Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2020 Sep 17.
Published in final edited form as: Womens Health Issues. 2017 Apr 17;27(4):449–455. doi: 10.1016/j.whi.2017.02.006

Do Maternity Care Coordination Services Encourage Use of Behavioral Health Treatment among Pregnant Women on Medicaid?

Megan Shepherd-Banigan a,*, Marisa E Domino b,c, Rebecca Wells d, Regina Rutledge b, Marianne M Hillemeier e, Courtney H Van Houtven a,f
PMCID: PMC7497466  NIHMSID: NIHMS1560610  PMID: 28427755

Abstract

Objective:

Maternity care coordination (MCC) may provide an opportunity to enhance access to behavioral health treatment services. However, this relationship has not been examined extensively in the empirical literature. This study examines the effect of MCC on use of behavioral health services among perinatal women.

Methods:

Medicaid claims data from October 2008 to September 2010 were analyzed using linear fixed effects models to investigate the effects of receipt of MCC services on mental health and substance use–related service use among Medicaid-eligible pregnant and postpartum women in North Carolina (n = 7,406).

Results:

Receipt of MCC is associated with a 20% relative increase in the contemporaneous use of any mental health treatment (within-person change in probability of any mental health visit 0.5% [95% CI, 0.1%–1.0%], or an increase from 8.3% to 8.8%); MCC in the prior month is associated with a 34% relative increase in the number of mental health visits among women who receive MCC (within-person change in the number of visits received 1.7% [95 CI, 0.2%–3.3%], or from 0.44 to 0.46 mental health visits). No relationship was observed between MCC and Medicaid-funded substance use–related treatment services.

Conclusions:

MCC may be an effective way to quickly address perinatal mental health needs and engage low-income women in mental health care. However, currently there may be a lost opportunity within MCC to increase access to substance use–related treatment. Future studies should examine how MCC improves access to mental health care such that the program’s ability can be strengthened to identify women with substance use–related disorders and transition them into available care.

Depression, substance use, and other perinatal mental health disorders are prevalent during pregnancy and the postpartum period (O’Hara & Wisner, 2014). Given that these conditions are associated with negative outcomes for women and babies, it is imperative to identify solutions. Maternal care coordination (MCC) is one potential strategy to increase access to behavioral health treatment. MCC is a program run by local health departments that connects Medicaid-eligible women to services and education. We illustrate the prevalence of behavioral health problems and treatment among perinatal women, examine the negative effects of these problems on infants, and describe how MCC is a potential, but understudied, approach to increasing the use of behavioral health treatment among low-income women with infants.

The most widely recognized mental health disorder is perinatal depression, but other disorders, including substance use disorders, contribute to substantial maternal morbidity and are often comorbid with depression (Chapman & Wu, 2013; Howard et al., 2014; O’Hara & Wisner, 2014). Perinatal depression is a condition that occurs in all cultural groups, affecting approximately 15% to 20% of women within 3 months after birth (Gaynes et al., 2005), and persisting for up to 12 months after delivery (Sit & Wisner, 2009). The proportion of low-income and minority women diagnosed with any depression is greater than for women in the general population (Gress-Smith, Luecken, Lemery-Chalfant, & Howe, 2012; Kozhimannil, Trinacty, Busch, Huskamp, & Adams, 2011; Vesga-Lopez et al., 2008). The prevalence of substance use disorders among pregnant and postpartum women is also high (Vesga-Lopez et al., 2008). Data from the early 2000s show that among U.S. women with an infant under 1 year of age, 15% reported binge alcohol use and 8.5% reported illicit or nonmedical drug use in the past month (Chapman & Wu, 2013). Another study using nationally representative data collected between 2001 and 2002 found that 3.6% and 2.0% of U.S. women who had been pregnant in the past year reported an alcohol use disorder or a substance use disorder, respectively (Vesga-Lopez et al., 2008). These results likely understate prevalence because of social desirability response bias. Further, these conditions are often comorbid; women with a history of substance use are at higher risk for experiencing depressive symptoms after birth (Connelly, Hazen, Baker-Ericzen, Landsverk, & Horwitz, 2013) and postpartum depression is correlated with substance dependence in the postpartum period (Chapman & Wu, 2013).

Unfortunately, for many women, particularly those of low socioeconomic status (Ammerman et al., 2009; Kozhimannil et al., 2011), perinatal mental health and substance use disorders remain undetected and untreated (Geier, Hills, Gonzales, Tum, & Finley, 2015; Gopman, 2014; Vesga-Lopez et al., 2008). One study found that 60% of postpartum women reported experiencing depressive symptoms, yet only 20% of the full sample sought related care (Declercq, Sakala, Corry, & Applebaum, 2007). This is not surprising; in the United States, fewer than 50% of women are screened for postpartum depression (Gjerdingen & Yawn, 2007). Further, women with lower levels of education (Centers for Disease Control and Prevention, 2007), who are less engaged with the health care system (Centers for Disease Control and Prevention, 2007), and who have mental illnesses (Minkovitz et al., 2005) are less likely to attend postnatal care visits, resulting in a missed opportunity to identify and treat these conditions. Low levels of screening and diagnosis are problematic, because mental illness can recur and may develop into a chronic condition (Vliegen, Casalin, & Luyten, 2014).

