Trends in Suicidality 1 Year Before and After Birth Among Commercially Insured Childbearing Individuals in the United States, 2006-2017

Key Points

Question  What are the trends in suicidal ideation and intentional self-harm in a large national cohort of commercially insured childbearing individuals?

Findings  In this serial cross-sectional analysis of 595 237 childbearing individuals 1 year before and after giving birth, suicidal ideation and intentional self-harm increased significantly between 2006 and 2017. Non-Hispanic Black individuals, those with low-income, and younger individuals as well as those with comorbid anxiety, depression, or other serious mental illness had larger escalations.

Meaning  Clinical and policy interventions for addressing this health crisis should be tailored to meet the unique needs of childbearing individuals in the year before and following birth, particularly among high-risk groups.

Importance  Suicide deaths are a leading cause of maternal mortality in the US, yet the prevalence and trends in suicidality (suicidal ideation and/or intentional self-harm) among childbearing individuals remain poorly described.

Objective  To characterize trends in suicidality among childbearing individuals.

Design, Setting, and Participants  This serial cross-sectional study analyzed data from a medical claims database for a large commercially insured population in the US from January 2006 to December 2017. There were 2714 diagnoses of suicidality 1 year before or after 698 239 deliveries among 595 237 individuals aged 15 to 44 years who were continuously enrolled in a single commercial health insurance plan. Data were analyzed from October 2019 to September 2020.

Main Outcomes and Measures  The primary outcome was diagnosis of suicidality in childbearing individuals 1 year before or after birth based on the identification of relevant International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) and ICD-10-CM diagnosis codes during at least 1 inpatient or 2 outpatient visits.

Results  Of 595 237 included childbearing individuals, the mean (SD) age at delivery was 31.9 (6.4) years. A total of 40 568 individuals (6.8%) were Asian, 52 613 (8.6%) were Black, 73 172 (12.1%) were Hispanic, 369 501 (63.1%) were White, and 59 383 (9.5%) had unknown or missing race/ethnicity data. A total of 2683 individuals were diagnosed with suicidality 1 year before or after giving birth for a total of 2714 diagnoses. The prevalence of suicidal ideation increased from 0.1% per 100 individuals in 2006 to 0.5% per 100 individuals in 2017 (difference, 0.4%; SE, 0.03; P < .001). Intentional self-harm prevalence increased from 0.1% per 100 individuals in 2006 to 0.2% per 100 individuals in 2017 (difference, 0.1%; SE, 0.02; P < .001). Suicidality prevalence increased from 0.2% per 100 individuals in 2006 to 0.6% per 100 individuals in 2017 (difference, 0.4%; SE, 0.04; P < .001). Diagnoses of suicidality with comorbid depression or anxiety increased from 1.2% per 100 individuals in 2006 to 2.6% per 100 individuals in 2017 (difference, 1.4%; SE, 0.2; P < .001). Diagnoses of suicidality with comorbid bipolar or psychotic disorders increased from 6.9% per 100 individuals in 2006 to 16.9% per 100 individuals in 2017 (difference, 10.1%; SE, 0.2; P < .001). Non-Hispanic Black individuals, individuals with lower income, and younger individuals experienced larger increases in suicidality over the study period.

Conclusions and Relevance  In this cross-sectional study of US childbearing individuals, the prevalence of suicidal ideation and intentional self-harm occurring in the year preceding or following birth increased substantially over a 12-year period. Policy makers, health plans, and clinicians should ensure access to universal suicidality screening and appropriate treatment for pregnant and postpartum individuals and seek health system and policy avenues to mitigate this growing public health crisis, particularly for high-risk groups.

Suicide is a leading cause of maternal mortality in the US.^1-3 Suicide is the second leading cause of death among women aged 25 to 34 years and has steadily increased in prevalence since 2001.⁴ A 2017 report³ aggregating data from 9 maternal mortality review committees taking place between 2008 and 2017 highlighted maternal suicide as an emerging issue and documented that suicide accounted for 6.5% of maternal deaths. Comprehensive estimates of the prevalence and trends in suicide among childbearing individuals remain poorly understood.^5,6 The US Centers for Disease Control and Prevention maternal mortality statistics exclude suicide deaths, deeming them incidental or accidental rather than related to pregnancy.⁷ Researchers and maternal mortality review committees examine severe maternal morbidity, or near misses, to understand how to mitigate maternal deaths. Similar to maternal mortality, psychiatric near misses, such as suicidal ideation or intentional self-harm, remain excluded from standard severe maternal morbidity measures.⁸ We conducted the Maternal Behavioral Health Policy Evaluation (MAPLE) study to characterize prevalence and trends in suicidality (suicidal ideation and/or intentional self-harm) among childbearing individuals 1 year before and following birth.

