The extent of implementation and perceptions of maternity and social care professionals about two interprofessional programs for care for pregnant women: a mixed methods study

Abstract

Background

In 2018, the Dutch government initiated the Solid Start program to provide each child with the best start in life. Key program elements are a biopsychosocial perspective on pregnancy and children’s development and stimulating local collaborations between social and health domains, with a specific focus on (future) families in vulnerable situations. Two programs for interprofessional collaboration between maternity and social care professionals to optimize care for pregnant women in vulnerable situations were developed and implemented, in Groningen in 2017 and in South Limburg in 2021. This paper describes the extent of implementation of these programs and the perceptions of involved professionals about determinants that influence program implementation.

Methods

We conducted a mixed-methods study in 2021 and 2022 in two Dutch regions, Groningen and South Limburg. Questionnaires were sent to primary care midwives, hospital-based midwives, obstetricians (i.e. maternity care professionals), (coordinating) youth health care nurses and social workers (i.e. social care professionals), involved in the execution of the programs. Semi-structured interviews were held with involved professionals to enrich the quantitative data. Quantitative and qualitative data were collected and analyzed using Fleuren's implementation model.

Results

The findings of the questionnaire (n = 60) and interviews (n = 28) indicate that professionals in both regions are generally positive about the implemented programs. However, there was limited knowledge and use of the program in Groningen. Promoting factors for implementation were mentioned on the determinants for the innovation and the user. Maternity care professionals prefer a general, conversational way to identify vulnerabilities that connects to midwives’ daily practice. Low-threshold, personal contact with clear agreements for referral and consultation between professionals contributes to implementation. Professionals agree that properly identifying vulnerabilities and referring women to appropriate care is an important task and contributes to better care. On the determinants of the organization, professionals indicate some preconditions for successful implementation, such as clearly described roles and responsibilities, interprofessional training, time and financial resources.

Conclusions

Areas for improvement for the implementation of interprofessional collaboration between maternity care and social care focus mainly on determinants of the organization, which should be addressed both regionally and nationally. In addition, sustainable implementation requires continuous awareness of influencing factors and a process of evaluation, adaptation and support of the target group.

Introduction

Preconception, pregnancy and the first two years of life (the first thousand days) are crucial for children’s development and health, and a decisive period in the emergence of health inequities [1, 2]. Pregnant women in a vulnerable situation have a higher risk of adverse perinatal outcomes [3].

The concept of vulnerability during pregnancy seems to be complex and multifaceted and therefore difficult to address. A concept analysis by Briscoe et al. [4] defines vulnerability during pregnancy, birth and the postnatal period as “women are vulnerable when they experience ‘threat’ from a physical, psychological or social perspective, where ‘barriers’ and ‘reparative’ conditions influence the level of vulnerability”. A more recent Dutch definition describes vulnerability in pregnancy as an imbalance between psychosocial and/or medical risk factors, and protective factors [5]. There are various factors that can make pregnant women vulnerable, such as being a teenager, smoking, partner violence or living in a deprived neighborhood [6,7,8]. At the same time, there are also factors or circumstances that protect women from becoming vulnerable or reduce vulnerability, such as having a social network, being self-reliant or having good health literacy skills [5, 7].

Research shows that pregnant women in vulnerable situations with an accumulation of risk factors are more likely to receive less adequate care due to the complexity of case management [9]. This requires collaboration between the social and health domains that considers both medical and non-medical risk factors relevant to a pregnant woman and her social and physical environment. Two systematic reviews underlined the importance of interprofessional collaboration to improve health outcomes [10, 11] and identified key elements for successful collaboration among maternity care professionals, such as good communication, mutual respect and support for colleagues [10].

Internationally, multiple countries have implemented programs and policy reforms to reduce health inequities by integrating medical and social services in early life [9, 12,13,14]. A systematic review of preventive interventions targeting pregnant women with psychosocial vulnerabilities aiming to reduce preterm birth identified weak evidence on the effectiveness of interventions, such as home visits, group meetings, phone calls and massage [15].

