This project provides a broad look at SUD-related services available to pregnant and postpartum patients across both obstetric and SUD treatment settings in Baltimore, Maryland. To our knowledge, it is the first comprehensive, cross-disciplinary assessment of perinatal OUD service availability in a major US city.

Overall, our results show that most birthing hospitals in Baltimore often care for perinatal patients with OUD, but the majority of PNC practices and SUD programs in the city care for this population less frequently. This imbalance may be driven by several factors including the greater number of PNC and SUD programs available to patients relative to birthing hospitals, preferential referral to specialty outpatient programs, lack of patient engagement in care prior to delivery, and/or underdiagnosis/lack of identification of SUD in the outpatient setting. Additionally, our data show considerable variability in screening practices and in service availability (both by type of organization and by type of service). Birthing hospitals and SUD programs generally provide more SUD-related services than PNC practices. MOUD and linkage to treatment are generally more available than withdrawal management and harm reduction services.

Our data demonstrate a lack of consistent validated SUD screening, availability of MOUD, and naloxone accessibility in prenatal/obstetric settings, representing significant missed opportunities for treatment engagement and overdose prevention. While most birthing hospitals and PNC practices reported screening in some way for OUD, organizations frequently utilized informal screening and/or urine toxicology. Notably, our results show minimal improvement in the decade since prior statewide assessment demonstrated that a minority of obstetric settings used validated substance use screening tools, considered standard of care by many professional and governmental organizations [11,12,13, 34]. Informal approaches may lack reliability and introduce potential for bias, and urine toxicology may be used punitively [35,36,37]. Additionally, while all birthing hospitals reported the ability to continue MOUD for patients, two of seven birthing hospitals and three-quarters of PNC practices do not initiate MOUD. Perhaps most notably, around two-thirds of responding birthing hospitals and PNC practices reported not offering any naloxone to patients.

While most SUD programs offering MOUD reported no restrictions related to pregnancy status, our data reveal several potential barriers to treatment engagement. About a quarter of SUD programs reported requiring obstetric evaluation prior to MOUD initiation, which may be an obstacle to the OUD stabilization needed to facilitate engagement in prenatal care. Additionally, few programs offer tailored services that can increase the relevance and accessibility of treatment for pregnant patients or for parents of young children. While most residential programs and recovery residences did report accepting patients during pregnancy, far fewer allow children onsite. (Of note, the higher rate of permitting children among recovery residences in our sample [50%] may be inflated by the low response rate among recovery residences and overrepresentation of programs specifically catering to this population.) Even among outpatient SUD providers, only about half permit children to accompany parents to treatment. These limits likely lead to a significant loss of treatment options after delivery for many patients given competing childcare responsibilities [13, 29].

Withdrawal management services for the perinatal population were also a notable gap within the Baltimore landscape. With regards to OUD, several of the dedicated withdrawal management/stabilization programs surveyed indicated denying buprenorphine management to pregnant patients when it was otherwise available to the general population. Alcohol/benzodiazepine withdrawal management, a critical service to facilitate safe cessation of these substances, was also unavailable in many locations, including SUD programs overall and particularly among PNC practices. This lack of availability may in part be due to the fact that these services, particularly for cases of severe withdrawal, often merit inpatient-level monitoring; however, even two of seven birthing hospitals and two of five alcohol/benzodiazepine withdrawal management programs reported denying this service during pregnancy.

Among the most commonly endorsed needs by respondents were education/specialist consultation on SUD treatment in pregnancy within obstetric settings and on pregnancy-related medical concerns within SUD programs. These endorsements mirror MMRC recommendations to increase PNC and SUD provider comfort caring for patients with perinatal OUD and to reduce silos between these disciplines [5, 31, 32]. Despite the limited availability of naloxone for overdose reversal among birthing hospitals and PNC practices, very few endorsed that they would benefit from the ability to distribute naloxone directly to patients. Lack of interest in providing this service may reflect an unawareness of the prevalence of overdose during the perinatal period as well as the potential benefit of directly distributed naloxone, compared with prescribed naloxone, in accessibility and reduction of overdose deaths [38,39,40].

Overall, our results highlight strengths and weaknesses within the landscape of perinatal SUD services in Baltimore and potential avenues for quality improvement. Within Baltimore, results from this project have served as a critical guide to our team’s ongoing capacity building efforts citywide. In particular, dissemination of results to multiple local professional organizations/networks and mortality review committees has offered the opportunity for education about maternal overdose mortality as well as potential ways to strengthen care for this population. Additionally, results have enabled the project team to identify and conduct targeted outreach and technical assistance to organizations that were either already providing care to this population or had the potential to fulfill an unmet need; 88 clinicians across 10 SUD treatment and/or prenatal care organizations have been engaged thus far. Depending on specific organization needs, outreach offerings have included provider education on standards of SUD care in pregnancy, promotion of population-specific community resources (e.g., case management, legal support, doula care), supported linkage to obstetric resources, and longitudinal consultation on development of population-specific programming. While this landscape assessment is specific to Baltimore City, the patterns identified may be relevant more broadly. This project may serve as a model for other jurisdictions interested in better understanding their local landscape and strengthening services for patients using opioids during pregnancy and the postpartum period.

Data from our project should be interpreted in light of both its strengths and limitations. Our assessment was unique in its comprehensiveness–both due to the broad multidisciplinary sample of organizations included as well as the range of topics addressed. Additionally, our systematic contact confirmation and outreach efforts yielded a strong response rate of 61%. However, not all organizations providing services to pregnant patients with OUD in Baltimore City are represented in this data. Some organizational categories, such as certified-recovery residences, yielded lower response rates than others, and so data about these programs may not be as broadly applicable. Additionally, some SUD-related organizations or providers, particularly those providing more informal care (such as non-certified recovery residences or individual buprenorphine prescribers outside of larger SUD-focused practices), may not have been captured by our outreach lists. For a small number of organizations on the initial outreach lists, our team was unable to identify individual respondents, leading to their exclusion. Additionally, those organizations with greater focus on this specialized population likely had a higher response rate, which may artificially increase the reported rate of service availability. Importantly, reported service availability may not accurately reflect true access to services, quality of care provided, and/or the patient experience receiving care. All of these factors may result in lower actual accessibility of services than is represented in our data.