In addition to the negative health effects for women, a large body of evidence demonstrates that maternal mental health and substance use disorders can have a negative impact on early mother–infant bonding (O’Higgins, Roberts, Glover, & Taylor, 2013), attachment (Lefkovics, Baji, & Rigo, 2014), and caregiving activities, such as breastfeeding, sleep routines, and child health service use (Field, 2010). These disorders are also associated with greater risk of child neglect and abuse (Rutherford, Williams, Moy, Mayes, & Johns, 2011). In addition, maternal behavioral health conditions are associated with longer term consequences for child mental health outcomes, including emotional–behavioral difficulties (Betts, Williams, Najman, & Alati, 2015). However, evidenced-based psychiatric and substance use treatment for mothers can decrease symptoms and unhealthy behaviors and improve parental functioning (Cuijpers, Weitz, Karyotaki, Garber, & Andersson, 2015; Gopman, 2014). Given the importance of addressing mental health and substance use disorders during the perinatal period for women and children, connecting pregnant women with screening services and essential care is an imperative and achievable policy goal.

Care coordination is a key aspect of the medical home framework that maintains patient engagement across the continuum of care and may be an important health system resource for children and families (Johnson & Rosenthal, 2009). North Carolina health departments have provided coordinated maternity care since 1987, although only in 2011 did the state establish a state-wide, comprehensive pregnancy medical home for Medicaid beneficiaries (Wells, Cilenti, & Issel, 2015). As part of the pregnancy home model in North Carolina, MCC is available to Medicaid-eligible pregnant women from pregnancy through the last day of the month in which the 60th postdelivery day occurs. The North Carolina MCC program staff, including nurses, social workers, and paraprofessionals, provides formal case management services to eligible women to promote positive birth outcomes. Services include connecting women to appropriate resources, including prenatal and postnatal care, childbirth education, and referrals to community services such as housing, transportation, and mental health counseling (North Carolina Division of Medical Assistance, 2010). This set of services is different from standard perinatal/postpartum care in that women are actively identified early during pregnancy and are referred to a broader range of social and health services compared to referrals they might receive as part of standard perinatal and postpartum care.

MCC services target low-income pregnant women and aim to connect women with needed services in the health care system (Ammerman et al., 2009). Given this, we hypothesize that women with behavioral health issues will be identified by the case management team and supported to access and remain engaged in behavioral health treatment. However, the degree to which MCC workers recognize mental health needs and refer clients to appropriate services is not well-established. Existing research about pregnancy medical home models suggests that care coordination improves prenatal care service use and birth outcomes (Hillemeier et al., 2015; Nason, Alexander, Pass, & Bolland, 2003), but the effect of care coordination on maternal service use for perinatal depression and substance use treatment services is less clear (Shaw, Levitt, Wong, & Kaczorowski, 2006). This study investigates the effects of receipt of MCC on Medicaid-funded behavioral health service use among Medicaid-eligible pregnant and postpartum women in North Carolina (North Carolina Division of Medical Assistance, 2010). Specifically, we examine whether receipt of any MCC is related to use of mental health and substance use treatment services during pregnancy and in the first 3 months after delivery.

Materials and Methods

Data and Sample

This study uses Medicaid claims data for a random sample of 8,000 women whose births occurred between October 1, 2008, and September 30, 2010, and were funded by the North Carolina Medicaid program. This random sample comprises approximately 11% of the annual number of Medicaid births in North Carolina. The sampling rate was determined by computing constraints and was not selected a priori. Vital statistics records were matched to infant and maternal Medicaid claims. The North Carolina State Center for Health Statistics matched records using a combination of first, middle, and last name, date of birth, county of residence, and hospital of birth; match rates exceeded 90% in each year (Hillemeier et al., 2015).

The analytical sample for the present study included 7,406 births (69,266 person-months); women were excluded from the analysis sample if births could not be matched with maternal Medicaid claims and eligibility files or mothers did not have at least 2 months of continuous Medicaid coverage during pregnancy. Medicaid coverage included women with full Medicaid as well as those who received more limited coverage through the Medicaid pregnancy waiver, which provides health insurance for women from pregnancy until 3 months after delivery (North Carolina Division of Medical Assistance, 2016). The treatment group comprises all women with Medicaid coverage in the sample who received MCC services at some point during pregnancy and 3 months after delivery; the control group includes all women with Medicaid coverage who did not receive MCC services during the same time period.

This study was reviewed and approved by the Institutional Review Boards at the University of North Carolina Chapel Hill and at the Pennsylvania State University.

Variables

Dependent variables

Any Medicaid-financed a) mental health or b) substance use–related treatment visit in the month, and number of Medicaid-financed c) mental health or d) substance use–related treatment visits made by the mother in each month. Data were retrieved from Medicaid claims; mental health or substance use–related visits were identified based on criteria, including 1) a primary care code plus a mental health or substance use–related diagnosis, 2) any outpatient visit with a mental health or addictions specialty provider, or 3) any specialty mental health or substance use–related outpatient procedure code. Substance use–related diagnostic codes include alcohol and drug-related abuse or dependence, substance-related psychosis, or self-inflicted harm involving substances and do not include tobacco use; services funded outside of the Medicaid system were not captured in our data.