Interventions tailored to the individual woman’s needs, circumstances and context are more likely to have a positive effect.

In the Netherlands, the nationwide action program ‘Solid Start’ was launched by the Dutch Ministry of Health, Welfare and Sport in 2018 [1, 16]. The program aims to provide each child the best start in life by stimulating biopsychosocial perspective and local collaborations between social and health domains tailored to local circumstances, with a specific focus on (future) parents and young children in vulnerable situations, so called ‘social maternity care’ [8]. Making these collaborations is challenging because medical and social care have different legislation, funding, organizational structure and culture. Social maternity care is not part of routine prenatal care in the Netherlands (Table 1).

Table 1 Maternity care and social care for women in vulnerable situations

Especially in the north (Groningen) and southeast (South Limburg) of the Netherlands we see higher adverse perinatal outcomes than in other Dutch regions due to an accumulation of medical as well as non-medical risks such as lifestyle and socio-economic circumstances [8, 17,18,19]. Poverty and deprivation have substantial effects on perinatal outcomes [8]. This effect seems to weigh more heavily among Dutch Western pregnant women [19]. The combination of Western pregnant women and living in poverty is more common in both regions [20].

As a response to these poorer perinatal outcomes in Groningen and South Limburg, stakeholders (i.e. professionals, policy makers, researchers) developed and implemented two programs for interprofessional collaboration between maternity care professionals (i.e. midwives, obstetricians, maternity care assistants) and social care professionals (i.e. youth health care nurses, social workers) consisting of two components: 1) an identification tool for psychosocial risk factors and 2) work agreements for referral for social support (see Table 2). These programs are similar in their aims of identifying pregnant women in vulnerable situations and collaborating with Preventive Youth Health Care (YHC) to refer them, if necessary, to the social domain for additional tailored care. But they include different components and are therefore different in how these aims are achieved. Proper implementation of these programs has an effect on their effectiveness for women in a vulnerable situation [21].

Table 2 Programs in Groningen and South Limburg

An innovation process consists of four phases, namely dissemination, adoption, implementation and continuation. Each of these phases can be seen as steps in which, potentially, the desired change does or does not occur. The transition from one stage to the next is affected by various determinants, which can be divided into: factors related to the innovation (e.g. complexity, relevance), the user (e.g. skills, knowledge, social support), the organization (e.g. resources, feedback), and the socio-political context (e.g. legislation) [22, 23].

To better understand what works for whom under what circumstances in further developing and implementing social maternity care, it is important to assess, after dissemination, the extent to which maternity care professionals and social care professionals in Groningen and South Limburg have adopted the programs and what determinants influence the next step of implementation in actual care settings [24]. This knowledge can be used to design appropriate and effective innovation strategies adapted to these determinants. Therefore, the research question of this study is as follows: To what extent have maternity and social care professionals adopted and implemented the two programs, and what determinants influence program implementation?

Methods

Two interprofessional programs

In the municipality of Groningen, an interprofessional team consisted of midwives and obstetricians from two Maternity Care Networks, a staff member from an organization for maternity care assistance and staff members from the YHC organization agreed to implement an interprofessional program starting January 2017 [25]. The program comprises several steps and options and starts with identifying vulnerabilities by midwives with the ALPHA-NL identification tool [26]. This is a questionnaire that pregnant women complete and discuss during the second appointment with their midwife. If risk factors for vulnerability are identified and the woman consents to request interprofessional support, the midwife can 1) contact YHC by phone for further guidance and to request support for the woman and/or 2) discuss the case in a multidisciplinary consultation with medical and social care providers and/or 3) decide on supportive care indicated in 17 care pathways.

In South Limburg, we developed a program in 2021 in an interprofessional working group consisting of maternity care providers, nurses and staff members of YHC, social workers, social pediatricians, policy makers and researchers. To identify factors for vulnerability, the regular questionnaire for psychosocial and medical history was updated. The Positive Health dialogue tool (Spider web and Alternative dialogue) was introduced at the start of the prenatal consultations and at 30–34 weeks of gestation. The Positive Health dialogue tool can be used to address individuals’ health from a broader perspective, and discuss their needs, values and preferences [27, 28]. Each midwifery practice was linked to a coordinating YHC nurse who can be consulted for further exploration, support and, if necessary, referral to additional services in the social domain. The program was implemented in a pilot study involving 10 midwifery practices in South Limburg.

