Policymaker, health provider and community perspectives on male involvement during pregnancy in southern Mozambique: A qualitative study

Background: Increasing male involvement during pregnancy is considered an important, but often overlooked intervention for improving maternal health in sub-Saharan Africa. Intervention studies aimed at improving maternal health mostly target mothers hereby ignoring the crucial role their partners play in their ability to access antenatal care (ANC) and to prevent and treat infectious diseases like HIV and malaria. Very little is known about the current level of male involvement and barriers at different levels. This study explores the attitudes and beliefs of health policymakers, health care providers and local communities regarding men’s involvement in maternal health in southern Mozambique. Methods: Ten key informant interviews with stakeholders were carried out to assess their attitudes and perspectives regarding male involvement in programmes addressing maternal health, followed by 11 days of semi structured observations in health care centers. Subsequently 16 focus group discussions were conducted in the community and at provider level, followed by three in depth couple interviews. Analysis was done by applying a socio-ecological systems theory in thematic analysis. Results: Results show a lack of strategy and coherence at policy level to stimulate male involvement in maternal health programmes. Invitation cards for men are used as an isolated intervention in health facilities but these have not lead to the expected success. Providers have a rather passive attitude towards male involvement initiatives. In the community however, male attendance at ANC is considered important and men are willing to take a more participating role. Main barriers are the association of male attendance at ANC with being HIV infected and strong social norms and gender roles. On the one hand men are seen as caretakers of the family by providing money and making the decisions. On the other hand, men supporting their wife by showing interest in their health or sharing household tasks are seen as weak or as a manifestation of HIV seropositivity. Conclusion: A clear strategy at policy level and a multi-level approach is needed. Gender-equitable relationships between men and women should be encouraged in all maternal health interventions and providers should be trained to involve men in ANC. The study was carried out by Ghent University in collaboration with the International Centre for Reproductive Health – Mozambique (ICRH-M) and Universidade Eduardo Mondlane (UEM) between March and October 2017. ICRH-M is a Mozambican Non-Governmental Organization (NGO) and research institution. UEM is the main public university in Mozambique. The study was conducted in Marracuene and Manhica districts in Maputo Province. This region has around 334,000 inhabitants and 21 rural health centers. This study site was involved in previous ICRH-M studies, and therefore health providers and health managers working in this area have a constructive relationship with the principal investigator and ICRH-M researchers. The principal investigator (AG) is a Belgian doctoral student with research experience in Mozambique. She was assisted during all focus group discussions (FGDs) and couple interviews by a research assistant (HC) and two fieldworkers. The research assistant was a final year medical student doing an internship at the reproductive health research unit of the Universidade Eduardo Mondlane. Key informant interviews (KIIs) and observations were conducted by the researcher alone prior to the FGDs. Data collection took place between March and October 2017. Firstly 10 key informant interviews were carried out over a 2 month period, followed by 11 days of semi-structured observations as preparation before the FGDs. Afterwards 16 FGDs were conducted with providers and community members to explore different aspects of male involvement, which were explored further in three in depth couple interviews. FGDs were spread over a period of 5 month period to allow for minimal interim analysis, followed by the in depth interviews. The key informant interviews were conducted with maternal health policymakers, researchers and NGO staff. Men and women specialized in maternal health policies, program implementation and research in Mozambique were eligible for participating in the interviews, which aimed to frame the topic within the political and structural context of Mozambique. Experts working in private clinics or commercial organizations were excluded. Participants were identified using a ‘snowballing’ approach. The first round of contacts was identified by personal contacts of the authors (AG, OD, SG, NO and KR) and reviewing attendee lists of national maternal and child health conferences. In addition interviewed contacts were asked to nominate other appropriate key informants. Key informant interviews all took place face-to-face in a private room. All interviews were recorded and transcribed, except for one interview where only notes were taken (no recording was allowed by the participant). Subsequently, AG conducted sit-in observations at antenatal clinics and attended health promotion sessions at the different study sites to explore the workload of providers, number of men attending ANCs, power dynamics within the consultation and quality of care provided. Seven different study sites were purposively selected to include health centers with different characteristics (high versus low workload, urban versus rural). Both a checklist (see Additional file 1) and written narratives were used to collect the data of the observations. In total 159 antenatal consultations were observed. Afterwards, FGDs were conducted with providers and community members. For the FGDs with providers, the heads of the health centers were contacted to discuss a date and time for conducting the FGDs. All health