Background: Malaria in pregnancy (MiP) is an important public health problem across sub-Saharan Africa. The package of measures for its control in Ghana in the last 20 years include regular use of long-lasting insecticide-treated bed nets (LLINs), directly-observed administration (DOT) of intermittent preventive treatment with sulfadoxine-pyrimethamine (IPTp-SP) and prompt and effective case management of MiP. Unfortunately, Ghana like other sub-Saharan African countries did not achieve the reset Abuja targets of 100% of pregnant women having access to IPTp and 100% using LLINs by 2015. Methods: This ethnographic study explored how healthcare managers dealt with existing MiP policy implementation challenges and the consequences on IPTp-SP uptake and access to maternal healthcare. The study collected date using non-participant observations, conversations, in-depth interviews and case studies in eight health facilities and 12 communities for 12 months in two Administrative regions in Ghana. Results: Healthcare managers addressed frequent stock-outs of malaria programme drugs and supplies from the National Malaria Control Programme and delayed reimbursement from the NHIS, by instituting co-payment, rationing and prescribing drugs for women to buy from private pharmacies. This ensured that facilities had funds to pay creditors, purchase drugs and supplies for health service delivery. However, it affected their ability to enforce DOT and to monitor adherence to treatment. Women who could afford maternal healthcare and MiP services and those who had previously benefitted from such services were happy to access uninterrupted services. Women who could not maternal healthcare services resorted to visiting other sources of health care, delaying ANC and skipping scheduled ANC visits. Consequently, some clients did not receive the recommended 5 + doses of SP, others did not obtain LLINs early and some did not obtain treatment for MiP. Healthcare providers felt frustrated whenever they could not provide comprehensive care to women who could not afford comprehensive maternal and MiP care. Conclusion: For Ghana to achieve her goal of controlling MiP, the Ministry of Health and other supporting institutions need to ensure prompt reimbursement of funds, regular supply of programme drugs and medical supplies to public, faith-based and private health facilities.
The study design was ethnographic. It employed non-participant observations, conversations2 and in depth interviews (IDIs) to obtain data from healthcare managers, health workers, NHIS personnel, pregnant women and a cross section of community members. The study team collected data from April 2018 to March 2019, while follow-up interviews and conversations were conducted in September 2019. The choice of ethnography enabled the study to achieve its objective of exploring behavioural factors influencing access to IPTp-SP from different perspectives such as healthcare managers, health providers and pregnant women. Using ethnography to explore specific, as well as complex phenomena in the health care environment is common and relevant to understanding behavioural factors influencing health care provision and access to health care [68, 69]. Various scholars have conducted ethnographic studies in different locations in hospitals such as in wards, on health care interventions and health care policy issues [70–73]. Such studies have contributed to understanding behavioural factors influencing health care provision and utilization [74]. The research team comprised of a post-doctoral researcher (MA) and eight research assistants (RAs).3 MA trained the RAs and supervised data collection. Each RA was assigned to a facility and a community for data collection. The team conducted non-participant observations intermittently in the 8 facilities and 8 communities. The study was conducted in five districts,4 three in the Ashanti and two in the Volta regions of Ghana. Eight health facilities (five government and three faith-based) and 8 communities were chosen for the study. Details of the selection process are indicated in subsequent paragraphs. Ashanti region was selected to represent the middle belt of the country, while Volta region was selected to represent the southernmost belt of Ghana. The two regions are linguistically different, Twi is spoken in the Ashanti region and Ewe is spoken in the Volta region. Ashanti Region reported the second highest percentage (98.8%) of women receiving ANC care from skilled providers in 2014, while the Volta region reported the second lowest percentage (93.9%) of women receiving ANC from skilled providers [37]. The district hospitals in the five districts qualified automatically to participate in the study. Also, interactions and interviews with pregnant women in some of the study communities revealed that they preferred to visit particular health facilities for ANC services. Three of such facilities, which are faith-based were included in the study. Thus, a total of 8 health facilities were selected for the study. Some women preferred the three facilities (2 in the Ashanti region and 1 in the Volta region), because they were closer to their communities than the district hospitals. The women’s assertion of nearness to facilities was further confirmed by transect walk in all the communities to confirm the location of health facilities. The study team visited the study health facilities and went through ANC case records. The team counted the total number of MiP cases documented in the ANC records for each community that accessed ANC in a study facility, from January 2015 to March 2018. The community with the highest recorded number of pregnant women who were diagnosed of malaria in each facility was chosen for the study. Four study communities in the Ashanti and 4 in the Volta Region were selected for the study (Table 1). Hospital records were used as a criteria to select study communities, because it increased the team’s chances of including pregnant women who had experienced malaria during pregnancy in communities that had a history of MiP case. Two other reasons for using hospital records to guide the selection of study communities were: (1) earlier literature suggested that Ghana had high ANC attendance rate and high IPTp-SP + 1 uptake [36, 37]; (2) health providers in study facilities confirmed that malaria cases among pregnant women had reduced drastically compared to the general population, due to an increase in uptake of IPTp-SP. Study health facilities and communities in the Ashanti and Volta Regions with pseudonyms *Study facilities in the Ashanti region have been given the pseudonyms: ASF01, ASF02, ASF03 and ASF04. Study communities in the Ashanti region have been given pseudonyms: ASC01, ASC02, ASC03, ASC04 #Study facilities in the Volta region have been given the pseudonyms VRF01, VRF02, VRF03, VRF04. Study communities in the Volta Region have been given pseudonyms: VRC01, VRC02, VRC01, VRC01 The research team held a sensitization workshop for the Volta Regional health directorate and the district health management teams (DHMT) of participating districts in the Volta region. The regional health director and the DHMT directors approved the study. A sensitization workshop was not carried out in the Ashanti region, because some of the team