Fragmented Medicaid coverage and risk of adverse psychiatric outcomes within 9 months after delivery

OBJECTIVE: Mental health conditions (MHCs) and substance use disorders (SUDs) are among the leading causes of maternal mortality in the late postpartum period.1,2 A recent Centers for Disease Control and Prevention study found that 23% of maternal deaths were associated with MHCs and SUDs.1 Patients with Medicaid insurance may be at increased risk of these diagnoses. Historically, many Medicaid beneficiaries lost coverage 60 days after delivery when pregnancy-related eligibility ended. From 2015 to 2018, >40% of women covered by Medicaid became uninsured or changed insurance after delivery.3 Insurance "fragmentation" may be an additional risk factor for adverse outcomes. We evaluated the effect of "Medicaid fragmentation" on the risk of suicide attempts and drug overdose among readmissions within 9 months after delivery. STUDY DESIGN: We conducted a retrospective cohort study of women aged 15 to 54 years in the 2010-2019 Nationwide Readmissions Database (NRD), a publicly available all-payer database that captures 60.4% of hospitalizations nationally.4 The NRD collects data on a state level capable of tracking patients across hospital admissions within a state within a given year. We identified delivery hospitalizations and ascertained first rehospitalization within 270 days from delivery discharge associated with a composite of suicide attempt, drug overdose, and/or a primary MHC or SUD diagnosis. We included only deliveries with discharge through March 31 to allow a full 270-day follow-up. Using adjusted logistic regression models, we determined the likelihood of 270-day readmission with the composite with unadjusted odds ratios (ORs) and adjusted ORs (aORs) as measures of effects. The adjusted models included demographics (age, median ZIP code income quartile based on residence, and payer), hospital factors (location and teaching status), and clinical factors potentially associated with readmission outcomes (chronic hypertension, pregestational diabetes mellitus, and psychiatric and SUD diagnoses). We analyzed all deliveries and performed analyses restricted to patients with Medicaid at delivery with first readmission between 60 and 270 days (2-9 months) after delivery, evaluating "Medicaid fragmentation" in a case-control fashion as the primary exposure of interest for (1) suicide attempt and (2) nonfatal overdose. "Medicaid fragmentation" was defined as either no payer or a non-Medicaid payer at rehospitalization. Readmissions from 60 days onward were specifically analyzed to capture elapsed Medicaid coverage. All analyses were performed using SAS (version 9.4; SAS Institute, Cary, NC).RESULTS: Of 8.9 million deliveries, we identified 29,799 readmissions within 9 months after delivery with the composite diagnosis (33.3 per 10,000 delivery hospitalizations), including 3688 readmissions with suicide attempts (4.1 per 10,000 delivery hospitalizations) and 8261 readmissions with drug overdose (9.2 per 10,000 delivery hospitalizations) (Supplemental Figure). Medicaid insurance was associated with increased odds of readmission (aOR, 2.3; 95% confidence interval [CI], 2.2-2.5), as were SUD (aOR, 4.0; 95% CI, 3.7-4.2), and psychiatric diagnoses (aOR, 5.3; 95% CI, 5.0-5.6) (Table). More than 66% of women who were readmitted had Medicaid at delivery. Of women with Medicaid at delivery and a 2-to 9 month postpartum readmission, 25% of those women experienced an interruption to Medicaid coverage. In this group, fragmented Medicaid insurance was associated with higher odds of suicide attempts (aOR, 1.25; 95% CI, 1.03-1.52). In adjusted case-control models, psychiatric diagnoses were associated with lower odds of drug overdose, and SUD was associated with lower odds of suicide attempts. CONCLUSION: Most women who will be readmitted for a mental health indication after delivery have Medicaid at delivery. Postpartum insurance instability is common in the Medicaid population. Insurance fragmentation may disrupt access to behavioral health screening and treatment, and it is associated with an increased risk of suicide. Of note, 1 limitation of our study is that we were unable to assess risk by race or ethnicity, as the NRD does not contain this information; other works show that among Medicaid-paid births, coverage instability disproportionally affects Hispanic women.3 Another limitation is that we were unable to determine whether patients were in treatment of MHC or SUD diagnoses before readmission given that this database only includes inpatient discharge data. A third limitation is that modeling the exposure of "Medicaid fragmentation" required a case-control design among those readmitted with a composite readmission diagnosis. Patients with delivery MHC and SUD diagnoses may have had lower odds for outcomes, such as overdose and suicide attempt, respectively, because they were readmitted specifically for MHC and SUD diagnoses. Thus, these lower odds likely represent an artifact of the case-control structure and do not indicate that patients with MHC and SUD diagnoses are at lower population-level risk of these outcomes. Our analysis underscores the importance of consistent healthcare coverage after delivery and risks associated with discontinuity of coverage and lends support to ongoing efforts to extend Medicaid coverage within 12 months after childbirth.5