The Association between NICU Admission and Mental Health Diagnoses among Commercially Insured Postpartum Women in the US, 2010–2018

Maternal mental health (MH) conditions represent a leading cause of preventable maternal death in the US. Neonatal Intensive Care Unit (NICU) hospitalization influences MH symptoms among postpartum women, but a paucity of research uses national samples to explore this relationship. Using national administrative data, we examined the rates of MH diagnoses of anxiety and/or depression among those with and without an infant admitted to a NICU between 2010 and 2018. Using generalized estimating equation models, we explored the relationship between NICU admission and MH diagnoses of anxiety and/or depression, secondarily examining the association of NICU length of stay and race/ethnicity with MH diagnoses of anxiety and/or depression post NICU admission. Women whose infants became hospitalized in the NICU for <2 weeks had 19% higher odds of maternal MH diagnoses (aOR: 1.19, 95% CI: 1.14%–1.24%) and those whose infants became hospitalized for >2 weeks had 37% higher odds of maternal MH diagnoses (aOR: 1.37 95% CI: 1.128%–1.47%) compared to those whose infants did not have a NICU hospitalization. In adjusted analyses, compared to white women, all other race/ethnicities had significantly lower odds of receiving a maternal MH condition diagnosis [Black (aOR = 0.76, 0.73–0.08), Hispanic (aOR = 0.69, 0.67–0.72), and Asian (aOR: 0.32, 0.30–0.34)], despite higher rates of NICU hospitalization. These findings suggest a need to target the NICU to improve maternal MH screening, services, and support while acknowledging the influence of social determinants, including race and ethnicity, on health outcomes.


Introduction
Maternal mental health (MH) conditions represent a leading cause of preventable maternal mortality in the US [1], implicated in 1 in 9 maternal deaths [2]. Depression and anxiety affect approximately 15% of childbearing women nationally [3,4], and MH symptoms during the perinatal period, such as suicidal ideation and self-harm, have increased over time [5,6]. Traumatic, financial, emotional, or stressful partner-related life events during the antenatal period increase the risk of postpartum depression [7]. such as race/ethnicity (Black, Hispanic, and Asian versus white), retained an association with maternal MH after controlling for NICU experiences.

Materials and Methods
This retrospective cohort study evaluated the prevalence of maternal MH condition diagnoses of anxiety and/or depression identified within the year following birth among women aged 15-44 with and without an infant admitted to the NICU using Optum's deidentified Clinformatics ® Data Mart Database (CDM). CDM includes a statistically deidentified large claims data warehouse of administrative health claims from all 50 states. We identified postpartum women from 2010 to 2018 and only included those who had continuous enrollment in a single employer-based health plan for at least 1 year before and 1 year after a live birth. For these women, using a family identifier variable, we linked family members, and using year of birth, we identified newborns within a family. We restricted our analytical cohort to postpartum women with linked newborns (Figure 1). such as race/ethnicity (Black, Hispanic, and Asian versus white), retaine with maternal MH after controlling for NICU experiences.