Key explanatory predictors

Predictors include receipt of MCC services in the a) current month and b) past month. The initial MCC session occurred via a 15-minute in-person encounter during pregnancy, but subsequent MCC services may have occurred in person or via telephone (Wells et al., 2015). MCC services were identified by procedure codes (e.g., T1017) combined with provider type and specialty codes (Hillemeier et al., 2015).

Covariates

Although behavioral health services are likely only beneficial in a diagnosed population, diagnoses from administrative data such as Medicaid claims under-report population prevalence because diagnoses are only present when a provider has recognized a condition and reported it to a payer on an insurance claim. Therefore, analyses that focus only on a population with administrative behavioral health diagnoses excludes many individuals with a target condition for treatment. We, therefore, included the full sample of pregnant and postpartum women as described, but control for the presence of a diagnosed behavioral health condition. Although adjusting for a diagnosed behavioral health condition might decrease the observed treatment effect, these variables are important within-individual confounders and by including them we present more conservative treatment effect estimates instead of partially controlling for observed confounding. As will be explained in the next section, we estimated within-person effects and thus did not control for time-invariant characteristics, such as age or race/ethnicity. To avoid confounding with treatment, we used a 1-month lagged indicator of a mental health or substance use–related diagnosis in the claims files. To account for time trends, we included pregnancy-month indicator variables for the timing of each month of Medicaid coverage from month 1 of pregnancy to 3 months after delivery (months 1–14). Panels were unbalanced because women were enrolled in Medicaid for varying amounts of time; women were part of this analytical cohort for at least 2 months and at most 13 months. The average period of time that women remained in our analytical cohort was 9.4 months.

Statistical Methods

Person-month respondent characteristics were described using standard descriptive statistics.

Women self-select into MCC services both on observable and unobservable factors; as a result, we expect substantial differences between treatment groups. To address reasons for self-selection into both MCC and behavioral health services, we applied econometric statistical techniques. Specifically, fixed effects models using ordinary least squares (OLS) estimation (Cameron & Trivedi, 2010) were estimated to test the within-individual relationship over time between receipt of MCC services and subsequent use of mental health and substance use–related treatment services during pregnancy and up to 3 months postpartum. Mental health and substance use–related treatment service variables were modeled as separate outcome variables.

The fixed effect parameters represent the average change in the number of behavioral health visits within each individual per month owing to receipt of MCC services in the current and the past month. By testing changes within each individual over time, longitudinal fixed effects models control for time-invariant (unchanging) differences in observed and unobserved characteristics between MCC recipients and controls, including mechanisms of self-selection into MCC and into behavioral health services. This approach, therefore, controls for potential biases that would otherwise arise owing to self-selection on unobserved time-invariant characteristics. All models used complete case analysis and robust standard errors to account for potential heteroscedasticity.

Analyses were run in STATA IC 13. We performed several sensitivity analyses to test various model specifications and the robustness of our primary analysis, including 1) fixed effects count models with Poisson and negative binomial distributions, 2) two-part models to account for the high number of nonusers, 3) OLS fixed effects with trimester versus month indicator variables, and 4) an inverse probability of treatment weighted (IPTW) analysis of the use of MCC at any time on any mental health and substance use–related treatment and the number of treatment services received during the timeframe. Propensity scores were derived from all potential confounders available in the dataset, including mother demographic characteristics, mother health variables, previous birth outcomes, Medicaid eligibility category, health insurance, and local health department variables. Root mean squared errors (RMSE) were used to compare the predictive ability of each of the first three models. The OLS fixed effects models had the lowest RMSEs, suggesting that the linear models fit the data better, despite the lack of normality of our dependent variables. Models with trimester and monthly indicators had the same RMSE. When we included trimester, rather than monthly indicators, MCC no longer had an effect on mental health service use or substance use–related visits; this finding suggests that these models may be sensitive to specification. We present the results from the fixed effects OLS model because the IPTW model did not allow us to investigate the timing of MCC effects across pregnancy/postpartum. We also present the results from the fixed effects OLS models with the month indicator variables because it was important to account for incremental variation by month throughout pregnancy; however, we interpret the results with caution.

Results

Table 1 displays the baseline covariates for women in this sample. Note that presented statistics are month-level estimates of claims-derived diagnoses and do not reflect the actual prevalence of behavioral health diagnoses in the population. On average, 3.67% of women had a mental health diagnosis in the past month and 1.55% had a substance use–related disorder diagnosis in the past month. Person-month receipt of behavioral health treatment was a rare outcome; on average, fewer than 3% of women in each month had a mental health treatment visit and fewer than 1% of women in each month had a substance use–related treatment visit. In each month, 17.66% of women in the sample received MCC services; MCC was accessed more frequently by women who were on average younger, African American, received Temporary Assistance for Needy Families and Supplemental Security Income, and had a mental health or substance use–related diagnosis.

Table 1.

