Decision-making on whether to use assisted vaginal birth is complex, occurring in emergency settings influenced by preferences, knowledge, and experiences of health workers, women, and the health care environment. Health workers may not have the knowledge and skills to decide to or to conduct assisted vaginal birth [3]. Where health workers have been trained, they may hesitate to perform assisted vaginal birth, due to the perceived complexity of assisted vaginal birth, fears of failed attempts, litigation following poor outcomes, or HIV transmission [3]. Addressing these challenges requires improving health workers’ knowledge and skills, task clarity, mentoring, task sharing, and ensuring resources and equipment availability [3]. Moreover, unsupportive work environments can hinder use, especially when clinical protocols for transitioning from assisted vaginal birth to cesarean birth are unclear [3]. Thus, clear protocols, supportive supervision, and local ownership over assisted vaginal births initiatives are critical [3].

Women’s perceptions and experiences are also essential to consider [23]. Women may have fears about assisted vaginal birth due to pain relief absence, exclusion from decision-making, and lack of partner or family support [3]. Negative interactions with health workers can also hinder use, often stemming from power imbalances, poor communication, and mistrust of health systems [3]. Therefore, ensuring pain relief availability and fostering positive communication between health workers, women, and partners or families are critical. Figure 1 demonstrates the complexity of assisted vaginal birth uptake, emphasizing the importance of examining various influencers from different stakeholders before implementing any intervention.

Interventions proposed to optimize use of assisted vaginal birth are often part of broader interventions to improve labor management [3], an approach also encouraged by the WHO [13]. These may include education and training, appointing opinion leaders, providing equipment, implementing guidelines, task shifting and sharing, and audit and feedback [3]. The education and training should cover both assisted vaginal birth and broader physiological labor and birth management [3]. To date, most interventions have focused on addressing health facility and health worker factors, with little consideration of women, partners, or community views. WHO has emphasized the need to understand women’s knowledge, belief, and fears about assisted vaginal birth. Information sharing with women and joint decision-making are paramount to women’s involvement, and exploring how best to achieve this is crucial [13,23]. Figure 2 depicts the potential interventions that may be helpful to optimize use of assisted vaginal birth [3,13] In this study protocol, we provide options to explore which interventions may be suitable to address the identified facilitators and barriers in any given setting. We propose using behavior change frameworks to map barriers and facilitators identified in the formative research, which will help identify potential strategies.

Growing evidence indicates that understanding and addressing individual and organizational behaviors are critical to improve implementation of evidence-based practices [25,26]. One model that captures the complexity of different behaviors is the COM-B (Capability, Opportunity, and Motivation) model of behavior change [26]. COM-B posits that individuals require capability, opportunity, and motivation to perform a desired behavior [26]. Capability points to individual capacity, including knowledge and skills, which influences the desired behavior [26]. Motivation includes drivers of the desired behavior, such as goals, emotions, beliefs, and impulses [26]. Opportunity refers to external factors outside of the individual that may influence the desired behavior [26]. Opportunity includes existing policies, communities’ perceptions, and environmental structures [26]. Both an individual’s capability and opportunity influence their motivation, which in turn affects their intended behavior [26]. Similarly, the intended behavior then impacts their overall capability, motivation, and opportunity [26]. The Theoretical Domains Framework (TDF) expands the COM-B model to categorize the barriers and facilitators to change [27]. The barriers and facilitators under the COM-B and TDF domains can be mapped to the Behavior Change Wheel to identify potential interventions that address barriers and reinforce facilitators [26]. The Behavior Change Wheel comprises 9 types of interventions and 7 policies that can be implemented to promote behavior change [26].

In this protocol, we propose the use of the COM-B and TDF frameworks to guide the study design and instruments to identify barriers and facilitators influencing the optimization of assisted vaginal birth. The “desired behavior” is defined as appropriate use of assisted vaginal birth. In addition to these frameworks, this protocol development was guided by the potential interventions identified by Torloni et al [3], the ecological framework in Figure 1, and the WHO technical brief on optimizing use of assisted vaginal birth [13].

The following participants are proposed for the primary qualitative research to better understand the care context:

We propose to talk to women who gave birth in the past 6 months—both who have and have not experienced assisted vaginal birth in the past 6 months—and women’s representatives from women’s organizations and agencies. A 6-month postpartum time frame is set to ensure that women have sufficient recovery time while still recalling their birth experience. There will be no exclusion criteria imposed, such as clinical complications, mode of birth, intrapartum stillbirth, and so on. In settings where some degree of assisted vaginal birth is practiced, we suggest conducting IDIs with women who have assisted vaginal birth in the past 6 months, FGDs with women who have given birth in the past 6 months, and IDIs with women’s representatives. However, we acknowledge that it may be challenging to find women who underwent assisted vaginal birth in settings with limited or no use of the procedure. In this type of setting, we propose to conduct only FGDs with women who have given birth in the past 6 months and IDIs with representatives from women’s organizations and agencies.