Materials and Methods
This retrospective cohort study evaluated the prevalence of matern diagnoses of anxiety and/or depression identified within the year follow women aged 15-44 with and without an infant admitted to the NICU deidentified Clinformatics ® Data Mart Database (CDM). CDM includ deidentified large claims data warehouse of administrative health cla states. We identified postpartum women from 2010 to 2018 and only inc had continuous enrollment in a single employer-based health plan for at le and 1 year after a live birth. For these women, using a family identifier va family members, and using year of birth, we identified newborns within stricted our analytical cohort to postpartum women with linked newborns We identified women whose infants had a NICU admission (CPT co 99477, 99478, 99479, 99480) and who had MH diagnoses of anxiety and/o Appendices A-C) up to one year or anytime in the year following deliv We identified women whose infants had a NICU admission (CPT codes 99468, 99469, 99477, 99478, 99479, 99480) and who had MH diagnoses of anxiety and/or depression (see Appendices A-C) up to one year or anytime in the year following delivery using standardized International Classification of Disease, 9th and 10th Revision, Clinical Modifi-Children 2022, 9, 1550 4 of 13 cation diagnosis codes present at least once in inpatient claims or twice in outpatient claims. We selected sociodemographic covariates based upon factors established to influence the relationship between NICU hospitalization and maternal MH [10][11][12][13][14][16][17][18][19]. Covariates included age (≤18, 19-26, 27-34, 35-39, ≥40), race/ethnicity (Asian, Black, Hispanic, Unknown race/ethnicity, white), region (Midwest, West, Northeast, South), insurance type (Point of Service, Exclusive Provider Organization/Health Maintenance Organization, Preferred Provider Organization, other), other MH conditions, and substance use disorders. We identified other MH conditions not related to anxiety or depression, such as bipolar disorder, schizophrenia, and other conditions, as well as substance use disorder conditions such as alcohol, tobacco, cannabis, and other conditions using ICD-9 and ICD-10 codes (see Appendices B and C, respectively). We used a similar algorithm of one inpatient or two outpatient claims to assess the prevalence of other MH conditions or substance use disorders. The University of Michigan Institutional Review Board (HUM00188304) approved this study.

Statistical Analysis
Using means or proportions with the associated 95% confidence intervals we evaluated demographic and clinical characteristics including age, race/ethnicity, region, insurance type, other MH conditions, and substance use disorders for all women, and for postpartum women whose infants had a NICU admission for <14 days or ≥14 days for the years between 2010 and 2018. Using all study period data (2010-2018), we examined trends in NICU admission rates, overall, by maternal MH status and by postpartum individual's race/ethnicity. NICU admission was the primary predictor of interest and maternal MH diagnoses of anxiety and/or depression the primary outcome; secondarily, we explored maternal MH by NICU length of stay and race/ethnicity. We used generalized estimating equation (GEE) models with an exchangeable covariance structure to control for repeating deliveries (women who gave birth more than once during the years 2010-2018) to explore the association between maternal MH diagnoses of anxiety and/or depression and NICU admission adjusting for the following covariates chosen a priori: delivery year, postpartum individual's age, race/ethnicity, insurance plan type, region, other MH conditions, and substance use disorders. We conducted all analyses and data management using SAS v9.4 (Cary, NC, USA).
of Black women had a diagnosis of postpartum anxiety and/or depression in 2010 and 11.9% (5% CI: 10.9%-13.0%) in 2018. Between 2010 and 2018, rates of NICU admission increased among all infants, regardless of race/ethnicity, including: Black infants from 10.2% (95% CI: 9.3%-11.1%) to 12.1% (95% CI: 11.1%-13.2%); Asian infants from 8.7% (95% CI: 7.8%-9.5%) to 10.9% (95% CI: 10.0%-11.8%); Hispanic infants from 9.0% (95% CI: 8.3%-9.7%) to 10.6% (95% CI: 9.9%-11.4%); and white infants from 8.6% (95% CI: 8.3%-8.9%) to 10.0% (95% CI: 9.7%-10.4%; Figure 3). After adjusting for delivery year, age, race/ethnicity, region, insurance, other MH conditions, and substance use