (which would typically prompt more urgent referrals occurring prior to our study period). Data was collected from the EHR via an institutional clinical data warehouse containing live Epic (Verona, WI) data from an academic health system. The database model captures patient-specific variables including demographics, diagnoses, procedures, medication and laboratory test results, as well as structured content extracted from clinical notes. The data warehouse is hosted under a Business Associate Agreement (BAA). This platform aims to facilitate advanced data science research on sensitive clinical datasets by providing a secure and collaborative big data analytics environment. Collected data included demographics (age, parity, race/ethnicity, insurance type, language), obstetric and childbirth variables (induction and/or augmentation of labor, duration of 1st and 2nd stage labor, mode of childbirth), psychiatric history (prior diagnoses, medications, hospitalizations, substance use), complications (including perineal trauma, postpartum hemorrhage, chorioamnionitis, admission to intensive care), use and type of labor analgesia, private, academic or community obstetric services, and neonatal outcomes (birth weight, hospitalization in a neonatal intensive care unit (NICU)). The ICD-10 codes used to populate psychiatric variables such as F53 (mental and behavioral disorders associated with the puerperium) are provided in Table 1. The primary outcome of this study was the proportion of new mental health referrals made (to social work and/or psychiatry services) from the 6-week postpartum EPDS screen up to 1 year after childbirth. Secondary outcomes were the proportions referred to each specialty after 6 weeks and during the peripartum period (antenatal to 1 year postpartum); patient variables associated with specialty referral from 6 weeks to 1 year postpartum using a multivariable logistic regression model; timing of postpartum referrals after 6-week EPDS screening; and psychiatric (medications, hospitalizations, ICD-10 mental-health diagnosis codes) and infant associated outcomes within 1 year of childbirth. Statistical methods Continuous variables were compared using ANOVA for normally distributed variables, and Mann Whitney U test for non-parametric comparisons. An uncorrected chi-square test was used to compare the frequencies of categorical variables between the two groups (EPDS score <10 vs ≥10) or Fisher’s exact test. Data are presented as mean (standard deviation, SD), median [interquartile range, IQR] for continuous variables, or number (percentage) for categorical variables. Univariable and multivariable log-binomial models were used to assess the risk factors for referral to social work or psychiatry specialties. Multivariable models were adjusted for potential confounders with a p-value < 0.2 considered significant in univariable analysis. Missing data was assessed for all analytic variables. Data was assumed to be missing at random (MAR) conditional on observed covariates. Missing data were handled using multiple imputation by chained equations (MICE) under MAR assumption. Five imputed datasets were generated in STATA 19 using predictive mean matching for continuous and multinomial logistic regression for categorical variables. Estimates were combined using Rubin’s rules. A sensitivity analysis was done using complete case analysis. All analyses were performed using two-tailed hypothesis testing with a P-value of less than 0.05 as statistically significant. All statistical analyses were performed using STATA (Version 19.5, College Station, TX). Results 41,790 patients delivered at a quaternary healthcare system between 2014 and 2023. A 6-week postpartum EPDS score was collected from 8,155 patients and included in the analysis. The STROBE flow diagram is shown in Figure 1. Table 3 summarizes the demographic, obstetric, anesthesia, psychiatry, and neonatal variables in patients that screened negative and positive for PPD. Overall, 1,107 (13.6%) screened positive for PPD (EPDS score ≥10) from 6 weeks to 1 year postpartum. Median (IQR) EPDS score in the patients that screened negative and positive were 3 [1–5], and 12 [11–16], respectively. Patients with an EPDS score ≥10 were more frequently non-White, Hispanic, single, ASA grade ≤3, multiparous, delivered a non-viable fetus, delivered a baby with very low birth weight, and had a psychiatric diagnosis following postpartum EPDS screening. Table 4 shows the proportion of new mental health referrals made to social work and/or psychiatry services from the 6-week postpartum EPDS screen (positive or negative) up to 1 year after childbirth (12.6%; 1024/8155). 45% and 7.5% of patients with a positive and negative EPDS screen were referred, respectively. Table 5 shows the overall number of referrals to social work and/or psychiatry services during the peripartum period (antenatally up to one year following childbirth) and an associated EPDS screen conducted postpartum. Patients were more likely to be referred to either social work or psychiatry services antenatally regardless of EPDS screening status. 