The MAPLE study evaluated suicidality trends among individuals aged 15 to 44 years in the year before and following birth using Optum Clinformatics Data Mart. These data include medical claims for a large national commercially insured population across all 50 US states. We identified hospital deliveries from January 2006 to December 2017 using standardized International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) and ICD-10-CM diagnosis and procedure codes.⁹ We compiled demographic and clinical characteristics for all individuals and separated by those with and without suicidality diagnoses. We identified clinical comorbidities, including depression, anxiety, bipolar disorder, psychosis, and substance use disorders, using Healthcare Cost and Utilization Project algorithms based on ICD-9-CM and ICD-10-CM codes.^10,11 We identified suicidality by a diagnosis of suicidal ideation and/or intentional self-harm occurring in either 2 outpatient or 1 inpatient diagnostic fields in the Clinformatics Data Mart. The University of Michigan Institutional Review Board approved this study, which included a waiver of informed consent due to use of deidentified patient data.

We used a series of generalized estimating equations, clustering at the individual level, to assess changes in suicidality between 2006 and 2017 among the overall sample and subsamples stratified by demographic measures and psychiatric comorbidity. Overall sample models for suicidal ideation, intentional self-harm, and suicidality included year indicators only. Subsample models for suicidality included year indicators, the subsample demographic characteristic or psychiatric comorbidity, and the interaction between year and subsample indicators. We used predictive margins to compare suicidality between 2 time points, 2006 and 2017.

We used 2-sided statistical tests and an α level of .05 to determine statistical significance. We performed claims data management in SAS version 9.2 (SAS Institute) and statistical analyses in Stata version 14.1 (StataCorp).

This study included 698 239 deliveries between 2006 and 2017 (range, 52 844 to 64 791 per year) among 595 237 commercially insured childbearing individuals. The cohort mean (SD) age at delivery was 31.9 (6.4) years. A total of 40 568 individuals (6.8%) were Asian, 52 613 (8.6%) were Black, 73 172 (12.1%) were Hispanic, 369 501 (63.1%) were White, and 59 383 (9.5%) had unknown or missing race/ethnicity data. A total of 2714 deliveries among 2683 individuals coincided with indications for either suicidal ideation or intentional self-harm (range, 103 to 320 per year). Overall diagnoses of suicidal ideation increased from 0.1% per 100 individuals in 2006 to 0.5% per 100 individuals in 2017 (difference, 0.4%; SE, 0.03; P < .001) (Figure). Diagnoses of intentional self-harm increased from 0.1% per 100 individuals in 2006 to 0.2% per 100 individuals in 2017 (difference, 0.1%; SE, 0.02; P < .001). Finally, diagnoses of suicidality increased from 0.2% per 100 individuals in 2006 to 0.6% per 100 individuals in 2017 (difference, 0.4%; SE, 0.04; P < .001).

Suicidality diagnoses commonly appeared in the fourth or higher diagnostic field in the Clinformatics Data Mart (4860 of 9414 suicidality diagnoses [51.6%]), but higher proportions of suicidality diagnoses appeared in the primary diagnostic field over time (total, 606 of 9414 diagnoses [6.4%]; 2006, 5 of 157 diagnoses [3.2%]; 2017, 217 of 1748 diagnoses [12.4%]). Of 2714 suicidality diagnoses, 1225 (45.1%) occurred during the predelivery period and 1592 (58.7%) occurred in the postdelivery period. We identified relatively few instances (103 of 2714 [3.8%]) where suicidality appeared in both time periods for a given individual.

