Moving toward integrated care for perinatal opioid use disorder and mental health in an urban safety-net hospital: A formative implementation evaluation

Participants

A total of 19 healthcare system employees, 15 external organization employees and five patients participated. We recruited a diverse range of employees within the system, including clinic managers, physicians, registered nurses, people working in clinical informatics and billing, mental health professionals and social workers. At external organizations we primarily spoke with leaders, such as a founder, director, or supervisor. We had one traditional healing coordinator participate. To ensure patient participants and community organizations reflected the diversity of the patient population at HHS, we purposely recruited organizations that served Black, Native American, and immigrant communities.

Patient participants identified themselves as Black/African American (n = 1), Native American (n = 1), multiracial (n = 1), and White (n = 2). Three participants identified as female, one genderfluid, and one as transgender. One participant was currently pregnant with their first pregnancy, and one participant each had one, two, four, or five children. Delays in approval to recruit patients, and restrictions on approaches to doing so, combined with patient participants being lost to contact between expressing interest at initial engagement and the interview, led to a relatively small number of patients being interviewed during the period of the study compared to other groups.

Procedures

Employees of HHS and community organizations were recruited directly via email invitations. Interviews lasted between 20 and 60 min. External organization employees received a $50 gift card for participation. Healthcare system employees participated during work hours and thus were ineligible for gift cards.

Eligibility criteria for patients were being over the age of 18, self-identified as a woman or birthing person, currently pregnant or had given birth within the past 5 years, and self-identified as having struggled with opioid use. We were prepared for non-English-speaking participants though received none. Flyers were distributed to multiple departments in the hospital and providers/staff were asked to post them in waiting rooms and hand them out to patients in a non-coercive manner. They offered the potential participant to either call the study team themselves to inquire about participation, or if the potential participant agreed for them to send the first name and phone number or email only via secure (Health Insurance Portability and Accountability compliant) messaging to the research team. The research coordinator then called participants, screened them, and set an interview appointment with a reminder call or email prior. Only first names were recorded to ensure participant confidentiality and a waiver of documentation of consent was approved by the HHRI Board and IRB. Patient participants received $50 gift cards for participating. After the end of the interview period, patient participants were offered the opportunity to participate in a survey, with a $20 incentive, refining their priorities and preferences for structure of the planned services. Data collection occurred January–September 2024.

Interview protocol

Three separate interview guides were developed. The healthcare and external organization employee guides identified demographics, population needs, organizational workflow and service descriptions, and feedback on the prototype model of care. The patient interview guide identified demographics, feedback on past and current experiences with healthcare and treatment, and feedback on the prototype model of care. After developing the interview guide and Data Collection Tool, the study team (GB, TS, and BA) completed mock interviews and pilot interviews and refined the tools.

The research team utilized the rapid qualitative analysis (RQA) approach, following established procedures developed by Hamilton, ¹⁷ and detailed by Hagaman and Rhodes. ¹⁸ RQA was completed in four steps. First, during participant interviews, as the PI facilitated interview questions, a second team member recorded participant responses using the Interview Observer Data Collection Tool to simultaneously organize the participant’s responses within each of the primary research questions. Second, the interview recording was then reviewed alongside this worksheet to ensure that all information was captured. Third, data was then synthesized in a Rapid Assessment Procedure (RAP) Sheet according to research questions. Finally, all data collection procedures were reviewed by the team at weekly meetings.

Measures

We utilized REDCap, a secure, online, data collection tool, to obtain survey data. Based on findings from the qualitative interviews, we developed a survey for a second round of patient input, with specific detailed questions such as ordering the different components of the model prototype by priority, as well as days and times for clinic groups, motivations and barriers for attending, and other resources they were currently utilizing.

Qualitative data analysis

Data from RAP sheets were used to develop codes in an inductive content analysis approach. ¹⁹ The final codebook resulted in 30 distinct codes, which were grouped into the six PRISM elements. ¹⁶ Codes and themes were reviewed and discussed by all data collection team members (GB, TS, and BA).