59% and 21% of patients with a positive and negative EPDS screen were referred to either of these services, respectively. Table 6 shows the multivariable regression multiple imputed model used to identify factors which are independently associated with referral to social work or psychiatric services. Risk factors for referral to these services included positive EPDS screening score ≥10, widowed, divorced, and separated marital status, ASA grade ≥3, delivery of a non-viable fetus, hospitalization of baby in NICU, a psychiatric diagnosis and psychiatric medication use. A complete case analysis was also performed (Table 7) with consistent findings to the multiple imputed model, except for suicide attempt being an additional risk factor. Table 8 shows the distribution and timings of patient referrals made to social work and psychiatry services over a 1-year period following childbirth. For patients who screened positive (EPDS score ≥10) and were referred, most referrals (social work 84.4%; 347/411 and psychiatry services 53.2%; 115/216) occurred within the first two months postpartum. A minority of referrals occurred after 6 months postpartum (social work 2.7%; 11/411, psychiatry services 9.3%; 20/216). For patients who screened negative (EPDS 0–9) and were referred, most referrals also occurred within the first 2 months postpartum (social work 48.9%; 161/329 and psychiatry services 32.1%; 84/262). However, more than 20% of referrals occurred after 6 months postpartum (social work 21.8%; 72/329, psychiatry services 31.7%; 83/262). Table 9 shows the outcomes following a postpartum EPDS positive and negative screening. More patients with an EPDS score ≥10 were referred to a social worker or medical psychiatry services postpartum, diagnosed with a psychiatric condition, started on psychiatric medication, hospitalized by psychiatry, and/or had a child requiring admission to hospital or a new disease diagnosis. Discussion In our retrospective cohort study, we identified 12.6% of patients with a documented referral to social work and/or psychiatry services following the 6-week postpartum EPDS screen up to 1 year after childbirth. Of the 13.6% of patients who screened positive for PPD (EPDS ≥10), we noted a critical gap; only 45% were referred to social work and/or psychiatry services. While universal screening enables the identification of those at risk of PPD, our findings show that there is a discrepancy between detection and initiation of care, and a potential missed opportunity for improving maternal morbidity and outcomes.13,14 The referral patterns across the peripartum period show that although there were more referrals in the negative EPDS group overall, the timing of referral may be more informative than the absolute number of referrals. Most patients with a negative EPDS screen at 6 weeks who were seen by psychiatry or social work services had been referred antenatally rather than postpartum (951 vs 526). This may suggest that antenatal assessment identified some at-risk patients before postpartum screening, and highlights the importance of antenatal monitoring and referral pathways.12,17 Patient variables associated with specialty referral from 6 weeks to 1 year postpartum included widowed, divorced and separated marital status, ASA grade ≥3, delivery of a non-viable fetus, hospitalization of a baby in NICU, psychiatric diagnosis, and medication use. As referral data was collected solely from within the EHR, we do not know how involved patients were in the referral process. However, we suspect barriers to treating this population exist, not just at the individual level, but at interpersonal, institutional, community and policy levels, which may account for the discrepancy between the proportion of patients that screened positive and the proportion that were seen by specialty services.14,20,21 The majority of women who experience PPD do not enroll in psychiatry services following referral.22 Individual barriers such as lack of knowledge, time management concerns, beliefs on seeking help and resistance to medications play a role, however institutional level barriers such as inadequate referral protocols or Medicaid coverage, community level barriers such as lack of childcare and transportation, and policy level barriers such as insurance status and poverty must also be considered.14,20,21,23,24 While the majority of referrals to social work and psychiatric services occurred within 2 months of childbirth for all screened patients at the postpartum clinic, our study demonstrated that over 20% of patients with a negative EPDS screen were referred to additional services following 6 months postpartum. This finding suggests that patients with a negative screen at 6 weeks may have ongoing social work