care centers in Manhica and Marracuene were listed and the study sites were purposively selected to include health centers with different characteristics and good road access. FGDs were planned during lunch break or after working hours. Men and women were mixed in the focus group discussions with providers, as we believed in this group the gender dynamics are of interest and would not affect the openness of the participants. For the FGDs in the community, community leaders were contacted in advance with an invitation letter. Communities with different characteristics were selected (distance to a health facility, rural versus urban, seasonal wave of work or year-round employment, …) for the study. All respondents for the FGDs in the community were purposively selected to represent certain segments of the population—namely, pregnant women or pregnant < 2 years ago, male partners, community leaders, health activists, CHWs, and traditional birth attendants. FGDs in the community were divided into male and female groups, since we believed this composition may make participants more likely to discuss topics openly together than if groups were mixed. FGDs were conducted in a private place, inside the community office or under a tree away from other activities, and the date and time were decided by the community leaders. For both provider and community FGDs, the number of participants per group ranged from 5 to 8. FGDs in the community were conducted in the local language (Changana), while provider FGDs were conducted in Portuguese. All FGDs were facilitated by the researcher, assisted by two local fieldworkers. The researcher is fluent in Portuguese and the fieldworkers were fluent in both Portuguese and the local language. FGDs in the community with men were assisted by a male fieldworker and FGDs in the community with women by a female fieldworker. FGDs were spread over a period of 5 months to allow for minimal interim analysis. Data was collected until data saturation. The interview guides for the key informant interviews and FGDs can be found as additional files (see Additional file 2). Subsequently, themes that emerged during the community FGDs were discussed with three couples, the couples were purposively selected in order to include couples at different stages of their reproductive life. As FGDs within the community were conducted with men and women separately, we wanted to conduct these interviews with men and women together to generate new insights regarding the dynamics within couples. In depth couple interviews were conducted in the local language (Changana). All interviews and focus group discussions were transcribed verbatim in Portuguese, except for the FGDs conducted in the local language, which were translated into Portuguese during transcription. Transcription from the local language into Portuguese was conducted by the research assistant (mother tongue Changana) and an extra interpreter as a double check. Thematic analysis was used as data analysis method and the framework approach was used as a tool, including 7 stages of analysis: transcription, familiarization, coding, developing a working analytical framework, applying the analytical framework, charting data into the framework matrix, interpreting the data [22]. R Qualitative Data Analysis (RQDA) software was used for coding. All data was coded by both AG and HC, afterwards all codes were discussed, and they agreed on a set of codes and categories (first four steps of the framework analysis). After reviewing a number of theories on access to healthcare and health promotion programming [23–27] the socio-ecological framework was identified as the most appropriate model to guide the analysis. Four units of analysis were identified: individual, interpersonal, community and health system related factors (see Fig. 1). The last three steps of the framework approach were supervised by OD. Field notes from observations were also analyzed and added as analytical memos that facilitated interpretation of certain phenomena that emerged during data analysis. All analysis was carried out in Portuguese and some quotations were translated into English by the researcher in the present article. Only age ranges were reported along with quotations to guarantee anonymity. Translations of quotations were double checked by a bilingual colleague (mother tongue Portuguese). Socio-ecological framework with emerging themes in RQDA-plot For the observations all health directors and health providers of the participating health centers were asked for permission after explaining the aim and procedures of the study. The women and their partners (if present) were asked orally if they consented to be observed during consultation before entering the consultation room. The provider explained to them the observer was a Belgian midwife, conducting a study about the Mozambican health system. For the FGDs with providers the district officers were contacted for authorization and organizing the FGDs. FGDs in the community were organized by the community leaders. Information about the objective of the study and procedures was provided to all respondents (KIIs, FGDs and in-depth interviews) orally and in writing. Participants were asked if they consented to interviews being recorded using a tape recorder. Confidentiality, anonymity and ground rules were discussed before starting an interview or FGD. Participation in the study was voluntary and all participants of the key informant interviews, FGDs and in-depth interviews gave their written consent. Participants not capable of signing could provide their fingerprint. No incentive was provided, other than refreshments during FGDs. Ethical approval for the study was obtained by the medical ethical commission of Ghent University (EC/2018/1319), the National Health Bioethics Committee of Mozambique and the Health Bioethics Committee of UEM and HCM (Maputo Central Hospital) (CIBS UEM&HCM/0008–17).