Narratives on why pregnant women delay seeking maternal health care during delivery and obstetric complications in rural Ghana

Background: Despite the many maternal healthcare policy programmes in Ghana such as free the antenatal care (ANC) and the fee-exemption policy under the National Health Insurance Scheme, among others, the country has yet to make substantial improvements in addressing low skilled care utilisation in pregnancy and delivery. From previous studies, maternal mortality has been linked to women’s healthcare decision-making power at the household level in many low and middle-income countries. Thus, a pregnant women’s ability to choose a healthcare provider, act on her preferences, and to be sufficiently financially empowered to take the lead in deciding on reproductive and pregnancy care has significant effects on service utilisation outcomes. Therefore, we explored rural community-level barriers to seeking care related to obstetric complications and delivery from the perspectives of mothers, youth, opinion leaders and healthcare providers in Nadowli-Kaleo and Daffiama-Bussie-Issa districts in the Upper West Region of Ghana. Methods: This exploratory qualitative study was based on the narratives of women, health providers and community stakeholders regarding the expectant women’s autonomy to decide and utilise maternal care. To achieve maximal diversity of responses, purposive sampling procedures were followed in selecting 16 health professionals, three traditional birth attendants and 240 community members (opinion leaders, youth and non-pregnant women) who participated in individual depth interviews and focus group discussions. Results: Women’s lack of autonomy to seek care without prior permission, perceived quality care of traditional birth attendants, stigmatisation of unplanned pregnancies and cultural beliefs associated with late disclosure of childbirth labour all delayed mothers timely use of skilled care in the study communities. These barriers compounded problems arising from communities that are geographically isolated from hospital care. Conclusions: Decisions about seeking maternal care were usually made by the expectant woman’s husband and family without providing adequate support to pregnant women during the latter stages of pregnancy and delivery. We conclude that this is primarily a cultural issue. The study recommends a change in the approach to community-level health education campaigns for maximum impacts through the increased involvement of men and families in health service delivery and utilisation. Nadowli-Kaleo district (NKD) and Daffiama-Bussie-Issa district (DBID) in the Upper West Region of Ghana are the study locations. According to the 2010 Population and Housing Census in Ghana, the two districts have a population of 61561 and 32827 respectively [18, 19], which account for 13.5% of the region’s population. The female to male ratio in DBID is 94.7 and 87.6 in NKD [18, 19]. However, there were more literate males than females aged 11 years old and more; 53.0% versus 47.0% in NKD and 48.2% versus 37.0% for DBID. The majority of the people are youth, and the proportion of males to females with secondary level education or higher was respectively 5 and 2.8%. Overall, six out of ten people in both districts could read and write a language [18, 19]. Economically, the districts are very deprived and dominated by subsistence farming, with more than 80% of the population not having formal sector employment [18, 19]. Subsistence farming accounts for over 70% of the labour force, and sorghum, millet, maize, groundnut and beans are the leading food crops grown. Also, nearly all families engaged in animal rearing and poultry farming with cattle, goats, sheep, pigs and poultry as the most common ones in the area [18, 19]. Women generally have little influence in decision-making pertaining to their health in these communities, particularly those who are uneducated and less educated [7, 12, 20–33]. There were cases of teenage pregnancies, which contributed to poor pregnancy outcomes and obstetric complications in most pockets of these districts [4]. This was a qualitative exploratory study that the authors conducted over 10 months from February to May 2016 and from January to May 2017. The study was conducted in eight communities in the two districts. The two districts in 2015 were made up of 16 sub-district health structures [Health Centres and community-based health and planning service (CHPS) compound or clinic], five in Daffiama-Bussie-Issa district and 11 in Nadowli-Kaleo district. Purposive sampling criteria was used in selecting communities within the districts to maximise the diversity of experiences. First, eight sub-districts were chosen at random from the list of all 16 sub-districts. From the eight, we identified communities with a Health Centre and those without a Health Centre or with CHPS compounds/clinics. At this stage, one community from each sub-district was purposively selected, with equal consideration for those served by a Health Centre facility and those with a CHPS facility or clinic. This procedure led to the selection of eight study communities – four in each study district. A total of 259 participants (240 FGD participants, 16 healthcare providers and 3 traditional birth attendants) took part in the research. Focus groups comprised non-pregnant women who had childbirth experiences, youth leaders (18–35 years of age) and community opinion leaders. The ages of participants ranged from 20 to 69 years. Apart from the healthcare providers, most of the participants had no formal education. From the eight study communities, we used purposive sampling technique to choose participants who were core members of community development and maternal care programmes in the various community levels [12, 33]. Purposive sampling was used to help select stakeholders in maternal and newborn health, that will provide relevant information to address the research aims. For the focus group cohort, the research team worked with local community leaders who gave permission to access their communities and served as informants and helped in recruiting research participants [33]. Two key recruitment techniques were adopted to recruit research participants. First, the research team repeatedly advertised study aims during the village market days using the “town criers” combined with “standing on strategic corners” of the communities. The research team carried out participant selection. A pre-determined inclusion criteria of participants’ willingness and availability, and being able to consent to participation personally, was used in composing the groups. The