Bipolar disorder is a severe mental illness that affects approximately 3% of the general population^1-3 and typically requires lifelong treatment, including pharmacotherapy.^2,4-6 It is most often diagnosed between the ages of 18-30² years and is episodic, with exacerbations often presenting in times of stress.⁷ Among women, the highest lifetime risk for first onset of bipolar disorder or mood episode recurrence occurs in the perinatal period, defined as during pregnancy or within one year postpartum.^8-10 Estimates of women with bipolar disorder that experience a mood episode during the perinatal period range from 60-70%.^11,12

A diagnosis of bipolar disorder places both the mother and baby at increased risk of adverse outcomes, particularly when left untreated.¹³ Women with untreated bipolar disorder have an increased risk of gestational hypertension and antepartum hemorrhage,² as well as self-injury, substance use, and suicide.^10,14,15 Infants may also be affected by a mother with untreated bipolar disorder; they are more likely to be born small for their gestational age,² have elevated levels of fetal stress hormones, and have impaired mother-baby bonding.^10,16 Bipolar disorder is a known risk factor for postpartum psychosis,¹⁷ which carries a risk of suicide and infanticide.^12,14,18 Given the risks of untreated bipolar disorder, the benefits of pharmacotherapy with mood-stabilizing drugs typically outweigh the risks.^19-21 Thus, treatment initiation or continuation is recommended for most perinatal women with bipolar disorder, and discontinuation of pharmacotherapy should be discouraged.⁶

Despite associations with negative consequences, bipolar disorder in the perinatal period is underdiagnosed and undertreated.^15,22 Professional societies and expert committees recommend universal screening of pregnant and postpartum women in the obstetric setting for perinatal mood and anxiety disorders, such as perinatal depression.^19-21,23,24 The Council on Patient Safety in Women’s Health Care, a multidisciplinary group of organizations that aims to improve health care for women coordinated by the American College of Obstetricians and Gynecologists (ACOG), recommends that obstetric providers should consider screening for bipolar disorder when women meet screening criteria for major depressive disorder, particularly before initiating pharmacotherapy treatment for perinatal depression.²⁰ Lack of bipolar disorder screening can result in women that are mistakenly thought to have depression, but actually have bipolar disorder, receiving antidepressant monotherapy.^25-27 Antidepressants are generally not recommended as monotherapy for bipolar disorder; starting an unopposed antidepressant in a woman with underlying bipolar disorder rather than unipolar depression can exacerbate disease by precipitating mania and psychosis.^28,29 Appropriate screening practices and subsequent assessment and connection to care as indicated may help ensure that perinatal women who screen positive for bipolar disorder receive established, evidence-based treatments.^2,4,15,30

Despite recommendations for obstetric providers to screen for bipolar disorder in the perinatal period,²⁰ there is a dearth of information on the rates of positive bipolar disorder screens, associated factors, and treatment participation. These data have important implications and the potential to help inform the process of incorporating bipolar disorder screening into the work flow of busy obstetric practices.

The objectives of this study were to describe: (1) proportion of perinatal women who screen positive for bipolar disorder in the obstetric setting; (2) demographic characteristics associated with positive bipolar disorder screens among perinatal women; and, (3) the associations of positive bipolar screens with clinical characteristics and healthcare utilization.

This study was conducted with a sample of English-speaking pregnant and postpartum women ages 18–45 years, recruited from May 2016 - June 2018. Women were recruited during pregnancy and up to three months after giving birth from 14 geographically diverse obstetric clinics across the Commonwealth of Massachusetts. The study was approved by University of Massachusetts Medical School Institutional Review Board. All women gave informed consent prior to participating.

Recruitment was conducted as part of a larger longitudinal study, the PRogram In Support of Moms (PRISM),^25,26 for which the methods are described elsewhere (Clinical Trials #: ).³¹ Women were recruited through patient lists that were compiled by participating obstetric practices and shared with the study team via secure file transfer. Study staff contacted and prescreened patients based on their demographic and clinical information, including age (18-45 years), gestational age (greater than four weeks post-conception through three months postpartum), and language (able to communicate in written and spoken English). Subsequently, a Research Coordinator telephoned prescreen-eligible patients to gain informed consent and to screen patients for full eligibility. Of contacted patients, 25.6% of patients consented to participate in the study. Of these consented participants, the response rate for the study data described herein was 87.9%.

Study data were collected via telephone interviews consisting of: (1) structured questions regarding obstetric and psychiatric care; (2) validated screening instruments for depression, bipolar disorder, and substance use; (3) demographic questions; (4) questions regarding factors associated with mental health treatment participation (e.g. “In the past 3 months, have you seen, or talked on the telephone, to any of the following professionals about your emotions or mental health?” “In general, how satisfied are you with the way the provider handled your emotional/mental health concerns during the past 3 months?” and “Are you taking any psychiatric medications now?”); and, (5) questions regarding unmet treatment needs (e.g., “During the past 3 months, was there ever a time when you felt that you needed help for your emotions, mental health but you didn't receive it?’).