Person-Month Mother Health and Demographic Characteristics*

Mother Characteristics
(%)		 Unit MCC
Recipients
(n = 12,235)		 Non-MCC
Recipients
(n = 57,031)		 Full Sample
(n = 69,266)†
Health-related characteristics
 Any MCC use in month % 100 0 17.66
 Behavioral health diagnoses
  Mental health diagnosis % 4.33 3.53 3.67
  Substance dependence diagnosis % 1.79 1.5 1.55
 Any use of behavioral health services in month
  Mental health‡ % 3.28 2.51 2.64
  Substance dependence % 0.87 0.71 0.73
 Count of behavioral visits in month
  Mental health M (SD) 0.06 (0.43) 0.05 (0.41) 0.05 (0.41)
  Substance dependence M (SD) 0.03 (0.63) 0.02 (0.43) 0.02 (0.47)
Demographics
 Average age (mo) M (SD) 23.27 (5.25) 24.67 (5.58) 24.4 (5.55)
 Mother race
  Asian % 0.85 1.36 1.27
  Black % 45.96 37.05 38.63
  Native American/Pacific Islander % 3.79 1.72 2.09
  White % 44.79 50.63 49.60
  Unknown % 3.82 6.30 5.86
  Missing % 2.94 0.79 2.56
 Ethnicity
  Mexican % 0.99 1.56 1.46
  Other-Hispanic % 3.24 4.90 4.61
  Non-Hispanic % 88.11 80.40 81.76
  Unknown % 6.87 10.20 9.61
  Missing % 0.79 2.94 2.56
  TANF ever % 5.98 5.49 5.58
  SII ever % 2.73 2.01 2.14
Open in a new tab

Abbreviations: MCC, maternity care coordination; SII, Supplemental Security Income; TANF, Temporary Assistance for Needy Families.

*

Reported statistics corresponds to estimation sample (n = 7,406; person-months = 69,266).

†

Recipient-months refer to months where MCC services are received. Non-recipient months include months for those who never received MCC services as well as months for previous or future MCC recipients with no MCC services in the current month.

‡

Refers to a diagnosis of at least 1 out of 5 of the following mental health conditions: depression, anxiety, bipolar, schizophrenia, trauma.

Table 2 displays the coefficients from the linear fixed effects models. When controlling for month of pregnancy and behavioral health diagnosis, receipt of MCC in the current month was associated with a 0.5% point increase in the within-person probability of any use of mental health treatment (0.005; 95% CI, 0.001–0.010; or an increase from 8.3% to 8.8% of women each month); this translates into a 20% relative increase in the within-person use of services.1 Receipt of MCC in the past month was associated with an increase of less than two-tenths of a visit in the number of monthly mental health treatment visits (0.017 visits; 95% CI, 0.002–0.033; or from 0.44 to 0.46 mental health visits). Although small in magnitude, this translates into an average 34% relative increase for individuals.2 We did not observe an association between MCC use and any or number of substance use–related treatment services. We did not observe an association between MCC received in the prior month and any mental health treatment service use or between MCC received in the same month and the number of mental health treatment visits. When controlling for MCC services and month of pregnancy, individuals with a mental health or substance use–related diagnoses were more likely to use treatment services. When controlling for MCC services and behavioral health diagnoses, month of pregnancy was not associated with use of behavioral health treatment services.

Table 2.

Change in Monthly Within-Individual Use of Behavioral Health Treatment Services Attributed to the Use of MCC From Linear Fixed Effects Models

Independent Variables Any Mental Health Treatment Service Use in Month Any Substance Use Treatment Service Use in Month No. of Mental Health
Treatment Visits in Month		 No. of Substance Use
Treatment Visits in Month
Change in Probability of Receipt of Any Visit (95% CI) Change in Number of Visits Received (95% CI)
MCC received in same month 0.005* (0.001–0.010) 0.002 (−0.001 to 0.004) 0.004 (−0.008 to 0.017) 0.008 (−0.006 to 0.022)
MCC received in prior 1 month 0.004 (−0.001 to 0.008) −0.0001 (−0.002 to 0.002) 0.017* (0.002–0.033) 0.006 (−0.013 to 0.025)
Mental health-related disorder (in past month)† 0.094* (0.071–0.117) 0.011* (0.002–0.019) 0.270* (0.190–0.350) 0.034 (−0.013 to 0.081)
Substance use-related disorder (in past month) 0.061* (0.032–0.089) 0.101* (0.070–0.132) 0.263* (0.110–0.417) 0.595* (0.333–0.857)
Time
 Pregnancy month 1 −0.003 (−0.011 to 0.005) 0.0002 (−0.003 to 0.004) −0.005 (−0.021 to 0.011) 0.004 (−0.07 to 0.0147)
 Pregnancy month 2 −0.004 (−0.012 to 0.004) 0.0005 (−0.003 to 0.004) −0.007 (−0.027 to 0.013) −0.001 (−0.013 to 0.010)
 Pregnancy month 3 −0.003 (−0.011 to 0.005) 0.002 (−0.002 to 0.006) −0.005 (−0.027 to 0.017) 0.001 (−0.011 to 0.012)
 Pregnancy month 4 −0.001 (−0.009 to 0.007) 0.002 (−0.002 to 0.005) −0.008 (−0.032 to 0.016) 0.0001 (−0.012 to 0.012)
 Pregnancy month 5 −0.003 (−0.011 to 0.005) 0.002 (−0.001 to 0.006) −0.009 (−0.033 to 0.016) 0.002 (−0.012 to 0.016)
 Pregnancy month 6 −0.001 (−0.009 to 0.007) 0.002 (−0.002 to 0.006) −0.005 (−0.030 to 0.020) 0.003 (−0.016 to 0.021)
 Pregnancy month 7 −0.0001 (−0.008 to 0.008) 0.004* (0.00–0.008) −0.002 (−0.027 to 0.023) 0.002 (−0.016 to 0.019)
 Pregnancy month 8 −0.001 (−0.009 to 0.007) 0.006* (0.002–0.010) −0.003 (−0.029 to 0.023) 0.006 (−0.011 to 0.023)
 Pregnancy month 9 −0.001 (−0.010 to 0.007) −0.001 (−0.005 to 0.003) −0.004 (−0.032 to 0.025) −0.006 (−0.022 to 0.010)
 Postpartum month 1 0.007 (−0.002 to 0.015) 0.001 (−0.002 to 0.005) 0.011 (−0.018 to 0.040) 0.011 (−0.002 to 0.024)
 Postpartum month 2 0.007 (−0.002 to 0.015) 0.002 (−0.001 to 0.006) 0.011 (−0.016 to 0.038) 0.023* (0.007–0.039)
 Postpartum month 3 0.012* (0.003–0.021) 0.002 (−0.001 to 0.006) 0.018 (−0.010 to 0.047) 0.019* (0.002–0.035)
Open in a new tab