Demographic characteristics differed proportionally for deliveries with associated diagnoses of suicidality 1 year before or after birth and those without (Table 1). Suicidality was more commonly diagnosed 1 year before or after birth for those aged 15 to 26 years. Higher proportions of suicidality diagnoses were also found one year before or after delivery for non-Hispanic Black individuals and for individuals with 400% or less of the federal poverty level (FPL). Of 2714 suicidality diagnoses, 1881 (69.3%) had comorbid anxiety diagnoses, 2334 (86.0%) had comorbid depression diagnoses, and 2482 (91.5%) had either comorbid anxiety or depression diagnoses. Further, there were 855 diagnoses of suicidality (31.5%) with comorbid bipolar disorder, 87 (3.2%) with a comorbid psychotic disorder, and 881 (32.5%) with either comorbid bipolar disorder or a comorbid psychotic disorder. Finally, 1001 (36.9%) of those experiencing suicidality had a least 1 diagnosis indicating the presence of a comorbid substance use disorder. Diagnoses of depression and/or anxiety preceded diagnoses of suicidality in 2587 of 3841 instances (67.4%). When considering bipolar, psychotic, and substance-related disorders in addition to depression and/or anxiety, there were prior mental health diagnoses in 2799 of 3841 instances of diagnosed suicidality (72.9%). Diagnoses of depression and/or anxiety increased steadily over the study period from a prevalence of 12.1% per 100 individuals in 2006 to 20.9% per 100 individuals in 2017. In contrast, bipolar and psychotic disorders remained relatively stable over the study period, identified among approximately 1% of the cohort. Substance-related disorders also remained stable, affecting approximately 3% to 4% of the cohort over the study period (Table 1).

Suicidality prevalence varied by subgroups, and nearly all subgroups experienced an increase in diagnoses of suicidality between 2006 and 2017. The diagnoses of suicidality increased from 1.6% per 100 individuals to 9.5% per 100 individuals among individuals aged 15 to 18 years (difference, 7.9%; SE, 1.7; P < .001) and from 0.3% per 100 individuals to 1.9% per 100 individuals among individuals aged 19 to 26 years (difference, 1.6%; SE, 0.2; P < .001) (Table 2). Among all racial/ethnic groups except for Asian individuals, diagnoses of suicidality increased: for non-Hispanic Black individuals from 0.2% per 100 individuals to 0.9% per 100 individuals (difference, 0.7%; SE, 0.2; P < .001); for non-Hispanic White individuals from 0.2% per 100 individuals to 0.6% per 100 individuals (difference, 0.5%; SE, 0.1; P < .001); and for Hispanic individuals from 0.2% per 100 individuals to 0.6% per 100 individuals (difference, 0.4%; SE, 0.1; P < .001). Among individuals with household incomes of 400% or less of the FPL, diagnoses of suicidality increased from 0.2% per 100 individuals in 2006 to 0.8% per 100 individuals in 2017 (difference, 0.6%; SE, 0.1; P < .001) compared with a smaller increase among those with incomes greater than 400% of the FPL from 0.1% per 100 individuals in 2006 to 0.4% per 100 individuals in 2017 (difference, 0.2%; SE, 0.1; P < .001).

Diagnoses of suicidality with comorbid depression or anxiety increased from 1.2% per 100 individuals in 2006 to 2.6% per 100 individuals in 2017 (difference, 1.4%; SE, 0.2; P < .001), with similar patterns for comorbid depression and for comborbid anxiety. Diagnoses of suicidality with comorbid bipolar disorder or psychosis individuals increased from 7.1% per 100 individuals in 2006 to 16.9% per 100 individuals in 2017 (difference, 9.9%; SE, 2.0; P < .001). Diagnoses of suicidality with comorbid bipolar disorder increased from 6.9% per 100 individuals in 2006 to 16.9% per 100 individuals in 2017 (difference, 10.0%; SE, 0.2; P < .001) and suicidality with comorbid psychotic disorders 7.1% per 100 individuals in 2006 to 47.6% per 100 individuals in 2017 (difference, 40.5%; SE, 11.9; P < .001). Diagnoses of suicidality with comorbid substance use disorder increased from 1.8% per 100 individuals in 2006 to 9.3% per 100 individuals (difference, 7.5%; SE, 4.4; P = .09).

We found that diagnoses of suicidality in a large commercially insured US population increased steadily between January 2006 and December 2017. Maternal health experts note that the lack of consistent, inclusive data on suicidality before and after birth is alarming.⁶ Pregnancy-related suicides occur in 1.6 to 4.5 per 100 000 live births in state samples.^5,12 Further, suicide near misses, such as suicidal ideation and intentional self-harm, remain excluded from standard measures of severe maternal morbidity.¹³ Given the severe maternal mortality crisis among racial/ethnic minority individuals, especially Black women, it is imperative to include psychiatric risks in predictive models and practice guidelines. Similar to our findings, prior studies have found that most pregnancy-associated suicide deaths occur among older and non-Hispanic White women.^2,5,12 However, the present study adds evidence of a sharp increase in suicidality among younger and non-Hispanic Black women over a recent 12-year period that should inform intervention efforts.