Implementation of the programs took place by informing all relevant healthcare and social care professionals about the programs and their content. Manuals were developed and disseminated. In Groningen, midwives were already trained in 2017 in applying the ALPHA-NL identification tool. In South Limburg midwives and coordinating YHC nurses were trained in April–May 2021 in applying Positive Health.

Design

To answer our research question, we used a mixed method sequential explanatory design. Collected quantitative data were enriched with in-depth interviews to gain broad understanding [29]. The study consisted of two parts:

• 1. A quantitative survey in both regions among professionals involved in implementing the programs to identify the extent of implementation and perceptions about the implementation of the programs.

• 2. A qualitative study consisting of semi-structured in-depth interviews with professionals who were involved in the implementation of the programs. We used the results from part 1 as input for part 2.

The implementation model of Fleuren et al. [22] was used as a framework for the design of the study. This framework represents the main stages in innovation processes and related categories of determinants and is based on a literature review and a Delphi study among implementation experts. It a generic framework used in the Netherlands and other countries for the introduction and evaluation of innovations in healthcare and education since 1999. Based on this framework, a short instrument was developed to measure determinants of innovations that may influence their implementation: the Measurement Instrument for Determinants of Innovations (MIDI) [30]. In this study, we used the MIDI for the survey study (part 1) and the topic guide of the semi-structured interviews (part 2) and we used Fleuren's implementation model to structure and interpret the results.

Part 1: survey study

In part 1 we conducted a cross-sectional survey among involved professionals. We developed a questionnaire, including questions about the extent of implementation and the Dutch MIDI [30, 31]. To assess the extent of implementation among users, we asked two questions: ‘Do you know the program?’ (yes/no) and ‘Are you using the program as intended?’ (yes, as described in the manual/partly/no). We repeated these two questions for the different components within the two programs. The MIDI is based on Fleuren's implementation model [22]. This model describes 29 determinants that can enhance or hinder the implementation of an innovation, divided into determinants of the innovation (7 determinants), determinants of the prospective user (11 determinants), determinants of the organization (10 determinants) and determinants of the socio-political context (1 determinant). Understanding these determinants, which influence the actual use of an innovation, is important to develop a tailored follow-up strategy for sustainable implementation. The Dutch MIDI questionnaire provides 29 validated example questions. Depending on the type of innovation, target group and stage of implementation, relevant questions can be selected. The research team selected 22 questions from the Dutch MIDI questionnaire and rephrased them to suit our own innovation and context. The items were rated on a five-point Likert scale ranging from [1] ‘completely disagree’ to [5] ‘completely agree’ or on a dichotomous scale ‘Yes/No’.

Additionally, demographic variables and work characteristics were collected. In both regions, the survey was the same, except for the questions that specifically addressed program components.

Study population

In Groningen, where the program was broadly implemented, the study population consisted of midwives of all midwifery practices (n = 27), hospital-based midwives and obstetricians (in training) from two hospitals (n = 60), and social care professionals and YHC nurses (n = 48). In South Limburg the study population consisted of primary care midwives from the ten participating midwifery practices (n = 31) and coordinating YHC nurses (n = 3).

Recruitment

Data collection took place in November 2021. All professionals in the study population received an e-mail with information about the study and a link to the online questionnaire. Participants’ informed consent was obtained before they could start the questionnaire and participants could decide to stop filling out the questionnaire without explanation at any moment. In both regions, all potential respondents received three reminders by mail.

Data analysis

The questions on the extent of implementation of the programs were analyzed by combining the two questions and the response options. This resulted in four categories indicating the extent of implementation: [1] I do not know the program and do not use it; [2] I know the program but do not use it; [3] I know the program and partially use it as agreed upon in the manual; and [4] I know the program and use it as agreed upon in the manual.