Abbreviation: MCC, maternity care coordination.

*

p ≤ .05.

†

Refers to a diagnosis of at least 1 out of 5 of the following mental health conditions: depression, anxiety, bipolar, schizophrenia, trauma.

Results from the sensitivity analyses show that, when we included trimester, rather than monthly indicators, MCC no longer had a significant effect on mental health service use or substance dependence; this suggests that these models may be sensitive to specification. The results from the propensity score model show that MCC increases the probability of any mental health treatment during pregnancy/postpartum by 1.3% points (95% CI, 0.1–2.0) versus 0.5% points in the fixed effects analysis. The results from these models show no association between use of MCC services and probability of substance use–related treatment services or the number of mental health and substance use–related treatment visits.

Discussion

Mental health and substance use disorders represent a significant problem for women and their infants in the perinatal period (Gavin et al., 2005). Despite this fact, owing to numerous structural and social barriers (Murphy & Rosenbaum, 1999), these disorders may go undetected and untreated (Gopman, 2014; Vesga-Lopez et al., 2008). Although evidence is limited, some studies suggest that pregnancy may provide a window of opportunity for intervention as maternal concern for the fetus may motivate treatment seeking and adherence (Daley, Argeriou, & McCarty, 1998; O’Connor & Whaley, 2007; Stade et al., 2009). Therefore, MCC may present an opportunity to build on maternal motivation and, in conjunction with sufficient system supports, engage pregnant women in behavioral screening and treatment services. Extant research finds that MCC improves prenatal service use and birth outcomes (Hillemeier et al., 2015), but whether MCC might also be effective in connecting low-income women to behavioral health treatment had not been examined. MCC is a low-intensity intervention and maternity coordinators may not be trained to identify behavioral health disorders. In contrast, maternity coordinators are likely aware of the behavioral health needs of the population they serve and may be more knowledgeable about behavioral health services and resources available for women who struggle with these issues.

We find that receipt of MCC services is associated with an increase in mental health service utilization. This is important as, among MCC recipients in this population, 22% of women have been diagnosed with a mental health disorder from 12 months before through 6 months after birth (Hillemeier et al., 2015). A central goal of MCC services in pregnancy is to identify health-related issues and connect women with appropriate resources for treatment as rapidly as possible so as to maximize the chances of healthy birth outcomes. The study findings suggest that MCC workers who recognize mental health conditions may be effective in arranging for this treatment within 1 month of the MCC visit. Once such women are identified through MCC, they have an increased number of mental health visits in subsequent months as well, suggesting that they are more likely to receive ongoing mental health treatment.

It is interesting that significant effects of MCC are not seen for substance use treatment. There are a large number of mental health and substance use treatment service shortage areas in North Carolina (Thomas, Ellis, Konrad, & Morrissey, 2012; and elsewhere); additionally, services and providers that accept Medicaid may not be available (Thomas et al., 2012). Another barrier to treatment may be fear of punitive action or stigma, because substance use in pregnancy is known to be associated with adverse effects on fetal health and development. Specifically, state laws that mandate criminal penalties for women and their medical providers deter women from seeking health care (Roberts & Pies, 2011; Stone, 2015). Although there are currently no state laws to prosecute women who use drugs during pregnancy in North Carolina, a woman may lose parental rights if her infant tests positive for an illegal substance (Miranda, Dixon, & Reyes, 2015). However, Senate Bill 297, introduced during the 2015 and 2016 North Carolina Legislative Session, aims to make prenatal narcotic drug use a criminal offense (Prenatal Narcotic Drug Use/Criminal Offense, 2015). Based on findings from other states, if such a law were passed, it would likely further reduce treatment seeking among pregnant women with behavioral health conditions in North Carolina. Our findings portray a lost opportunity; among MCC recipients in North Carolina, 10% of women had a claim-based substance dependence diagnosis between 12 months before and 6 month after birth (Hillemeier et al., 2015). This gap highlights the need to identify why MCC services seem to be less effective for connecting these women with substance dependence treatment and what additional services and social policies regarding treatment (American Congress of Obstetrics and Gynecology, 2015) are necessary to meet the needs of women during pregnancy.