To our knowledge, this is the first study to quantify the diagnosed risk of suicidality in the year prior to and following birth among individuals with mood disorders. Mood disorders represent a key risk factor for suicide.^3,14 Medication discontinuation, lack of ongoing treatment, and intimate partner violence are further risk factors for suicide among those with mood disorders.¹⁴ Further, that nearly one-quarter of those with bipolar disorder and half of those with psychotic disorders experienced suicidality in the year before or following birth is striking. Although differential effects of treatment, or lack thereof, are outside the scope of our analysis, this remains an important direction for future work.

Our study had limitations. We could not discern whether the increases in suicidality identified in this study represent increases in prevalence or in detection. Some portion of the increase identified could have resulted from decreased stigma regarding suicide in the general population and a concomitant increase in clinicians assessing and patients reporting suicidality. Documented increases in both maternal mood disorders and suicide among women of reproductive age support the findings of increases in the intermediate outcome of suicidality. Future work should assess whether the prevalence of suicidality increased, the clinical detection of suicidality improved, or both.

This study used data from a single large commercial insurance provider. Detection of suicidality may differ in other populations and settings. Medicaid populations may face higher risks of suicidality given the higher prevalence of many risk factors,^15,16 making our estimates conservative. However, Clinformatics Data Mart includes beneficiary data from numerous health plans with different benefit structures serving geographically and economically diverse populations. Additionally, examining suicidality in the year before and after birth necessarily excluded individuals who died from suicide within that time frame. Although this cohort likely included a small number of deceased individuals, survivor bias may have influenced our findings, also yielding more conservative estimates.

In this cross-sectional study of US childbearing individuals, identification of suicidal ideation and intentional self-harm occurring in the year preceding or following birth increased substantially over the 12-year study period. It remains unclear how much of this increase represents increases in disease burden compared with improvements in clinical detection. Nevertheless, policy makers, health plans, and clinicians should ensure access to universal suicidality screening and appropriate treatment for pregnant and postpartum individuals and seek health system and policy avenues to mitigate this public health crisis, particularly for high-risk groups.

Accepted for Publication: September 11, 2020.

Published Online: November 18, 2020. doi:10.1001/jamapsychiatry.2020.3550

Corresponding Author: Kara Zivin, PhD, MS, MA, Department of Psychiatry, University of Michigan, North Campus Research Complex, 2800 Plymouth Rd, Bldg 16, Room 228W, Ann Arbor, MI 48109 (kzivin@umich.edu).

Author Contributions: Dr Zivin had full access to all of the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.

Concept and design: Admon, Dalton, Kolenic, Ettner, Haffajee, Zivin.

Acquisition, analysis, or interpretation of data: All authors.

Drafting of the manuscript: Admon, Kolenic, Tabb.

Critical revision of the manuscript for important intellectual content: Admon, Dalton, Ettner, Tilea, Haffajee, Brownlee, Zochowski, Tabb, Muzik, Zivin.

Statistical analysis: Admon, Dalton, Kolenic, Ettner, Tilea, Brownlee.

Obtained funding: Zivin.

Administrative, technical, or material support: Zochowski, Zivin.

Supervision: Dalton, Zivin.

Conflict of Interest Disclosures: Drs Ettner and Tilea received grants from the National Institutes of Health during the conduct of the study. Dr Dalton received grants from the National Institutes of Health during the conduct of the study as well as grants from the Agency for Healthcare Research and Quality, the American Association of Obstetricians and Gynecologists, the Arnold Foundation, and the Blue Cross Blue Shield of Michigan Foundation and personal fees for serving on the clinical advisory board from Bind Health Insurance and for serving as a contributing editor from The Medical Letter outside the submitted work. Dr Zivin received grants from the National Institutes of Health, the Department of Veterans Affairs, and Blue Cross Blue Shield of Michigan Foundation during the conduct of the study. No other disclosures were reported.

Funding/Support: This research was supported by grant R01MH120124 from the National Institute of Mental Health and the Office of Research on Women’s Health.

Role of the Funder/Sponsor: The sponsors had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

Meeting Presentation: This article was presented at the virtual AcademyHealth 2020 Annual Research Meeting; July 30, 2020.