Recognizing the importance of the Adequate Childbirth Program intervention to the health of Brazilian women and children, the National School of Public Health (ENSP)—Oswaldo Cruz Foundation (FIOCRUZ) proposed an evaluation titled “Healthy Birth: a prospective study to evaluate the implementation and effects of a multifaceted in-hospital intervention.” In order to analyze the implementation and effects of the PPA in a sample of 12 hospitals utilizing mixed methods analysis [3].
The survey was carried out at two different times for data collection. The first [M1] focused on assessing the degree of implementation and the intervention’s effect; the second moment [M2] focused on assessing the sustainability of the intervention. More details on data collection, contextual aspects, and protocols established by the “Healthy Birth” survey can be found in Torres et al. [3] and Domingues et al. [16].
Study type and subsample
Integrated to this evaluation, this article develops a single case study with an exploratory qualitative approach, which chooses the intervention Projeto Parto Adequado as a case under analysis. For this qualitative analysis, a subsample of 08 (eight) hospitals (Hosp01; Hosp02; Hosp04; Hosp05; Hosp06; Hosp09; Hosp10; Hosp12) was intentionally selected (Fig. 2). The inclusion criteria for this subsample were the location of the maternity hospital according to the country’s macro-region, the type of maternity unit belonging or not to health insurers, and the maternity’s performance in achieving goals to reduce cesarean sections (data provided by ANS/PPA monitoring). We excluded four hospitals (Hosp03; Hosp07; Hosp08; Hosp11) due to geographic location and management [3].
Selection of participating professionals
The technique used to select such participating professionals was Snowball sampling [17], a non-probabilistic sampling that focuses on reference chains, used mainly in exploratory studies. Identifying key informants or seeds for the interviews began with the managers, who identified the project leaders/coordinators in the hospital. We selected 34 professionals directly linked to the management and coordination of hospitals, 10 (ten) managers, and 24 leaders/coordinators. There was no refusal from the participants in any of the hospitals. Managers were considered actors with greater decision-making power, and leaders/coordinators as those who directly support managers in the work decision-making process.
Data collection scenario
The training of the interviewers who carried out the data collection was carried out in two ways: in person for 02 (two) coaches and remotely for another 02 (two) coaches. In these trainings, there was a reading of the instrument, presentation of techniques for conducting interviews, field observations, and the importance of records.
Regarding the characterization of the 04 (four) interviewers who carried out the data collection: they were all women with academic training in Nursing, Midwifery, and History. The historians were researchers in public health; the nurse and the Midwife were researchers in women’s health; all had previous experience with data collection for research. These interviewees were drawn to the field mainly because they were already involved in women’s health research or care practice. They were presented as such to the interviewees.
Before entering the field for data collection, pilot tests of the scripts were carried out in a maternity hospital in Rio de Janeiro/Brazil, part of the Adequate Childbirth Project. However, this maternity hospital was excluded from the sample of hospitals in the Nascer Saudável evaluation. This phase made it possible to make adjustments and validate the interview scripts. Furthermore, the research team established a previous contact connection with the management of the institutions to agree on the dates on which the interviewers would be placed in the field for data collecting and any ethical issues that would be released.
The immersion in the hospitals took place during a period of 05 (five) days, from July to October 2017, in which the research team carried out interviews with managers coordinators/leaders. A semi-structured script was used for managers and leaders/coordinators, which included the following axes: decision process; deployment strategies; participation of the assistance team; women’s participation; monitoring; strategy results. This script was developed for Healthy Birth research [3].
The interviews took place in the hospitals themselves; privacy was maintained in rooms that could have closed doors and away from other people, with only the interviewer and the interviewee remaining in the room. The interview length (average of forty minutes) was not pre-determined and varied according to each interviewee's level of involvement in the project and subjective features. The interviews were audio-recorded. During this process, there was no need to repeat interviews. Field notes were taken at the end of each hospital field.
Subsequently, the interviews were transcribed by an independent professional. Such transcripts were not returned to the participants for evaluation, comments, and corrections. However, for the order of internal validation of the transcripts, there was a review by the research team.
About data saturation, it is worth noting that qualitative research, unlike quantitative research, is not based on how many individuals should be heard but on the intensity of the phenomenon and scope to which the actors are linked to the researched object. This emphasizes the need to fine-tune the interviewee selection process. However, the recommended minimum of at least 20–30 interviews was used for qualitative investigations [18].
Data storage
Work was carried out on constructing and analyzing the database in the MaxQda software [19]. During the database construction stage, the interviews were inserted, organized, and encrypted in the software according to the moment of the evaluative research, the ID of the hospital where the interviewee was linked, the professional category belonging, and finally, a professional identifier number. All citations in this article only include the professional position held by the interviewed leadership, in order to avoid possible identification.
Data analysis
According to Uwe Flick [20], the data were submitted to Thematic Content Analysis. Focusing on the axis of the semi-structured decision-making process, the following dimensions were explored: whether the maternity hospital was already developing some initiative related to childbirth care before the PPA; how maternity got to know the PPA; what motivated the participation of maternity in the PPA; and finally, who decided for maternity to participate in the PPA.
The open categorization was then carried out, generating broad segments. With the refinement of these segments, a list of codes was generated, a step called axial coding. An inductive association was made to create the categories [20]. The entire data analysis course was permeated by the theoretical framework of the Diffusion of Innovation by Everett M. Rogers [15].
Although there was no feedback from the participants regarding the findings, for the interpretation of the data, the interviewees’ speeches and their respective coding were validated by members of the research group, in which there were also reflexive co-participations on the interpretive procedures of the entire analytical phase [20].
As a methodological guide for qualitative research, it used the consolidated criteria for reporting qualitative research (COREQ) [21].