Organizational perspective: Break down the silos

There were three themes related to organizational perspectives which centered on addressing barriers of frontline staff and coordination across departments and specialties. First, all groups described organizational barriers that prevented patients from accessing substance use-specific care at their current point of care (Table 1, #1). Second, hospital employees highlighted organizational logistics that could impede program implementation, including siloed billing structures run by individual department managers which, along with a scarcity mentality, sometimes resulted in ‘turf wars’ or protective maneuvers. Other logistical concerns included a lack of clear policies and procedures regarding substance use during pregnancy, including child welfare reporting and other healthcare recommendations (e.g., encouraging lactation vs formula-feeding). At the same time, the need for coordination and desire for change across departments and specialties was shared by hospital employees and external organizations.

Patient perspective: Help them feel safe and supported

There were 12 codes related to patient perspectives, mostly focused on the need for providing patients with choices, patient-centered care and addressing patient barriers to care.. All groups agreed on the importance of providing patient choices. First, several strategies to support patient accessibility were discussed by both hospital employees and external organizations (Table 1, #2). All groups also discussed flexible and strategic scheduling as a facilitator of care. A variety of care formats were also an important factor to facilitate care mentioned by all groups (Table 1, #3). Relatedly, all groups discussed several desired services or programs as important for PPW with OUD. Available childcare options were a priority to address patient barriers and support participation in treatment programs.

Several barriers to care were discussed by participants, including patient barriers related to attrition, stereotyping and stigma (Table 1, #4). Lack of trust and stigma concerns between patients and providers meant there was a need for increased compassion from providers to build trust (Table 1, #5 & 8). Parenting support needs were also discussed including the overwhelming nature of child protective services (CPS) documents, developing and following a parenting plan, and the fear of CPS reporting. All groups identified stigma as an important barrier to accessing services. One example includes the importance of what happens post-delivery to the babies born to PPW with OUD. Because many of them experience neonatal abstinence syndrome (NAS) and are hospitalized for an extended period in the neonatal intensive care unit (NICU), the experience of parents hoping to take their babies home is critical. However, the setup of the NICU as an open space with minimal privacy was found to have the potential to an trigger self-stigmatizing thoughts and/or other-stigmatizing experiences. One employee believed that the rate of children sent to foster care (rather than parents being able to take the baby home) at this hospital compared to other services in the area could be directly related to this NICU set-up, and efforts were underway to find private rooms for these babies to facilitate parent engagement in the Eat, Sleep, Console approach to NAS. ²⁰ The stigma of substance use in pregnancy and postpartum was inextricably connected with racial discrimination, although while participants thought inclusion of cultural elements, such as availability of hospital chaplains and culturally specific foods, would be useful, most seemed to think that a program that served all cultures was needed. This finding may have been particularly salient due to the hospital being a safety-net health system with a large percentage of Black, Indigenous and People of Color and immigrant patients, whereas hospital systems in other settings may find different discriminatory themes could apply. Finally, support for families, namely lack of support for partners, was noted as a barrier to change and as a service some would like to see (Table 1, #6).

External organizations noted that the lack of patient-centered care related to the patient’s cultural background could prevent access to care (Table 1, #7). Strategies for addressing cultural considerations were discussed. Hospital employees reported on the importance of addressing patient needs to support implementation of a new program (Table 1, #9).

External environment: Fix a broken system

There were four codes for the external environment related to community resources, cultural environment, and competition. First, hospital employees and external organizations discussed community resources related to care coordination and navigation as a common barrier. The impact of the external cultural environment was also discussed as an important consideration when evaluating care. Most specifically, Minnesota is one of approximately 24 states where prenatal substance use is reportable to child protective services. ²¹ Prior to 2021, this was a requirement of all mandated reporters and was a gross misdemeanor with possible jail time if violated. In 2021, a coalition of healthcare providers and researchers facilitated a change in the law, making reporting optional rather than mandatory as long as the patient is receiving some sort of healthcare; however, there was a lack of awareness among most healthcare providers about this change to the law and a continued wariness about not reporting due to concerns about harm to the fetus. This culture of reporting has led to a lack of pregnant people seeking prenatal care out of fear of CPS.