Other than women, FGDs will also be conducted with women’s partners, family members, and community members. The selection of community member types will depend on local contexts. Partners, family members, and community members will participate in combined FGDs; however, to account for gender dynamics, we propose to conduct separate sessions for men and women. The study will also involve IDIs with health workers and facility administrators at health facilities. It may also be useful to conduct IDIs with health professional associations where appropriate, and decision should be made within the research team. Finally, IDIs will be conducted with policy makers at the ministry of health or local health offices. Relevant policy makers will be those overseeing maternal and newborn health and have experience in health policy formulation around labor management, guideline development, or resource supply and distribution. We will invite participants aged 18-49 years to participate. However, depending on the context, it may be appropriate to invite women younger than 18 years, such as in settings where adolescent pregnancy is high or young pregnant women are considered emancipated [24].

Maximum variation sampling will be used [32]. Maximum variation sample encourages recruitment and sampling based on diversity and uses specified parameters to stratify the sample [24,32]. Including individuals from diverse backgrounds is critical to ensure that research findings accurately reflect the perspectives of local stakeholders. Adapting a previously developed generic formative research protocol for cesarean birth that has demonstrated feasibility and usefulness [24], we have developed a sampling grid for this study (Table 1). The sampling grid will guide the recruitment, sampling approaches, and number of participants to be recruited. However, data collection should continue until data saturation is reached [33,34]. Data saturation will be determined when no new codes or themes emerge in at least 2 consecutive interviews or focus groups. The determination of saturation will be made through regular debriefings by the qualitative lead and the research team, ensuring consistency in interpretation.

The research team should ensure diverse participant representation, such as women from urban and rural areas, with different parity, modes of birth, various ethnicities, religions, and ages. Depending on the scope of planned implementation, it may not be necessary to conduct IDIs and FGDs with women in all study facilities. If the number of facilities involved in the planned implementation is small (eg, less than 5), conducting IDIs and FGDs with women in all facilities may be appropriate. However, if the number of health facilities is large (eg, more than 5), selecting some facilities may be more feasible and appropriate. Suppose a subsample of health facilities is chosen. In that case, the selected health facilities should reflect diversity based on geographical region, urban or rural settings, level of care (eg, tertiary, secondary, and primary), and number of annual births.

IDIs with health workers should be conducted at all health facilities involved in the study, as the health facility–specific barriers and enablers may vary. Facility administrators will be sampled based on their roles and, at a minimum, should include the head of obstetrics and the matron-in-charge of maternity. A diverse group of health workers will be recruited based on gender, age, and working years. In particular, the research team will ensure a balanced representation of gender among the health workers who participate in the study. Depending on the scope of planned implementation, policy makers should be sampled from both the ministry of health and the regional or district health offices.

Recruitment posters in local languages will be displayed at the health facility’s waiting area, consultation rooms, community halls, public health centers, and health offices to recruit potential women, community members, health workers, facility administrators, and policy makers. The posters will contain information about the study, eligibility criteria, and how to join the study. Depending on the settings, the study team may use the phrase “women and other pregnant people” in recruitment posters to ensure inclusivity.

Additionally, the study team will approach women who gave birth in the past 6 months during their postpartum or vaccination visits, with 1 non–health facility staff member present on-site to engage potential participants. This helps prevent power imbalances between health workers and women, which could lead to feelings of coercion in participation. Using facility records, women who indicated receiving assisted vaginal birth in the past 6 months will also be contacted to invite them to participate in the study. Conversely, the research team will approach women’s representatives, partners, family members, community members, and policy makers using snowball sampling through recommendations from gatekeepers, networks, and participating women. Health workers and administrators will be recruited at the health facility, with the research team reaching out during department meetings or via telephone or email using staff records. However, the recruitment strategy for health workers should be discussed with the maternity ward heads to decide on the best approach.

Potential participants interested in participating will be given a plain language statement (Multimedia Appendix 5), informed consent form (Multimedia Appendix 6), and time to read and decide to participate. After a few days, research team members will contact the potential participants to confirm their interest in participating and arrange an IDI or FGD at a convenient time and location. The location and time will be scheduled based on participants’ convenience and, ideally, IDIs and FGDs with women will be conducted after 4-6 weeks postpartum. The potential participants will be asked to bring the signed informed consent form to the location of the interviews. IDI and FGD guides can be found in Multimedia Appendices 7-9.