disorders, postpartum women whose infants experienced NICU admission had 23% higher odds (aOR: 1.23, 95% CI: 1.19%-1.27%) of an MH diagnosis of anxiety and/or depression than those women without NICU admission (Table 2a). Furthermore, after restricting the analysis to women who did not have a MH diagnosis in the prenatal period, for women whose infant's NICU experience lasted less than two weeks, we observed a 19% increase in odds of postpartum MH condition diagnoses of anxiety and/or depression (aOR: 1.19, 95% CI: 1.14%-1.24%; Table 2b) compared to women with no NICU admission. For women whose infant's length of stay exceeded two weeks, we observed a 37% increase in odds of postpartum anxiety and/or depression (aOR: 1.37, 95% CI: 1.13%-1.47%) compared to women who did not have an infant with a NICU admission (Table 2b). After adjusting for delivery year, age, race/ethnicity, region, insurance, other MH conditions, and substance use disorders, postpartum women whose infants experienced NICU admission had 23% higher odds (aOR: 1.23, 95% CI: 1.19%-1.27%) of an MH diagnosis of anxiety and/or depression than those women without NICU admission (Table 2a). Furthermore, after restricting the analysis to women who did not have a MH diagnosis in the prenatal period, for women whose infant's NICU experience lasted less than two weeks, we observed a 19% increase in odds of postpartum MH condition diagnoses of anxiety and/or depression (aOR: 1.19, 95% CI: 1.14%-1.24%; Table 2b) compared to women with no NICU admission. For women whose infant's length of stay exceeded two weeks, we observed a 37% increase in odds of postpartum anxiety and/or depression (aOR: 1.37, 95% CI: 1.13%-1.47%) compared to women who did not have an infant with a NICU admission (Table 2b). In adjusted analyses, compared to white women, other race/ethnicities had lower odds of a postpartum MH condition diagnosis of anxiety and/or depression: Black (aOR: 0.76, 95% CI: 0.74%-0.79%), Hispanic (aOR: 0.69, 95% CI: 0.66%-0.71%), and Asian (aOR: 0.35, 95% CI: 0.34%-0.37%), shown in Table 2b. This trend of lower odds of a postpartum MH condition diagnosis in racial/ethnic minorities was similar when the analyses were not restricted to women who did not have a MH diagnosis in the prenatal period.

Discussion
This multi-year study of a commercially insured US population found that postpartum women whose infants had NICU hospitalization had higher odds of an anxiety and/or depression diagnosis in the year following this experience compared to those who did not have an infant hospitalized in the NICU. Trends in NICU admission increased slightly during the study period, consistent with national averages of NICU stays [21]. Incidence of postpartum MH diagnoses of anxiety and/or depression increased during the study period. Odds of MH diagnoses of anxiety and/or depression increased as the length of stay in the NICU increased. These findings echo results of smaller studies that found NICU admission influenced MH symptoms of depression and anxiety among delivering women [10][11][12][13][14][15][16][17][18][19]. Our findings extend the results of these smaller studies expanding the sample size, region, and racial/ethnic diversity of women included for analysis in NICUs in the US. Through conducting an additional analysis that only includes postpartum women without prenatal MH conditions, we further confirm the relationship between NICU admission and incidence of MH diagnoses of anxiety and/or depression among postpartum women.
Longer length of NICU stay translated to higher risk for postpartum MH diagnoses of anxiety and/or depression in the sample of women without prenatal MH conditions. Postpartum women with infants in the NICU for less than two weeks had increased odds of MH diagnoses of anxiety and/or depression. These odds nearly doubled when NICU length of stay exceeded two weeks. Our study found that any NICU hospitalization increased postpartum MH diagnoses, but given the range of infant health severity in the NICU, a longer infant length of stay appears to be associated with a higher risk of maternal MH conditions. Using NICU resources to screen for and address the MH of postpartum women may prove beneficial, and women with infants in the NICU for prolonged periods of time may have an increased need for support [24]. Several studies indicate the benefit of individualized interventions using trauma-informed modalities to address the health of infants and their parents [37].