We acknowledge several limitations. First, although the fixed effect OLS models demonstrated the best fit for the number of behavioral health visits received, OLS is not an ideal distribution for count data. However, given the need to control for selection into the MCC sample, our preference was to use person-level fixed effects that complicate the calculation of nonlinear count data models. Second, the results were sensitive to model specification; for example, when trimester indicators were included instead of month indicators, the MCC and MCC lag coefficients were no longer significant. Therefore, the results must be interpreted with caution. Third, we chose to use fixed effects to limit bias owing to unobserved heterogeneity. However, fixed effects coefficients estimate changes within individuals across time; therefore, they are less efficient and provide conservative estimates, making it difficult to identify program effects, such as those possibly related to use of substance dependence treatment services. Fixed effects also do not control for unobserved time-varying confounders, such as possibly increased motivation for self-care associated with pregnancy, so there may be some residual bias associated with within-individual variation if omitted variables are correlated with MCC. Fourth, month-level behavioral health service use was fairly low; therefore, it is possible that we lacked the power to detect some associations. Fifth, services reimbursed outside the Medicaid program or lacking behavioral health diagnoses or provider codes would not be captured in this analysis. Finally, MCC services are used more often by Medicaid-eligible women of relatively lower socioeconomic status and higher health needs; it is therefore possible that MCC services might be less effective for increasing the use of mental health treatment services among Medicaid-eligible pregnant women with a relatively higher socioeconomic status.

This study also has several strengths. To begin, we used a large, state-wide dataset composed of a random sample of low-income, Medicaid-eligible pregnant women to address an issue that has not been examined in the empirical literature. Second, we applied analytical methods to account for between individual confounding and in doing so eliminated many sources of unobserved bias. Third, we conducted sensitivity analyses using IPTW models. As has been found in prior research (Beadles et al., 2015), results were similar for all outcomes except for count of mental health visits. However, we decided to use the longitudinal fixed effects models because these models control for time-invariant confounding and we were thus able to investigate the timing of MCC effects throughout pregnancy/postpartum. As propensity score methods do allow one to examine within-individual changes over time, we would not be able to examine the effect of MCC services at two times (last month and current month) on behavioral health service use.

Implications for Policy and/or Practice

The results from this study show that, for pregnant Medicaid-eligible women, one benefit of receipt of MCC service receipt may be an increase in the use of mental health services. This has important potential downstream effects for the future health of a woman and her family. North Carolina was one of the first states to establish a state-wide, comprehensive pregnancy medical home for Medicaid beneficiaries during the era of health care reform. Other states, such as Wisconsin, Michigan, Indiana, and Oregon, are following suit. This research provides additional evidence about the benefit of state-level policies that support low-income pregnant women and their children through locally implemented case management models.

Conclusions

We find that MCC may be an effective way to quickly address perinatal behavioral health needs and help low-income women to engage in mental health care. Extant research suggests that receipt of mental health treatment may have positive spillover effects for a woman and her infant (Cuijpers et al., 2015; Field, 2010; Gopman, 2014; Rutherford et al., 2011). Further, treatment for mood disorders may have positive effects on treatment for other behavioral health conditions, such as anxiety and trauma. Future research should aim to replicate these findings and extend this research to examine underlying mechanisms so that MCC services can be strengthened to improve the identification of women with behavioral health disorders and transition them into available care.

Acknowledgments

This research was funded by the Robert Wood Johnson Foundation Public Health Practice-Based Research Network and the Maternal and Child Health Bureau, grant R40 MC21519. Megan Shepherd-Banigan is supported by a VA OAA HSR&D PhD Fellowship TPP 21–027. Dr. Hillemeier also received support from the Pennsylvania State University Population Research Institute, NIH grant 2R24HD041025-11. The contents of this report are solely the responsibility of the authors and do not necessarily represent the official views of the Department of Veterans Affairs. We have no potential conflicts of interests.

Biography

Megan Shepherd-Banigan, PhD, MPH, received her PhD in Health Services Research from the University of Washington. She is a fellow at the Center for Health Services Research in Primary Care at Durham VA. She studies mental health service policy and delivery systems.

Marisa E. Domino, PhD, is a Professor of Health Policy and Management at the University of North Carolina-Chapel Hill. She received her PhD from Johns Hopkins University. Her research interests include the economics of mental health and care for low-income populations.

Rebecca Wells, PhD, MHSA, is a Professor of Management, Policy & Community Health at the University of Texas School of Public Health. She received her PhD from the University of Michigan. Her research interests include safety net systems and inter-organizational cooperation.

Regina Rutledge, PhD, MPH, is a Research Public Health Analyst at RTI International. She received her PhD in Health Policy and Management from the University of North Carolina. Her research interests include publicly funded insurance programs, family planning services, and rural health.

Marianne M. Hillemeier, PhD, MPH, MSN, is a Professor of Health Policy and Administration at The Pennsylvania State University. She received her PhD in Sociology from the University of Michigan. Her research interests include disparities in health care access for women and children.

Courtney H. Van Houtven, PhD, MSc, is an Associate Professor of the School of Medicine at Duke University. She received her PhD from the University of North Carolina-Chapel Hill. Her research interests encompass long-term care financing, informal care, and end-of-life care.

Footnotes

1

Average treatment effect (ATE): 0.005 increase in probability of service use associated with MCC service receipt/0.0264 individuals in sample who used any mental health services.

2

ATE: 0.017 increase in the number of mental health visits associated with MCC/0.05 visits per month.