Women were screened for depression using the Edinburgh Postnatal Depression Scale (EPDS),³⁷ a validated 10-item screening questionnaire that is widely used to assess depression during pregnancy and the postpartum period.³² Scores range from 0 to 30; possible depression is suggested by scores ranging 9 to 13. Scores of 10 or greater were considered positive depression screens.³² All women were screened for lifetime presence of bipolar disorder with the Mood Disorder Questionnaire (MDQ)³³ and for active substance use with the Parents, Partners, Past, and Pregnancy screen (the 4Ps).³⁴ Both the MDQ and 4Ps tools result in a binary positive or negative score. The MDQ is a 5-item screening questionnaire that is a widely accepted and validated tool for screening for bipolar disorder, including during pregnancy.^16,35,36 The first question screens for the lifetime prevalence of 13 symptoms associated with bipolar disorder (see Supplementary Table 1). Questions two and three determine the timeframe of the symptoms (they must occur together) and impairment to life, as based on the Diagnostic and Statistical Manual of Mental Disorders (DSM) 5 criteria.³⁷ The fourth question was to used to determine if the participant any family history of bipolar disorder and fifth question asks if the participant has ever been personally diagnosed with bipolar disorder by a health professional (Supplementary Table 1). To be included in the larger PRISM study, women must screen positive for depression on the EPDS, and negative on both the MDQ and 4Ps; however, for this study we looked at all women that prescreened as eligible, enrolled, and had complete data on the EPDS and MDQ, with a specific analytic focus on the MDQ screening results.

Several approaches have been used to score the MDQ.^35,36,38,39 The original method used the following threshold to denote a positive screen: (1) presence of seven or more symptoms in question one (Supplementary Table 1), (2) concurrent occurrence of said symptoms (question two), and (3) a resultant moderate or serious impairtment to their life (question three).³³ Revised scoring methods have been devised and tested to improve the sensitivity, while maintaining the high level of specificity.^35,40 Because of this, we opted to use a revised scoring method in this study. This method uses only the first question of the MDQ to determine screening status, meaning that women had to report a minimum of seven symptoms to result in a positive screen.^36,40 We chose this method because it achieves a sensitivity of 89% and a specificity of 84%.³⁵ Additionally, its Positive Predictive Value is 0.43, also greater than that of the original method.³⁵ Other groups have had similar success using this revised scoring method, in terms of rates of sensitivity and specificity of detecting bipolar disorder against the gold standard of the Structured Clinical Interview for DSM-IV Axis I disorders (SCID-I).^36,40

Women in the perinatal period are at heightened risk for new onset bipolar disorder or disease recurrence^8-10 and prior estimates suggest that three-quarters of those women will go on to have an illness which requires life-long management.^18,42,43 Given their frequent contact with healthcare providers during pregnancy, the perinatal period is a time of great potential for identifying new diagnoses or to reach at risk women who might not otherwise be identified. Obstetric providers are encouraged to universally screen for perinatal mood and anxiety disorders with validated tools, including to screen for bipolar disorder prior to initiating pharmacotherapy. However, what is less clear is how to connect perinatal women to mental health resources and care when indicated.

In comparison to previously published literature, positive bipolar disorder screening rates were higher in our sample of perinatal women. The few prior studies conducted in this domain have found estimates of 5-12%, depending on the MDQ scoring and recruitment methods used,^16,35,40 compared to our 18.8%. Of note, our sensitivity analysis calculations (using the original MDQ method) did yield an estimate that is more comparable to prior studies (8.7%), suggesting the importance of the MDQ scoring methods. A higher proportion of women who screened positive for bipolar disorder had lower education status and prenatal care covered by public insurance, both shown to be associated with adverse pregnancy outcomes irrespective of comorbid mental health conditions.^44,45 One of the most striking associations with screening positive for bipolar disorder in this study was the high rate of positive current substance use screens (18.5 times higher among women with bipolar disease). Though it is well known that women with bipolar disorder have high rates of comorbid substance misuse,^46-48 our findings bolster the importance of universally screening patients for active substance use during pregnancy, particularly when screened positive for other psychiatric illnesses like bipolar disorder. Universal screening for substance use disorders in pregnancy is currently recommended by ACOG, the Centers for Disease Control and Prevention, and American Medical Association.^20,49

We found that women with positive bipolar disorder screens, compared to those with negative screens, more frequently discuss their mental health with healthcare professionals, were offered treatment or treated previously, and were in some form of psychiatric treatment at the time of the study. However, about 1 in 4 women (24.0%) with positive bipolar disorder screens reported feeling that they needed psychiatric help but were not receiving it, which was significantly higher than those that did not screen positive (7.0%). Interestingly, women with and without a pre-existing bipolar diagnosis were reported feeling that they needed more psychiatric help at very similar rates (22.6% vs. 24.6%, respectively). This is not surprising in light of the aforementioned suboptimal rates of psychiatric treatment during the perinatal period even in those with previously diagnosed bipolar disorder, and it aligns with prior research.^2,4,15,30 However, this may also be due to alternative, unexamined factors. For example, other studies have found that pregnant women are not always given a choice by their psychiatric provider as to whether they may continue medication during pregnancy²² and, therefore, often go without any psychiatric treatment or support throughout their pregnancy.¹⁵ These treatment patterns would be expected to lead to patient dissatisfaction and to worse outcomes.