participants were approached directly by the research team, who explained the rationale of the study, established their eligibility and obtained their voluntary consent before they were included in focus groups. Twenty-four (24) units of FGDs and 22 individual depth interviews were completed. For the composition and participants, three focus groups of ten members in each group, were constituted for each community: a women’s group (n = 80), a youth group (n = 80; men = 40, women = 40), and opinion leaders’ group (n = 80; men = 58, women = 22). Therefore, a total of twenty-four (24) units of FGDs were conducted. The groups’ composition were partly informed by the literature and the need to obtain women’s experiences by providing them with the freedom to express their views in women’s only groups. For the healthcare staff cohort, recruitment through “door knocking” was complemented with the provision of information sheets highlighting the study objectives randomly distributed in the various health facilities. Sixteen (16) healthcare professionals working at the Ghana Health Service and involved in maternal service delivery also participated, comprising a medical director, pharmacist, head of the maternity department of the hospital, and eight heads (midwives) of sub-district health facilities (health centres and community-based health and planning service compounds) from the eight research sites. Five were professional midwives, while the other three were community health nurses/enrolled nurses. Five other nurses (three nurses plus the two district directors of health services), who were not directly involved in antenatal care services delivery were purposively selected to provide information on their general experiences of maternal and newborn healthcare utilisation. Three traditional birth attendants serving the study communities but residing outside them were also contacted, and those who gave their voluntary consent were included in the research. An open-ended/semi-structured question guide was used in collecting data for the study. The research questions were designed to gather answers related to family and community support to mothers, pregnant women’s freedom to decide on healthcare utilisation, the level of family support provided to the pregnant woman and barriers to access and utilise timely and skilled maternal care. Discussions also explored women’s freedom to use family resources during pregnancy and labour without fear or intimidation and the implications of these factors on skilled maternal care utilisation (see Additional file 1). Individual depth interviews (IDIs) and focus group discussions (FGDs) were the techniques used to collect data from healthcare providers/traditional birth attendants and community members, respectively. JS and MYM received substantial training in different data collection techniques before the study, and we collected all the data, in collaboration with JC, SW and JY. The multiple sources of data using different techniques allowed for triangulation of the perspectives. Audio-tape recordings were clustered responses/quotes from FGD units. Each interview was carried out through face-to-face interaction, providing the researcher with the opportunity to clarify vague or unclear responses, and to pick up on social cues. Interviews with health professionals were completed in English while focus group discussions were conducted in the local language (“Dagaare”). FGDs sessions were held for the separate categories of participants to avoid intimidation and cultural nuances that may impact on discussants’ freedom to express opinions concerning the culture and household set-up in the Upper West Region of Ghana. The results of the focus group sessions and individual depth interviews pertaining to the research question are reported in this paper. The team had a minimum of a postgraduate degree in public health, and two researchers (JS and MYM) are proficient in the local language, “Dagaare”. JS and MYM received extensive training on ethics in research, data integrity and confidentiality issues of participants and had a firm grounding in interview questioning and data management before the study. All authors are experts in qualitative research. The research team were in regular communication to ensure the research was conducted as designed. Play-back approach of audio recordings to interview and focus group participants was used to confirm their submissions before leaving the research community. Reviews of any adverse impacts such as participant recruitment and selection criteria and those that declined were noted, discussed and adjusted during the study. During the data collection, the research team ensured data relevance and integrity processes were observed. All data were securely stored in a password-protected computer to ensure data privacy. Data processing and analysis involved prolonged engagement with the data by all authors. Reliability of the findings was assessed by two external experts who are familiar with qualitative research. After the data collection, audio recordings of the FGDs were transcribed verbatim in Dagaare and translated to English and validated by language specialists from the Ghana Insitute of Languages to ensure rigour. However, interviews with healthcare providers were transcribed directly in English. The hand-written transcripts were typed and imported into NVivo version 11. Transcripts were studied to identify patterns which emerged as broad topics. Content analysis was done electronically and manually to index themes and factors and direct quotes selected to illustrate the factors and broader themes. Processing of each dataset was handled separately. Patterns within the responses in the transcripts was analysed deductively by coding concepts and main ideas. The research team conducted the coding independently and reconciled any differences that emerged. Data analysis was an iterative process and coding was carried out individually among all researchers, and any differences reconciled appropriately. The study received ethics approval from Charles Sturt Human Research Ethics Committee [Protocol numbers: {“type”:”entrez-nucleotide”,”attrs”:{“text”:”H16013″,”term_id”:”880833″,”term_text”:”H16013″}}H16013 and {“type”:”entrez-nucleotide”,”attrs”:{“text”:”H16178″,”term_id”:”880998″,”term_text”:”H16178″}}H16178] and Navrongo Health Research Centre [Protocol number: NHRCIRB345]. The Regional Health Directorate of Upper West Region and the two districts also gave written support for the conduct of the study. Written informed consent (signed or thumbprint) was obtained from all participants, and participation was entirely voluntary.