Northern New England has among the highest rates of opioid dependence in the U.S, with
prevalence highest and growing among those aged 18-35 years. Regional rates of perinatal
opioid use disorders (OUD) reflect this public health crisis; Northern New England region has
the highest incidence of opioid-related births in the nation. Perinatal opioid use disorders
impact 5-8% of pregnancies in Maine, New Hampshire, and Vermont, three states that have been
severely impacted by the current opioid epidemic. Consequences of inadequately treated OUD
include premature delivery and other perinatal complications, prolonged newborn
hospitalization for neonatal abstinence (NAS), and maternal morbidity and mortality from
infectious disease and overdose. In 2013, costs associated with NAS treatment in the U.S.
reached $1.5 billion. The epidemic has grown multifold since 2013.
Medication assisted treatment (MAT) with either methadone or buprenorphine is the recommended
standard of care during pregnancy. Professionals, including the American College of
Obstetricians and Gynecologists (ACOG), the American Academy of Pediatrics, the Society for
Maternal-Fetal Medicine, and the American Society for Addiction Medicine, have called to
adopt care models that promote early identification and treatment for pregnant women with
OUD. Referral to specialty programs for MAT has been the accepted standard of care, with
demonstrated safety for women and their infants. However, in response to the escalating
opioid crisis, the ACOG began offering buprenorphine training programs to its members.
Consequently, a number of maternity care practices throughout Northern New England now
provide integrated MAT services. But even where such integrated programs exist, questions
persist on the optimal care model for providing MAT to pregnant women with OUD.
There are patient- and provider- factors associated with variability in effectiveness of what
MAT model works best for whom. For women, advantages of buprenorphine over methadone include
a lower risk of overdose, fewer drug interactions, the accessibility of office-based
treatment delivery in the context of maternity care and demonstrated shorter NAS course. The
disadvantages of buprenorphine relative to methadone include potential hepatic dysfunction,
lack of long-term data on consequences of fetal exposure for infants, potential limited
efficacy in patients with high addiction severity, requirement of moderate withdrawal
symptoms prior to initiation to avoid iatrogenic withdrawal, and an increased risk of
diversion (i.e., sharing or sale). Despite buprenorphine's demonstrated neonatal advantages,
it is not effective for all women. The structure of methadone treatment (daily meeting) may
also better align with support needs for some women. For providers, the choice of what
medication-assisted approach to offer to patients is often restricted by availability and
access to specialty care services.
There is limited literature comparing the effectiveness of integrated versus referral MAT
care models for postpartum retention in treatment and women's experiences in these two
models. In particular, women with high levels of addiction severity or co-occurring mental
health conditions may have prenatal care needs from women with less complex behavioral health
concerns. Studies have also not assessed the impact of maternal opioid addiction severity on
newborn outcomes and maternal long-term recovery. The best neonatal outcomes will be achieved
by providing the most appropriate and effective treatment for mothers.
The challenge for patients, providers, and other stakeholders is determining the optimal
approach for delivering MAT during pregnancy and after delivery to improve outcomes.
Currently, regional obstetrical practices provide either: (1) Integrated care, in which MAT
and associated psychosocial services are delivered on site with obstetrical care, and (2)
Referral-based care, in which women receive MAT and obstetrical care in separate, specialized
locations. Both models have different potential advantages and disadvantages for mothers with
OUD and their babies regarding access, availability, acceptability and quality of MAT,
obstetrics and other needed services, and long term follow up and treatment after delivery.
While pregnancy motivates women to initiate MAT, relapse to use of opioids and cessation of
MAT frequently occur during the first postpartum year, placing both mother and infant at
significant risk. Payers and policymakers are also seeking answers about where to invest
healthcare resources to increase access to treatment for pregnant women, especially in the
current opioid crisis. A disproportionate number of women with OUD are insured through the
Medicaid system; Medicaid policy affects women's access to a wide range of services, from
reimbursement for same-day services or care coordination, to whether a woman continues to be
eligible for benefits after delivery.
Study Aims: This study aims to answer these important patient, provider, and policy questions
by comparing the real-world effectiveness of two models of MAT delivery currently in clinical
use in Maine, New Hampshire, and Vermont with respect to patient experience of care and
perinatal, neonatal, and longer-term substance use treatment outcomes.
Two main Comparative Effectiveness Research (CER) Patient-Centered Outcomes Research
Institute (PCORI) priority questions will be addressed:
Do clinical and patient-reported outcomes for pregnant and parenting women differ
between integrated and referral-based MAT practice models?
Within models, which psychosocial services are most associated with MAT continuation,
and for which groups of patients?
To answer these questions, the following specific aims will be addressed:
Aim 1 (Clinical Outcomes). To use clinical record data to evaluate the comparative
effectiveness of Integrated and Referral-Based MAT care models on maternal and neonatal
outcomes.
Aim 2 (Patient-reported Outcomes). To use patient reported data to evaluate the comparative
effectiveness of Integrated and Referral -Based care models on patient-centered outcomes.
Aim 3 (Heterogeneity of Effects). To examine differences in treatment retention within
condition by subgroups of patients based on (1) psychiatric comorbidity, (2) type of
medication used for MAT and (3) addiction severity.
Aim 4 (Specification of Services). To determine which services (psychosocial services, care
coordination, parenting education) are associated with better maternal and neonatal outcomes.
Aim 5 (Provider). To explore how provider attitudes about MAT and care of patients with OUD
vary by care model and are associated with maternal outcomes.
Study Description: This study will collaboratively engage 21 practices providing maternity
care across Northern New England, with examples of both integrated care and referral-based
models. The study population is pregnant women who receive prenatal care from any of these
practices and who meet Diagnostic and Statistical Manual of Mental Disorders (DSM-5)
diagnostic criteria for an opioid use disorder. A cluster-based, prospective observational
mixed-methods design will be used to compare outcomes for pregnant women with opioid use
disorder receiving prenatal care in obstetric practices that offer MAT through one of two
delivery models: Integrated or Referral-based. Clinical records data (n=2000) from pregnant
women with OUD receiving prenatal care at a partner practice will be used to examine Aim 1.
Aim 2 will be addressed with a patient-report subsample cohort (n=532) recruited in the 3rd
trimester of care and followed to 6 months postpartum. Aim 3 will use both kinds of data to
explore heterogeneity of treatment effects. For Aim 4, practice-level data will be collected
yearly to evaluate services provided across Integrated and Referral-based practices. In Aim
5, survey and qualitative interviews with providers will lend perspective on facilitators and
barriers to MAT in both Integrated and Referral-based practice settings. The analytic
strategy will account for clustering and patient baseline differences to compare outcomes
across assessment points. Whether the effect of treatment type differs according to
psychiatric status, type of MAT patients access, or addiction severity will also be tested.
Analysis of qualitative data will inform our interpretation of quantitative results and
enhance our understanding of patient experience, as well as barriers and facilitators to
receiving care within these care models. Patient representatives, practice-partners, state
and regional stakeholders, and scientific advisors will actively guide all stages of this
study and dissemination of results to relevant partner networks.