All survey data were analyzed using descriptive statistics by calculating frequencies and percentages using IBM SPSS statistics 25.0.

Part 2: semi-structured interviews

Study population and recruitment

In Groningen, we aimed to include one hospital-based midwife and/or obstetrician per hospital, a midwife from each midwifery practice, several YHC nurses and social workers. In South Limburg, we aimed to include one midwife from each participating midwifery practice, and several coordinating and community-based YHC nurses. All professionals were approached by mail and telephone to inform them about the study and to invite them to participate. They gave informed consent to participate in the study.

Data collection

Data collection took place from February -May 2022. Prior to the interviews a topic guide was prepared and discussed by the researchers. The topics in the topic guides were derived from the results of the survey study and the implementation model of Fleuren et al. [22] The topic guides are presented in Table 3. Two researchers conducted the interviews together. Depending on the participant’s preference, the interview could take place online or on-site. Online interviews were held via Microsoft Teams. All interviews were recorded by audio. At the end of each interview, a summary of the interview was given by the interviewer to check for additions. The audio recordings with Microsoft Teams had already been transcribed and were checked and adjusted using the audio recording. The other audio recordings were transcribed verbatim.

Table 3 Topics in interview study

Research team

The research team consisted of three senior and three junior researchers, all women, from different backgrounds. MH is a health care scientist, nurse and lecturer; DD is a midwife and lecturer, JO is a health care scientist in public health, midwife and lecturer, LQ is a midwife, medical anthropologist and sociologist, MvdHB is a midwife and lecturer, and EFdJ is a midwife, lecturer, and health care scientist.

These various backgrounds represented a diversity of professionals’ knowledge and research traditions and methods. Throughout the research team meetings, we reflected on the methodological choices that we made and the perspectives we chose. This provided us with a deep and broad understanding of our research topic.

Data analysis

We used an inductive content analysis to analyze the transcripts of the interviews by moving from inductive coding to a framework analysis [32]. We developed a coding scheme, and the open codes were placed into categories according to the implementation model of Fleuren [22] by three researchers. The analysis took place using MaxQDA (Groningen) or NVivo 11 (South Limburg).

Results

Respondents

In Groningen, 43 health care providers completed the online questionnaire: 18 primary care midwives (response rate 66,7%), three hospital-based midwives and seven obstetricians (in training) (response rate 20,8%) and 15 YHC nurses (response rate 31.3%). The primary care midwives had between one and thirty-two years of work experience in midwifery care and worked in a group practice (n = 10), a duo practice (n = 6) or a solo practice (n = 2). The hospital-based midwives had between two and ten years of work experience, the obstetricians (in training) had between one and thirty-three years and the YHC nurses had between two and thirty-one years of work experience.

In South-Limburg, 17 health care providers completed the online questionnaire (response rate 50%), three coordinating YHC nurses (100%) and 14 midwives (45%). The midwives had between four and thirty years of work experience in midwifery care and worked in a group practice (n = 11) or in a duo practice (n = 3). Coordinating YHC nurses had two or three years of work experience in this role. We had no background information on the non-responders of the survey study.

We conducted a total of 28 in-depth interviews with professionals involved in the implementation of the programs. In Groningen, we conducted 15 interviews with eight primary care midwives, four hospital-based midwives or obstetricians, two social workers and one YHC nurse. In South Limburg, we conducted 13 interviews with nine primary care midwives, two coordinating YHC nurses and two community-based YHC nurses.

The extent of implementation Groningen and South Limburg

Table 4 shows the extent of implementation of the programs in Groningen and South Limburg. In Groningen, the majority of the respondents indicated that they were familiar with the program, but only one third indicated that they used the program as intended. With respect to components of the program, midwives used the telephone consultation, the multidisciplinary consultation and the care pathways. The ALPHA-NL identification tool is hardly used by midwives.