Finally, regarding competition in the external environment, the conversation tended more toward collaboration with complementary organizations rather than any direct competition (Table 1, #10). External organizations also noted the importance of community involvement (Table 1, #11).

Implementation and sustainability infrastructure: Resources needed

There were four codes related to implementation and sustainability infrastructure shared by participants. These codes were related to sharing best practices and dedicated teams. Hospital employees and external organizations described the importance of facilitating sharing of best practices through departmental training and collaboration to facilitate care. Hospital employees and external organizations noted staff education was needed (Table 1, #12). Providers outside addiction medicine, such as in psychiatry and obstetrics and gynecology, while open to prescribing MOUD, identified the need beyond formal training for informal, longer-term support in building their knowledge and confidence in prescribing buprenorphine and other medications for OUD.

Several specific roles within a dedicated team were mentioned by participants for supporting implementation and sustainability infrastructure. Peer recovery staff were prioritized by all groups (Table 1, #13). Finally, other staff specialties were prioritized by all groups; but hospital employees and external organizations identified difficulties in funding these roles (Table 1, #14).

Organizational characteristics: Come together

There were two codes describing organizational characteristics related to staffing and incentives, shared goals and cooperation, and systems and training shared by participants. Hospital employees described difficulties with staffing, specifically their difficulty maintaining full staffing (Table 1, #15), which was related to financial pressures, and frustrations around difficulty cooperating across departments, where efforts to address this population were being duplicated (Table 1, #16). While duplication of efforts was discussed as frustrating, it also demonstrated shared goals to improve care and communication in service of treating PPW with OUD better. All employees interviewed were excited at the prospect of change for this population and believed the proposed model of care would contribute to that improvement.

Patient characteristics: This is hard

There were five codes outlining patient characteristics which primarily focused on disease burden, competing demands, and demographics. Several common factors related to disease burden experienced by PPW with OUD were described by participants. Hospital employees, external organizations, and patients all described aspects of patient resources and logistics as increasing difficulty to access services. Hospital employees and external organizations also described patient psychosocial factors as a common barrier to accessing services (Table 1, #17). All groups mentioned prenatal care as important for facilitating care for PPW with OUD. All groups discussed the urgency of meeting patients’ basic needs to support recovery.

All groups noted the importance of services to meet the needs of patients with specific demographic characteristics. One participant recommended involving the hospital’s chaplains for mothers with certain religious or cultural affiliations.

Disagreement between participant groups

While participants generally agreed on content shared within each code, there were several areas in which participants within and between participant groups disagreed. For example, while all groups mentioned the importance of services designed to support specific patient demographics, whether or not male partners should be supported in the same clinic and PPW with OUD was not consistent across respondents. Some participants felt that lack of support for male or non-birthing partners was a barrier to change and several hospital employees expressed a need for partner support; however, other hospital employees, external organizations, and patients advocated for a space only for mothers and birthing people. One concern was that mothers and birthing parents may feel less safe due to the high rate of intimate partner violence in this population generally. Additionally, when discussing scheduling, patients expressed differing opinions on the most accessible dates and times to facilitate care (e.g., days of the week, morning vs afternoon).

Overall, the findings of the study indicated: barriers at the individual and system levels for this patient population to engage in treatment; facilitators or strengths already existing at HHS; specific priorities of care and needs of healthcare system employees, patients, and the larger community to consider when developing and refining the care model; and suggested implementation strategies for overcoming the barriers and leveraging the facilitators to successfully implement the new model of care.