In our study, less than half (42%) of those who screened positive for bipolar disorder were receiving any psychiatric treatment (i.e., psychiatric medication with or without psychotherapy). Because we were using screening rather than diagnostic tools, these patients may or may not be undertreated as they require further diagnostic assessment. However, we found similar patterns among the subset of those with an existing bipolar diagnosis: less than two-thirds (61%) were receiving any current treatment. More critically, only 29% of those with an existing bipolar diagnosis were receiving treatment that included psychiatric medication; low treatment rates were also reflected in those who screened positive for bipolar disorder (20%). Further, treatments and psychiatric medications discussed in the questionnaire were not specific to bipolar disorder and, therefore, these could well be overestimates of what psychiatric providers would consider adequate treatment for bipolar disorder.

Given the low treatment rates we observed in patients with existing bipolar disorder, the illness appears to be undertreated and/or treatment adherence is suboptimal in the perinatal period and that there is a significant gap between needed care and the actual care received for bipolar disorder. This is alarming because bipolar disorder imparts a high-risk status to a woman, her pregnancy, and her child.¹⁶ Women may be asking for help but being stymied by long wait times to get in to see psychiatric providers or hesitancy amongst providers to treat women during the perinatal period. Our prior work indicates many barriers to psychiatric treatment during the perinatal period even when it is requested by the patients themselves, including that, in some cases, women are being instructed to discontinue their medications by mental health providers.⁵⁰ It is recommended that all perinatal women with bipolar disorder be treated with, at minimum, psychiatric mediations during the perinatal period.^4,6 Treatment of bipolar disorder in perinatal women is undoubtedly challenging due to: (1) hesitancy among medical providers to provide pharmacotherapy during pregnancy; (2) controversies and misconceptions regarding the risks and benefits of medication use in pregnancy and lactation; and, (3) stigma associated with pharmacotherapy during pregnancy.^15,22 However, given that psychiatric medications have the strongest evidence for bipolar disorder treatment, this is an important practice gap that needs to be addressed as part of comprehensive obstetric care.

A limitation of this study is that our data rely on the use of screening tools rather than diagnostic interviews; therefore, we were not able to confirm the potential cases of bipolar disorder. We conducted a sub-analysis of those women with a prior diagnosis of bipolar disorder to examine if the patterns and associations observed among MDQ positive screens held and found that, with few exceptions, those with and without a prior diagnosis were fairly similar with respect to the variables we examined. Additionally, the MDQ is a widely accepted and well validated tool to screen for bipolar disorder^16,35,36 and it is noteworthy that we used a scoring methodology with the highest validated sensitivity (89%) and specificity (84%) for diagnostic-interview confirmed bipolar disorder.^3,35,36,40 The sensitivity analyses using the original scoring method showed that, regardless of screening threshold used, the relationships between the variables in our data were robust and remained consistent, suggesting real differences between the groups. The MDQ has been associated with over-identification of bipolar disorder in some cases,^40,51 which likely also contributes to our higher rates of positive bipolar screens. Additionally, the symptomatology covered on the MDQ is known to overlap with other psychiatric illnesses, which may explain some of the high degree of correlation between the positive bipolar and substance use screens. However, when we examined these same correlations with just the subset of those with a known history of bipolar disorder, the relationships remained. Additionally, the use of screening tools rather than full diagnostic interviews allowed us to obtain data from a large sample of women with minimal time burden on the behalf of the participants and the study team.

To our knowledge, our study is one of the only to date that examines use of the MDQ and its multiple scoring methodologies and evaluates screening rates in perinatal women in an ongoing clinical trial that includes rich demographic and clinical data. While our data rely mostly on participant self-report, which is subject to recall bias, the use of multiple interviews with participants in the ongoing parent study allows for the opportunity for follow-up and future examination. Our study sample was recruited from a diverse set of practices in Massachusetts, representative of a wide array of patient experiences among the state’s perinatal population, as reflected in population level data.^41,52 However, generalizability is limited to Massachusetts, which has higher rates of insurance and better access to mental health providers than most states. Therefore, there is a need for further research in other settings and populations.

Creating streamlined approaches that help obstetric settings identify potential bipolar disorder and then connect patients to appropriate assessment and follow-up are difficult due to the already immense responsibilities of these providers and practices. However, doing so has the potential to affect the lives and mental health of women that may never otherwise get treatment. Integrating depression screening into obstetric care is feasible⁵³ and could be expanded to include detection of bipolar disorder symptomatology and subsequent referral for assessment and treatment as indicated. Our data presented here support the recommendations from the Consensus Bundle on Maternal Mental Health, suggesting that all perinatal women should be screened for bipolar disorder before initiating any treatment for depression. We also found high rates of positive MDQ screens, consistent with other research.^16,51 In addition, our work builds on this existing literature as we have also identified an assessment and treatment gap, suggesting that more work is needed to not only help obstetric practices screen²⁰ but also to respond to positive screens and connect women with further assessment and treatment when indicated.