Table 4 The extent of implementation of two programs among professionals in Groningen (n=43) and South Limburg (n=17)

In South-Limburg almost all professionals know and use the program as intended. All midwives know and use the updated questionnaire for psychosocial and medical history and the appointments with YHC. The elements of Positive Health (Spider web and Alternative dialogue) are also used in whole or in part by almost all midwives in the first trimester of pregnancy. Positive Health is hardly used in the third trimester of pregnancy (30–34 weeks of gestation), while midwives do indicate that they know this appointment.

The three coordinating YHC nurses are all familiar with the program and work according to the agreements made between midwives and YHC. Positive Health (Spider web and Alternative dialogue) is not used by the coordinating YHC nurses.

Professionals’ perceptions about the program

First, we present the results of the survey study and the interview study among professionals in Groningen, and subsequently those of South-Limburg. We include some quotes from the participants of the interview study. The reference at the end of the quotes refers to the type of professional (midwife (MW), obstetrician (OB), youth health care nurse (YHCN)), the number of the interview (1,2,3,….) and the region of the respondent (Groningen (G) or South Limburg (SL)).

Table 5 presents the results of the survey study of professionals in Groningen and South Limburg.

Table 5 Determinants for implementing the programs in Groningen and South Limburg as measured by the Measurement Instrument for Determinants of Innovation

Groningen

Determinants of the innovation

Twenty-nine respondents found the program clear, and 24 respondents found it complete and relevant. This was in line with the findings of the interview study. The components of telephone counseling, multidisciplinary consultation and care pathways improve the identification of vulnerabilities and collaboration between care providers. Telephone counseling was perceived as low-threshold and registering pregnant women, who consented to it, took little effort and time.

“….it is very nice to be able to call this phone number, taking some of the work off your hands. Because in the beginning it was that you thought you had to solve everything for a pregnant woman with a problem, and that is not the case anymore with Telephone Counseling.” (MW4G).

Telephone counseling helped to gain a better understanding of the work and tasks of the YHC and thus fostered collaboration. However, telephone counseling was used mainly for complex and multiple problems and less in more straightforward situations that required a more preventive approach. Furthermore, participants in the interviews mentioned that YHC nurses did not always pick up the supportive role after referral via telephone counseling. The multidisciplinary consultation contributed to a better and interdisciplinary approach around a family.

Participants in the interview study indicated that the use of the ALPHA-NL was seen as a barrier to discussing vulnerability. Support and motivation for using ALPHA-NL was lacking. It was perceived as intrusive and could disrupt the trusting relationship with the woman. ALPHA-NL was not seen as adding value to standard care because women hardly ever completed the ALPHA-NL and the ALPHA-NL was not suitable for women with language barriers or with low health literacy. In addition, they felt the ALPHA-NL training did not contribute sufficiently to learning good conversation techniques to discuss psychosocial aspects and sensitive topics (such as domestic violence, financial problems and addiction problems) with women.

“The midwifery practices that implemented the ALPHA-NL found it quite an intrusive questionnaire., with probing questions. They did find it difficult.” (MW2G).

Agreements on the transfer and exchange of information about follow-up between professionals were included in the program. However, these agreements were often not described in sufficient detail and were not always followed up due to lack of time and discontinuity of caregivers.

Determinants of the user

Thirty respondents found it important to identify and refer women in vulnerable situations to appropriate care. They expected that the program could help to achieve this goal. However, the interview study showed that participants did not feel an urgency to use the program. They also thought that the program and the care pathways were unknown to a lot of colleagues and were hardly ever used. Some participants indicated that they were unfamiliar with the program. The survey study showed that 29 respondents experienced support for the program within their practice or organization, but participants pointed out that professionals did not remind each other of the program.

Respondents of the survey study experienced varying degree of knowledge and skills to put the program into practice. Newly employed midwives and obstetricians were not offered training. Nineteen respondents felt that using the components in the program is part of their job, but this did not apply to the ALPHA-NL. Participants indicated that other tools and approaches were used in practice to identify vulnerabilities.

“Women get a questionnaire before the first consultation, so we already have a little bit of an impression of an obstetric and psychosocial history. For example, their educational level or the use of psychological care in the past. During the first consultation we don't really have standard questions. I don't like that [standard questions] about the ALPHA-NL list either. I prefer to have a conversation. (MW3G).