References

  1. American Congress of Obstetrics and Gynecology. (2015). Committee opinion no. 633: Alcohol abuse and other substance use disorders: ethical issues in obstetric and gynecologic practice. Obstetrics and Gynecology, 125(6), 1529–1537. [DOI] [PubMed] [Google Scholar]
  2. Ammerman RT, Putnam FW, Altaye M, Chen L, Holleb LJ, Stevens J, … Van Ginkel JB (2009). Changes in depressive symptoms in first time mothers in home visitation. Child Abuse and Neglect, 33(3),127–138. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Beadles CA, Farley JF, Ellis AR, Lichstein JC, Morrissey JP, DuBard CA, & Domino ME (2015). Do medical homes increase medication adherence for persons with multiple chronic conditions? Medical Care, 53(2), 168–176. [DOI] [PubMed] [Google Scholar]
  4. Betts KS, Williams GM, Najman JM, & Alati R (2015). The relationship between maternal depressive, anxious, and stress symptoms during pregnancy and adult offspring behavioral and emotional problems. Depression and Anxiety, 32(2), 82–90. [DOI] [PubMed] [Google Scholar]
  5. Cameron AC, & Trivedi PK (2010). Microeconomics using STATA (Revised ed.). College Station, TX: STATA Press. [Google Scholar]
  6. Centers for Disease Control and Prevention. (2007). Postpartum care visits-11 states and New York City, 2004. MMWR: Morbidity and Mortality Weekly Report, 56(50), 1312–1316. [PubMed] [Google Scholar]
  7. Chapman SL, & Wu LT (2013). Postpartum substance use and depressive symptoms: a review. Women and Health, 53(5), 479–503. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Connelly CD, Hazen AL, Baker-Ericzen MJ, Landsverk J, & Horwitz SM (2013). Is screening for depression in the perinatal period enough? The co-occurrence of depression, substance abuse, and intimate partner violence in culturally diverse pregnant women. Journal of Womens Health (Larchmt), 22(10), 844–852. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Cuijpers P, Weitz E, Karyotaki E, Garber J, & Andersson G (2015). The effects of psychological treatment of maternal depression on children and parental functioning: a meta-analysis. European Child and Adolescent Psychiatry, 24(2), 237–245. [DOI] [PubMed] [Google Scholar]
  10. Daley M, Argeriou M, & McCarty D (1998). Substance abuse treatment for pregnant women: A window of opportunity? Addictive Behaviors, 23(2), 239–249. [DOI] [PubMed] [Google Scholar]
  11. Declercq E, Sakala C, Corry M, & Applebaum S (2007). Listening to Mothers II: Report of the Second National U.S. Survey of Women’s Childbearing Experiences. Journal of Perinatal Education, 16(4), 9–14. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Field T (2010). Postpartum depression effects on early interactions, parenting, and safety practices: A review. Infant Behavior & Development, 33(1), 1–6. [DOI] [PMC free article] [PubMed] [Google Scholar]
  13. Gavin NI, Gaynes BN, Lohr KN, Meltzer-Brody S, Gartlehner G, & Swinson T (2005). Perinatal depression: A systematic review of prevalence and incidence. Obstetrics and Gynecology, 106(5 Pt 1), 1071–1083. [DOI] [PubMed] [Google Scholar]
  14. Gaynes B, Gavin N, Meltzer-Brody S, Lohr K, Swinson T, Gartlehner G, … Miller W (2005). Perinatal depression: Prevalence, screening accuracy, and screening outcomes. Research Triangle Park, NC: Agency for Healthcare Research and Quality (US). [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Geier ML, Hills N, Gonzales M, Tum K, & Finley PR (2015). Detection and treatment rates for perinatal depression in a state Medicaid population. CNS Spectrums, 20(1), 11–19. [DOI] [PubMed] [Google Scholar]
  16. Gjerdingen DK, & Yawn BP (2007). Postpartum depression screening: Importance, methods, barriers, and recommendations for practice. Journal of the American Board of Family Medicine, 20(3), 280–288. [DOI] [PubMed] [Google Scholar]
  17. Gopman S (2014). Prenatal and postpartum care of women with substance use disorders. Obstetrics and Gynecology Clinics of North America, 41(2), 213–228. [DOI] [PubMed] [Google Scholar]
  18. Gress-Smith JL, Luecken LJ, Lemery-Chalfant K, & Howe R (2012). Postpartum depression prevalence and impact on infant health, weight, and sleep in low-income and ethnic minority women and infants. Maternal and Child Health Journal, 16(4), 887–893. [DOI] [PubMed] [Google Scholar]
  19. Hillemeier MM, Domino ME, Wells R, Goyal RK, Kum HC, Cilenti D, … Basu A (2015). Effects of maternity care coordination on pregnancy outcomes: Propensity-weighted analyses. Maternal and Child Health Journal, 19(1), 121–127. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Howard LM, Molyneaux E, Dennis CL, Rochat T, Stein A, & Milgrom J (2014). Non-psychotic mental disorders in the perinatal period. Lancet, 384(9956), 1775–1788. [DOI] [PubMed] [Google Scholar]
  21. Johnson K, & Rosenthal J (2009). Improving care coordination, case management, and linkages to service for young child: Opportunities for states. Available at: www.commonwealthfund.org/publications/fund-reports/2009/apr/improving-care-coordination-case-management-and-linkages. Accessed: March 20, 2017.
