In this retrospective cohort study of deliveries in Maryland between 2016–2018, patients who experienced SMM during their delivery hospitalization had 3.7, 2.8 and 3 times the odds of postpartum hospitalization with a mental health condition diagnosis, substance use disorder diagnosis and both, respectively, compared to patients who did not experience SMM during delivery. Rates of postpartum hospitalization with such diagnoses were significantly higher among patients with SMM compared to those without SMM during the early and late postpartum periods. Our findings confirm prior research documenting higher rates of postpartum hospitalization among postpartum patients who experience SMM compared to those who did not experience SMM (23), and are supported by clinical plausibility with respect to the associations between SMM and presence of mental health condition and substance use disorder diagnoses. Research has shown that traumatic birth outcomes increase the risk of mental health conditions, particularly posttraumatic stress disorder, during the postpartum period (24–26). Moreover, a study in Sweden found a positive association between SMM at delivery and postpartum treatment for psychiatric disorders (27). Experience of adverse or unexpected outcomes such as preeclampsia, preterm birth, or cesarean delivery have been identified as risk factors for postpartum mental health conditions (28–30). Similarly, the postpartum period is recognized as a particularly vulnerable time for individuals with substance use disorders at risk of relapse, which may be further compounded by the experience of birth trauma or adverse pregnancy outcomes (31). Fear for personal safety or safety of the neonate and negative perceptions of birth, multiple interventions during labor and birth, and anesthesia complications are all associated with postpartum mental health conditions and can be unique vulnerabilities for those with histories of substance use disorders (31, 32). SMM frequently includes multiple interventions as well as anesthesia and is often characterized as a traumatic event by those who experience them, which can explain the relationships we observed in our study.
Efforts to reduce preventable SMM and its effects should include recognition and management of mental health and substance use in pregnancy and postpartum, and particularly the late postpartum period. In our study, postpartum hospitalization rates for patients with mental health and substance use disorders were higher during the late postpartum period compared to the early postpartum period. Similarly, studies of pregnancy-associated mortality have identified higher rates of deaths due to overdose and suicide in the late postpartum period (33, 34). Together, these suggest the late postpartum period is a particularly vulnerable time for these behavioral health conditions.
Patients with SMM often receive inadequate information about their morbidity (35). Recommendations following an SMM event include offering patients a debriefing by their clinician before hospital discharge, social support, referrals and warm hand-off to mental health and substance use providers and services, including specialized treatment (35). Such interventions have been found to be moderately effective in reducing mental health symptoms among patients with traumatic births (36, 37). The US Preventive Services Task Force recommends that clinicians provide or refer postpartum persons at increased risk for perinatal depression to counseling interventions due to sufficient evidence of their effectiveness (38). Increased risk includes those with a history of depression, current depressive symptoms, socioeconomic risk factors, or a history of significant negative life events. Findings from this analysis suggest that experience of SMM should also be considered as a risk factor.
Our findings demonstrate no significant difference in the in adjusted odds of postpartum hospitalization with a mental health condition between non-Hispanic Black and White patients who experience SMM, but lower odds of postpartum hospitalization for substance use disorders and both mental health conditions and substance use disorders. Additionally,, odds of hospitalization with a mental health condition or substance use disorder were significantly lower among Hispanic patients and other racial/ethnic minority groups compared to white patients. More research is needed to determine whether these differences are due to true lower rates of mental health conditions and substance use disorders or differences in rates of diagnosis and recognition.
Our study has limitations. While there are indications that SMM precipitates postpartum mental health conditions and substance use, these relationships may be bidirectional, such that a history of mental health conditions or substance use predisposes pregnant individuals to SMM, which in turn exacerbates the risk of the same in the postpartum period (16). Clearly establishing the directionality and causality of this relationship will require further, more robust longitudinal studies. In addition, examining the relationship of SMM with specific types of mental health conditions as well as co-occurring mental health conditions and substance use should be the focus of research when a larger sample size than ours is available. We used the first five diagnosis codes for each patient to ascertain our outcomes of interest – this cut-off point was used to examine postpartum hospitalizations where mental health conditions and substance use disorders represented important reason(s) for the hospitalizations. However, we may have missed patients who were admitted to the hospital for these conditions but, for which, only direct pregnancy complications were noted with the first five conditions codes. We may have also missed patients who do not disclose their mental health condition or symptoms, or their substance use to a healthcare provider, or those who are less likely to be screened for such conditions including birthing people of color and those with low incomes (39). The data are limited to conditions identified during hospitalizations, whereas many of these conditions do not result in hospitalization and may be identified in outpatient settings and treated through medication, therefore we are unable to adjust for prenatal depression and substance use, unless identified during the delivery hospitalization; moreover, only SMM during delivery hospitalization is examined, thus excluding SMM events that occur in antepartum and postpartum hospitalizations. In addition, residual confounding may be present due to the exclusion of other chronic health conditions and the limited sociodemographic variables available through hospital records. Furthermore, reliance on administrative data, which are primarily collected for billing purposes, has other limitations because all health conditions present during a hospital admission may not be consistently and accurately reported and the algorithm used to identify SMM based on this data is not well suited to identify all cases of SMM events involving hemorrhage (22).
However, the study has some important strengths. It uses a large, statewide database analyzed longitudinally over several years before the start of the COVID-19 pandemic – this was only possible given the inclusion of a unique patient identifier in the Maryland SID. The use of ICD-10 codes for outcome ascertainment has the strength of consistency in identifying cases over the study period and offers some reassurance that they represent true cases of patients with mental health conditions or substance use disorders.