Determinants of the organization

Aspects such as time, staffing, collaboration, resources and coordination scored mostly neutral in the survey study. Collaboration and contact between YHC nurses, and midwives and obstetricians take time and effort.

“I sometimes spend 20 min on the phone trying to arrange things, it takes a lot of time and perseverance.” (OB1G).

Reasons included varying work schedules and part-time work, as well as willingness to invest time in collaboration. In addition, there was a lack of systems for secure digital data exchange between collaborating healthcare providers. Overall, there was a need for a stronger role of YHC in this collaboration.

Fourteen respondents confirmed that changes within the organization other than the implementation of the program could hinder the use of the program. This was confirmed in the interviews, where healthcare providers mentioned that less priority was given to the program due to organizational changes, staff changes and many other tasks.

Determinants of the socio-political context

All respondents of the survey study indicated that the program was in line with current laws and regulations, or that they had no opinion on this. Interview participants also had no comments on this.

South Limburg

Determinants of the innovation

Participants found the program relevant (n = 14), clear (n = 13), not complicated (n = 13) and complete (n = 12). This was in line with the findings of the interview study. However, YHC nurses working in the community reported in the interview survey that they had heard about the program but were not familiar with it.

All recognized that the definition of vulnerability provides a good framework for care and contributes to a common language between professionals.

But it is important to record it uniformly in the medical file. The definition helps with that. (MW8SL).

Midwives cited the updated questionnaire for psychosocial and medical history and collaboration with the YHC nurse as success factors of the program.

…that we can call [coordinating YHC nurse] in our region and that you can discuss with her. That was not the case before and that is very helpful and an improvement. (MW4SL).

YHC nurses perceived the work of the coordinating colleagues and their central role in referral as positive. Midwives reported the added value of Positive Health as a conversation tool and to promote women's self-direction. However, they questioned the use of Positive Health with all pregnant women and its application twice during pregnancy. They suggested the use of Positive Health on indication.

YHC nurses reported that they used other interview tools to assess women's resilience and therefore did not see an added value in Positive Health to their daily work.

Determinants of the user

Respondents to the questionnaire and participants in the interview reported believing in the program, experiencing good effects and being intrinsically motivated to work with the program. They agreed that the program contributes to better identification and referral of women in vulnerable situations and considered the components of the program to be part of their task. The results of the survey showed that respondents have sufficient knowledge and skills (n = 13) and self-efficacy (n = 13) to put the program into practice. Still, results from the interviews showed some barriers. First, midwives and YHC nurses struggled with the content and delineation of their professional responsibilities and roles. YHC nurses who participated in the interview study had different opinions about the coordinating role of YHC in the social domain. Some believed that this was not the responsibility of the YHC nurse, but of social workers in the social domain. Second, midwives mentioned the need for theoretical knowledge about the content and impact of vulnerability on mother and child, conversation skills on 'difficult topics' and how to deal with different barriers of the women.

According to participants sustainable implementation of the program can be facilitated by making personal training and improving competencies to work with women in vulnerable situations a part of the program.

Determinants of the organization

Midwives and YHC nurses mentioned the collaboration as positive. According to the participants, the interprofessional training for midwives and YHC nurses contributed to good collaboration and getting to know each other.

I think the interprofessional training Positive Health has been of great value. If you know each other, the lines of communication become shorter. (YHCN3SL).

YHC nurses mentioned positive collaboration with some local administrations and services. However, waiting lists, lack of ownership, insufficient and unstructured feedback after referral made collaboration difficult.

Results of the survey study showed that only two respondents felt that there was sufficient time to work according to the program. In addition, 14 professionals reported that there were other changes within the organization that hindered the implementation and execution of the program. This was confirmed by the participants in the interview study, who added the lack of financial resources to properly implement care for vulnerable pregnant women.

Past experiences—such as budget cuts—and the current high workload made that YHC nurses worried about the future. To promote acceptance among women and their families, midwives and YHC nurses felt that the collaboration should be profiled as "normal" care and as teamwork with attention for everyone's expertise. A last-mentioned barrier to program implementation was the use of different patient record software between medical and social professionals. This affected optimal collaboration and transfer of important data.