Before the IDIs and FGDs commence, participants will be informed again about the study in more detail through a plain language statement (Multimedia Appendix 5). If they agree, they will be asked to provide consent (written or verbal, depending on the context; Multimedia Appendix 6). Trained research team members with social science or qualitative research backgrounds, and who are not affiliated with participants’ care or workplace, will facilitate the IDIs and FGDs. The gender and age of facilitators will be considered to help participants feel comfortable; for example, FGDs with women will be facilitated by a female researcher whenever possible. FGDs will be stratified by participant type (eg, women-only groups and health workers–only groups) to reduce hierarchical dynamics. Facilitators will use strategies such as turn-taking, inviting quieter participants to speak, and cofacilitation to support inclusive discussion and reduce dominance by more vocal individuals. We propose to conduct IDIs and FGDs in person; however, if not possible, the research team may opt to conduct them on the web. IDIs and FGDs will be approximately 45-90 minutes long and audio recorded. We suggest including 5-10 women per FGD. Field notes will be recorded during data collection and used during analysis to complement transcripts.

In cases where participants are unaware of assisted vaginal birth, facilitators will use comics and frequently asked questions documents to introduce women to the procedure (Multimedia Appendices 3 and 4). Participants will be reimbursed for their time and contribution, if possible, with the amount and form determined by local customs to avoid incentivizing participation. Once IDIs and FGDs are completed, participants will not be followed up unless member checking is conducted. Member checking involves returning to participants to confirm whether their perspectives have been accurately captured during transcription and analysis [35].

All research team members should undergo training with practice sessions before data collection. This includes principal investigators, social scientists, research coordinators, research assistants, transcribers, translators, admin, and other team members. The training workshop will cover project background, research designs, study tools, ethical considerations, project implementation plan, and manuals. Research monitoring will also be conducted throughout the study to ensure that issues can be addressed promptly. Research leads should communicate regularly with the research team members to address any issues promptly during data collection.

Verbatim transcription of audio recordings should occur alongside the data collection, ideally by the interviewer or facilitator, to enhance the trustworthiness of the data. Each transcription will be reviewed for data quality, themes that need further exploration, feedback for future data collection, and data saturation. During transcription, audio recordings will be deidentified by assigning pseudonyms to participants. After transcription, original audio recordings will be deleted to prevent identification. All data will be permanently deleted or destroyed after a specified retention period. Local data protection laws and requirements of study sites and stakeholders (ie, ethics committee) will determine this period. Raw and deidentified data will not be shared externally, although deidentified data may be reused in future research.

Translation may be needed if research teams comprise people with diverse language abilities. The research team will determine the appropriate timing for translation, whether during transcription or after, based on the language used during the interview. The suitability of these approaches will depend on the research team's language skills, cultural considerations, and consideration that linguistic nuances could be lost during translation.

We suggest using a thematic analysis approach for the qualitative research, as described by Braun and Clarke [36]. We recommend starting with an analysis by participant groups to identify variations in perspectives based on their backgrounds and experiences. In the next stage, perspectives from the different groups can be compared to explore similarities and differences based on participants’ backgrounds. We suggest coding data line-by-line and assigning relevant text to specific codes. These emerging codes can be organized into hierarchical groups by grouping related codes into overarching themes and patterns. The emerging themes can then be reported narratively to report the study findings. The team will explore overarching themes across different study sites to site-specific differences that must be considered during implementation. Data analysis will be conducted separately for each participant group, with triangulation used to support in-depth understanding and to assess both convergence and divergence in perspectives. We recommend conducting data analysis simultaneously with data collection to facilitate adaptations and cross-checking, and collaboratively among research team members, incorporating expertise in qualitative research, midwifery, and obstetrics to ensure high-quality analysis. Qualitative analysis software can help manage the analysis process but is optional.

The results from thematic analysis can then be mapped to the behavioral change frameworks, COM-B and TDF. Each barrier and facilitator will be mapped to the 3 COM-B domains and 14 TDF domains to diagnose the core domain driving the behaviors. Based on the domain mapping diagnoses, the Behavioral Change Wheel linked with COM-B will be used to identify the most appropriate intervention to address this. For example, if the barriers were related to reflective motivation, training and enablement can be implemented as interventions to promote the behavior [26]. For an example of how to conduct this type of analysis, see the results from a qualitative formative study to inform intervention development for postpartum hemorrhage [37,38].

Data analysis workshops may be held to facilitate analysis, interpret findings within the local context, refine effective interventions, and discuss practical implementation strategies. The workshop findings will help inform the intervention’s development and implementation considerations. The research team will facilitate the workshop, involving all relevant stakeholders, such as health facility staff, community members, and women. The research findings can then be disseminated to community leaders and members through meetings, presentations, posters, and social media.