We found a higher proportion of Black, Hispanic, and Asian infants with NICU admission than white infants over time in this sample, mirroring other national research [21,[30][31][32][33][34][35]. Although Black, Asian, and Hispanic infants had higher rates of NICU admission between 2010 and 2018, white postpartum women had higher rates of MH diagnoses of anxiety and/or depression compared to Black, Asian, or Hispanic women. Further, NICU admission increased the risk for these postpartum MH diagnoses, but prevalence of NICU admission by race/ethnicity did not correspond with diagnosis of postpartum MH conditions by race/ethnicity. Black, Hispanic, and Asian women had a higher proportion of infants in the NICU, yet a higher proportion of white postpartum women received maternal MH diagnoses, in accordance with similar research [30][31][32][33][34][35]. This counterintuitive finding may reflect the reality of experienced MH conditions and that white women do indeed experience worse mental health after childbirth. However, the surrounding body of research in this area supports the interpretation that this may describe evidence of a racial disparity in MH diagnoses for women with infants in the NICU [30][31][32][33][34][35]. Thus, another way to interpret our findings is that Black, Hispanic, and Asian women had lower odds of receiving a postpartum MH diagnosis compared to white women, and that may not describe the actual prevalence of these conditions among these racial/ethnic populations.
Although much of the NICU admission rightly focuses on the health of the infant, our study calls for attention to the well-being of postpartum women. Prior literature provides detailed guidance on the best practices for detecting MH conditions among all delivering women with infants hospitalized in the NICU, suggesting that care for parents' emotional well-being in the NICU represents an important component to the care of their hospitalized infants [24]. Considering care of the dyad, including both clinical and psychosocial needs, remains integral to family health and well-being; multidisciplinary care in the NICU environment can address these care gaps [24]. The 4th Trimester Project (North Carolina) [38] and Firefly (Tennessee) [39] promote patient-centered care among underresourced women during the postpartum period. These state efforts are improving the health and well-being for postpartum women, their infants, and families. Implementation of comprehensive interventions at the state and community level could tailor to the unique needs of under-resourced communities nationally to improve quality of care in the NICU and into the first years of life for infants and families.
This study has multiple strengths, including documentation of trends in NICU admission and postpartum MH diagnoses of anxiety and/or depression over time using a large, national sample and observation of dyadic outcomes. Along with these strengths, this work also has limitations. First, we used a privately-insured sample. Thus, these data do not reflect dyadic outcomes among nearly half the births in the US, which have public insurance coverage or remain uninsured, in the postpartum period. This sample selection represents one possible explanation for the racially and ethnically disparate findings of this study. Second, this study used ICD codes. Therefore, these data could not reliably indicate screening rates for maternal MH conditions and elevated symptoms for MH problems, a limitation which could have implications for potential racial/ethnic disparities. Since there are studies that document racial/ethnic disparities in screening rates and inadequate diagnoses/treatment of mental health conditions for racial/ethnic minorities compared to white people, our study may have not captured the accurate association between race/ethnicity and postpartum mental health. Third, this study could not address the mechanisms driving the differential rates of MH diagnosis across racial and ethnic groups of women with an infant in the NICU.
Some evidence suggests that legislating screening for maternal MH conditions may reduce inequities associated with screening and improving screening rates [32]. Yet, few postpartum women receive screening, diagnosis, follow-up, and adequate treatment, in part, because of uncoordinated systems of care across clinical specialties [40], with glaring disparities documented by race/ethnicity [33][34][35]. Addressing disparities in comprehensive MH care among postpartum women requires coordinated care across specialty inpatient settings, such as NICU, and extending into ambulatory care settings [24]. Improving care for dyads in the NICU and beyond also requires acknowledging the role that systemic factors, including social determinants of health, play in disparate MH care seeking, delivery, and outcomes [41].
In conclusion, this study found that NICU hospitalization increases the odds of maternal MH diagnoses of anxiety and/or depression in the year following this event, with longer length of infant hospitalization contributing to significantly higher rates of these maternal MH diagnoses. Although Black, Asian, and Hispanic infants had higher rates of NICU hospitalization over time, white postpartum women had higher rates of MH diagnoses. The interdisciplinary nature of the NICU provides an excellent opportunity to further investigate the social determinants of health and promote maternal-infant health.