Determinants of the socio-political context

In the socio-political context, respondents indicated that the program is in line with current laws and regulations or they had no opinion on this. Hindering factors regarding the program mainly related to characteristics of the organization.

Lessons learned

Synthesizing the results of the survey and interview study of Groningen and South Limburg, we can identify some lessons learned regarding development and implementation of interprofessional collaboration between health and social care.

Maternity care professionals find that identifying and referring women in vulnerable situations for tailored care is part of their job, and it is satisfying when the collaboration to achieve this goes well.

Interprofessional working agreements seem to contribute to better identification of pregnant women in vulnerable situations and interprofessional cooperation between maternity and social care. The prerequisite is that the professional roles and responsibilities are clearly described.

To identify vulnerabilities, a broad conversational tool about health and transition to parenthood that fosters trustful relationships with women is preferred. A tool that connects to midwives’ daily practice promotes implementation.

The importance and usefulness of an unambiguous framework and clear definition for different degrees of vulnerability is confirmed by professionals and promotes collaboration and shared language.

Professionals value low-threshold contact for consultation and referral to the social domain through a single point of contact (by phone or face-to-face), achieving efficiency.

Preconditions such as interprofessional training, time and financial resources are needed to enhance implementation.

Discussion

In this mixed-methods study we evaluated the implementation of two programs to identify pregnant women in vulnerable situations and to refer these women when needed to social care for tailored support and care. The results of this study show that both programs have been disseminated and adopted by the professionals. However, not all parts of the programs are used as intended and agreed in the manual. Research participants in both regions are positive about the programs and the potential benefits for women in vulnerable situations and their (unborn) child, but they also mention areas for improvement for further development, implementation and continuation.

In our study, more professionals in Groningen did not know the program and thus did not use it than in South Limburg. Reason for this may be that in Groningen the program was already developed and communicated to professionals in 2017. After that, attention to implementation and evaluation of the program was not a priority. New professionals were not trained to work according to the program. It is well-known that the extent to which implementation of innovations occurs depends of the implementation strategy used [33]. A multi-pronged approach to implementing innovations, targeting specific groups of professionals, seem to be more successful [34, 35].

In Groningen, the ALPHA-NL identification tool was not supported and implemented by midwives and obstetricians, while midwives in South Limburg were very enthusiastic about the (digital) medical and psychosocial history questionnaire. The ALPHA-NL was seen as an intrusive identification tool with probing questions about psychosocial topics. Women hardly completed the ALPHA-NL, while midwives in South Limburg experienced that women did complete the (digital) medical and psychosocial history questionnaire. This can have several reasons. First, we know that determinants such as the questionnaire design (i.e. question wording and question context) and the situation (i.e. online, offline, self-administered, interview-administered) appear to influence the willingness to fill out a questionnaire and the degree of social desirability bias in sensitive questionnaires [36]. Both identification tools are self-administered and can be completed at home. A Dutch study showed that when a woman understands why the maternity care professional asks about non-medical risk factors of vulnerability (question context), the woman is less likely to mind being asked these questions [37]. This suggests that designing a questionnaire for sensitive topics calls for appropriate strategies that provide more valid data and reduce respondents' discomfort in answering a sensitive question.

The success of the identification tool in South Limburg can be explained by the fact that it is not a complicated innovation that requires a lot of adaptation and training, but an update of an already existing practice. Midwives experienced that women have no problem filling out the questionnaire and that the new working method is time-saving.

Maternity care professionals in Groningen indicate a preference for a broader conversation with women about their health and psychosocial vulnerabilities rather than or in addition to an identification tool. This is confirmed in a Dutch study among primary care midwives, which shows a need for a good conversation with the pregnant woman and to trust upon their own intuition during this conversation [37]. The study by Fontein-Kuipers and Mestdagh [38] among migrant women also shows that women can self-assess their vulnerabilities. However, to discuss these vulnerabilities and their impact on the pregnancy requires a trusting relationship between the woman and the maternity care professional. This is in line with the results of a previous study of Feijen-de Jong et al. [39] that a positive attitude of professionals to women in vulnerable situations may enhance implementation of interventions and may contribute to relation-centered care.