and/or psychiatric needs in the first year postpartum that require exploration. However, these later referrals may not reflect a missed PPD diagnosis, but rather new or evolving symptoms. Prior data suggests that PPD symptoms may persist or emerge later in the 9–10 month postpartum period and are frequently unreported.25 Delayed referrals may therefore reflect motherhood transitions in the first year postpartum, different clinical phenotypes of PPD, potential false negative screening, under reporting of symptoms to healthcare providers due to stigma, or additional screening opportunities in pediatric and primary care settings at 1,2,4 and 6 months.26–30 We noted a difference in timing between social work and psychiatry referrals. For positive EPDS screened patients, social work referrals were initiated within the first two months postpartum (84%); however, medical psychiatry referrals were less frequently made within this period, with 53% occurring in the same period. The reasons for this may be intentional, as within current collaborative care models, social workers often serve as the first responder to maternal distress; however, it may also be due to systemic waitlist barriers.14,20,31–34 For patients who had a positive EPDS score at the postpartum clinic, approximately a third had a psychiatric diagnosis recorded, yet fewer than 10% of those with a confirmed psychiatric diagnosis were established on pharmacotherapy. This treatment gap is concerning given that the perinatal period carries a high risk of relapse and recurrence for women with mood disorders, particularly when maintenance pharmacotherapy is not established or is discontinued.35–37 While the EPDS is the current best measure to assess PPD, it is only a screening tool and further psychiatric evaluation is required to confirm PPD and determine if treatment is required.10–13 The reasons for this gap are likely multifactorial. Some factors may be healthcare dependent, such as local variation in referral pathways and access to psychiatry services, individual clinician experience and confidence initiating treatment, and whether the patient is within the health care system. 14,20–22,38–40 Other factors may relate to patient preferences and beliefs, including stigma, concerns about medication use while breastfeeding, and preferences for non-pharmacological treatment.38,41–44 We also noted a slightly increased proportion of infants within this screening category admitted to hospital or being diagnosed with a new condition within the year following childbirth. We know that untreated or undertreated PPD can lead to impaired maternal-infant bonding and infant consequences.8,45 While screening for EPDS is essential, it is insufficient unless there are robust pathways to connect patients to care.13 ACOG recommends integrating mental health care into obstetric practice, and a practical solution may be to create standardized and automated pathways instead of relying on passive referral systems that risk missing patients, such as collaborative care models, providing brief interventions for mothers, integrated into the obstetric clinic and therefore preventing patients from being lost to follow-up.31,32,46,47 Future work is needed to understand how best to support this population. Key strengths of our study are the large sample size collected over a decade, therefore allowing us to collect a granular data set, which encompasses a diverse range of participant characteristics, as well as, being able to show both maternal and infant outcomes. We acknowledge that our findings are based on a single center which has individual screening practices and therefore may not be generalizable to the rest of the US.48 However our academic setting likely reflects comparatively good screening and referral practices compared to less well-resourced settings. Due to limitations of the EHR system and ICD-10 coding, misclassification is possible and we were unable to discern if patients attended their appointments and were adherent to treatment.49 Conclusion In summary, 12.6% of patients had a documented referral to social work and/or psychiatry services following the 6-week postpartum EPDS screen up to 1 year after childbirth. Positive EPDS screening was associated with increased social work and psychiatry referrals, psychiatric diagnosis, pharmacotherapy, hospitalization, and infant-associated outcomes. However, a critical gap persists between the proportion screening positive for depressive symptoms and those who receive specialty referral, diagnosis, and treatment. Peripartum referral patterns suggest that antenatal pathways may identify some at-risk patients before postpartum screening, while later referrals among patients who screened negative at 6 weeks suggest that social and psychiatric needs may arise later in the postpartum year. These findings support longitudinal mental health screening and follow-up across the peripartum period.