  22. Kozhimannil KB, Trinacty CM, Busch AB, Huskamp HA, & Adams AS (2011). Racial and ethnic disparities in postpartum depression care among low-income women. Psychiatric Services, 62(6), 619–625. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Lefkovics E, Baji I, & Rigo J (2014). Impact of maternal depression on pregnancies and on early attachment. Infant Mental Health Journal, 35(4), 354–365. [DOI] [PubMed] [Google Scholar]
  24. Minkovitz CS, Strobino D, Scharfstein D, Hou W, Miller T, Mistry KB, & Swartz K (2005). Maternal depressive symptoms and children’s receipt of health care in the first 3 years of life. Pediatrics, 115(2), 306–314. [DOI] [PubMed] [Google Scholar]
  25. Miranda L, Dixon V, & Reyes C (2015). How states handle drug use during pregnancy. Available at: https://projects.propublica.org/graphics/maternity-drug-policies-by-state. Accessed: February 1, 2017.
  26. Murphy S, & Rosenbaum M (1999). Pregnant women on drugs: Combating stereotypes and stigma. New Brunswick, NJ: Rutgers University Press. [Google Scholar]
  27. Nason CS, Alexander GR, Pass MA, & Bolland JM (2003). An evaluation of a Medicaid managed maternity program: The impact of comprehensive care coordination on utilization and pregnancy outcomes. Journal of Health and Human Services Administration, 26(2), 239–267. [PubMed] [Google Scholar]
  28. North Carolina (State). Legislature. Senate. Prenatal Narcotic Drug Use/Criminal Offense Senate Bill 297; 14 (Chapter) 14-34.11. 2015-2016 Reg. Sess (2015).
  29. North Carolina Division of Medical Assistance. (2010). Maternity care coordination. Available at: https://dma.ncdhhs.gov/medicaid/get-started/find-programs-and-services/maternal-support-services. Accessed: December 10, 2016.
  30. North Carolina Division of Medical Assistance. (2016). Medicaid income and resource requirements. Available at: https://dma.ncdhhs.gov/medicaid/get-started/eligibility-for-medicaid-or-health-choice/medicaid-income-and-resources-requirements. Accessed: February 1, 2017.
  31. O’Connor MJ, & Whaley SE (2007). Brief intervention for alcohol use by pregnant women. American Journal of Public Health, 97(2), 252–258. [DOI] [PMC free article] [PubMed] [Google Scholar]
  32. O’Hara MW, & Wisner KL (2014). Perinatal mental illness: Definition, description and aetiology. Best Practice & Research: Clinical Obstetrics & Gynaecology, 28(1), 3–12. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. O’Higgins M, Roberts IS, Glover V, & Taylor A (2013). Mother-child bonding at 1 year; Associations with symptoms of postnatal depression and bonding in the first few weeks. Archives of Womens Mental Health, 16(5), 381–389. [DOI] [PubMed] [Google Scholar]
  34. Roberts SC, & Pies C (2011). Complex calculations: How drug use during pregnancy becomes a barrier to prenatal care. Maternal and Child Health Journal, 15(3), 333–341. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Rutherford HJ, Williams SK, Moy S, Mayes LC, & Johns JM (2011). Disruption of maternal parenting circuitry by addictive process: Rewiring of reward and stress systems. Frontiers in Psychiatry, 2, 37. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Shaw E, Levitt C, Wong S, Kaczorowski J, & McMaster University Postpartum Research Group (2006). Systematic review of the literature on postpartum care: Effectiveness of postpartum support to improve maternal parenting, mental health, quality of life, and physical health. Birth, 33(3), 210–220. [DOI] [PubMed] [Google Scholar]
  37. Sit DK, & Wisner KL (2009). Identification of postpartum depression. Clinical Obstetrics and Gynecology, 52(3), 456–468. [DOI] [PMC free article] [PubMed] [Google Scholar]
  38. Stade BC, Bailey C, Dzendoletas D, Sgro M, Dowswell T, & Bennett D (2009). Psychological and/or educational interventions for reducing alcohol consumption in pregnant women and women planning pregnancy. Cochrane Database of Systematic Reviews, (2), CD004228. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Stone R (2015). Pregnant women and substance use: Fear, stigma, and barriers to care. Health & Justice, 3(1), 2. [Google Scholar]
  40. Thomas KC, Ellis AR, Konrad TR, & Morrissey JP (2012). North Carolina’s mental health workforce: Unmet need, maldistribution, and no quick fixes. North Carolina Medical Journal, 73(3), 161–168. [PubMed] [Google Scholar]
  41. Vesga-Lopez O, Blanco C, Keyes K, Olfson M, Grant BF, & Hasin DS (2008). Psychiatric disorders in pregnant and postpartum women in the United States. Archives of General Psychiatry, 65(7), 805–815. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Vliegen N, Casalin S, & Luyten P (2014). The course of postpartum depression: A review of longitudinal studies. Harvard Review of Psychiatry, 22(1), 1–22. [DOI] [PubMed] [Google Scholar]
  43. Wells R, Cilenti D, & Issel LM (2015). The political economy of a public health case management program’s transition into medical homes. Social Science and Medicine, 145, 98–106. [DOI] [PubMed] [Google Scholar]

RESOURCES