In South Limburg, the definition of vulnerability in pregnant women of van der Meer et al. (2019) has been integrated into the program. Professionals experienced this as suitable for creating a common framework for collaboration among professionals.

In both regions the close collaboration between maternity care professionals (midwives and obstetricians) and YHC is an important success factor. Despite several organizational barriers personal contact and knowing each other's expertise was valued and professionals expected that good collaboration would lead to better health outcomes. This is in line with the systematic review of Aquino et al. [10]. It states that collaboration is enabled by good communication, mutual respect and support for colleagues, liaison staff roles, co-location and joint working between midwives and health visitors (i.e. qualified nurses or midwives with an additional diploma or degree in specialist community public health nursing) [10]. A Cochrane Review suggests that interventions aimed at improving interprofessional collaboration through practice changes can improve health care processes and outcomes. However, the authors did not draw generalizable conclusions about the key elements of interprofessional collaboration and its effectiveness [11].

Sustainable implementation of social maternity care is challenging, as the content is still under development and the domains of social and maternity care are organized separately, with different organizations, cultures and funding structures. National steering on the transition to social maternity care is lacking and stakeholders seek central guidance, resources and expertise to support implementation [16, 40]. A Dutch study exploring the availability and use of interventions to be used in care for vulnerable pregnant women identified 40 (interprofessional) interventions that are not or only partially used by professionals, so their effectiveness especially in the long-term remains unknown [30]. A recent systematic review also shows limited evidence regarding the effectiveness and performance of conversational psychosocial assessment tools [41]. Women and professionals reported high acceptability of these tools, although barriers to implementation influence its application, such as lack of time and competences as well as discomfort (professionals), negative experiences, lack of preparation and lack of relevance (women). To implement tools widely and sustainably, more research is required focusing on tool performance, effectiveness and referral to supportive interventions.

This illustrates that implementing innovations needs attention. Several theories on implementation of innovations acknowledge change as a stepwise process that individuals, groups and organizations must go through to achieve desired practice [42]. Transition to the next step is affected by different determinants that must be identified to design implementation strategies tailored to the most prominent barriers for change. Therefore, sustainable implementation of the programs in Groningen and South Limburg requires solid implementation strategies, based on the knowledge gained in this study. Solutions for the encountered problems need to be developed with relevant target groups, applied, and evaluated in terms of implementation theory and in terms of impact for mother and child.

Strengths and limitations of the study

The mixed-methods design is a strength of our study and offered us more in-depth evaluation of the level of implementation and professionals' perceptions regarding the programs. The questionnaire survey provided general results on the MIDI-determinants. The interviews offered deeper and nuancing insights about what did and did not work and how we could improve the programs.

A limitation of the study is that it was not possible to calculate a response rate in Groningen, because the questionnaire was widely distributed among professionals in maternity and social domains. As a result, we also do not have information on non-responders. The survey study in South Limburg took place, within a year of the implementation of the program, among participating primary care midwifery practices and YHC nurses. These professionals can be considered the early adopters in the implementation of the program and extra motivated to optimize care to pregnant women in vulnerable situations. Thus, we have no insight into the perceptions of midwives from non-participating midwifery practices.

In conclusion, in two regions with a high prevalence of Western pregnant women living in poverty, two programs for social maternity care for pregnant women in vulnerable situations have been implemented. Professionals in both regions are generally positive about the programs implemented. A general, conversational way to identify vulnerabilities and a low-threshold, personal contact for referral and consultation to the social domain seem to fit the needs of maternity care professionals.

Areas for improvement for the implementation of interprofessional collaboration between maternity care and social care focus mainly on determinants of the organization, which should be addressed both regionally and nationally. In addition, sustainable implementation requires continuous awareness of influencing factors and a process of evaluation, adaptation and support of the target group.