members in this current study had conducted community entry activities in the region and in the study districts in earlier studies. However, the team sent letters to all the participating DMHTs in the Ashanti and Volta regions, to seek permission to conduct the study. The DHMTs approved the study by writing letters to that effect. Copies of the letters approving the study were sent to the heads of the study health facilities in each study district and leaders of the study communities. The study team met assembly members, chiefs, queen mothers and elders in each study community, to inform them about the study and to seek permission to conduct the study in their communities. A research assistant was assigned to a health facility to conduct non-participant observation, to conduct conversations and IDIs with health workers and women who were attending ANC. The RAs first immersed themselves in the study facilities and the communities [75]. The RAs took the following steps: familiarized themselves with study facilities and communities; developed trusting relations with potential study participants, identified places that pregnant women liked to visit and carried out transect walk. The approach provided MA and the RAs information on where and how to recruit study participants using convenience and snowball sampling methods. The RAs used convenience sampling to select pregnant women attending ANC. They approached women who were attending ANC and those who confirmed that they were residing in a study community were invited to participate in IDIs. The RA explained the study to the women and those who were interested offered their contact addresses. The RAs later visited them at home to obtain written consent and to conduct IDIs. Pregnant women were also recruited from study communities, using snowball method. The first woman who was approached in the community or at the health facility helped the RA to identify other pregnant women in her community. Convenience sampling was used to select women who participated in conversations. RAs approached women who were visiting ANC and those who were interested in conversing with RAs gave verbal consent. The RAs accompanied them while they accessed different ANC services. The purpose was to learn and to understand women’s experiences in going through the ANC process. Information that the study team gathered from immersing themselves in the study field and maps that were sketched from transect walk in the 8 study communities, guided the team in selecting case studies. Case studies were purposively selected among women who attended ANC every month, those who attended ANC irregularly, those who skipped ANC appointments and those who were being treated for malaria. Healthcare providers (nurses and midwives) who had been working in the ANC for at least 1 year, were purposively selected to participate in the study. The choice of health providers who had worked over a year, was to ensure that health providers who were selected had adequate knowledge on the subject of interest. Additionally, healthcare managers (senior medical officers and physician assistants in-charge of hospitals and health centres and healthcare administrators), ANC managers (senior midwives in-charge of the ANC), and laboratory officials were purposively selected to participate in the study (Table 1). RAs visited case studies at home, to observe whether they used LLINs, whether they attended ANC regularly and whether they took their medications. The RAs conducted IDIs with the women and community members using Ewe language in the Volta region and Twi language in the Ashanti region. The IDIs centred on knowledge, attitudes, beliefs and practices on malaria in pregnancy interventions and socio-cultural practices. The IDIs that were conducted with healthcare providers centred on maternal and MiP service provision, challenges and facilitators. The RAs interviewed or talked to procurement officers and laboratory personnel, to clarify issues on payment processes and stock-outs of drugs and other medical products. ANC unit managers commonly referred to as in-charge and healthcare managers were interviewed, to help understand managerial and administrative practices. The study team talked to some of the officials of the national health insurance scheme (NHIS), to clarify and verify issues raised by healthcare managers such as NHIS policies and delays in reimbursement for some ANC and MiP services (Table 2). English language was used in conducting IDIs and conversations with healthcare workers and NHIS officials, because it is the official language of Ghana (see Additional files 1, 2, 3, 4, 5, and 6). Data collection methods and categories of respondents Observations were conducted in 8 health facilities and 8 communities in the Ashanti and Volta regions for 12 months RAs conducted transect walk in study communities to identify key places, settlement patterns and physical access to health care facilities and other sources of health care. The RAs sketched maps from the information obtained from the transect walk, which guided the study in choosing three additional facilities, locating pregnant women and case studies. Interviews were recorded using digital recorders and transcribed verbatim to preserve interviewees’ original messages. Interviews in Ewe and Twi were transcribed into English to enable easy analysis and comparison. This study was conducted as part of a larger study on parasitic infections during pregnancy and community interventions programme in Ghana. The study set out to explore socio-cultural and community factors influencing utilization of MiP interventions in Ghanaian communities. Aspects of the study has been reported elsewhere [14]. However, initial analysis of the data revealed that multiple community and health system factors influenced utilization of MiP interventions. This informed the team to return to the study health facilities to explore and understand health system factors influencing utilization of MiP interventions. Grounded theory approach was used in data analysis [76]. The first set of data (interviews, observation notes and conversations) was triangulated and an iterative approach was used to code the data. MA and ED (ED is a qualitative data analysis expert, who was hired to assist in coding) developed a codebook and used it to code the data with the assistance of qualitative analysis software, Nvivo Version 11. The codes generated by MA and ED were compared, agreed and merged. Key informants such as healthcare managers and health workers were visited for further interviews in pursuance of emerging themes which included the following: (1) why and how managers made certain decisions on maternal healthcare; (2) why ANC providers who claimed they loved their jobs felt frustrated in providing maternal healthcare; and (3) why women did not trust the health care system and were skipping appointments. Coding, analysis and data collection went on until saturation was attained. Saturation was attained when no new information on the major themes were obtained, which is in line with Charmaz’s [76] recommendation. The results of this manuscript are based on major themes that emerged from the analysis. Ethical clearance was obtained from the University of Health and Allied Sciences’ Research Ethics Committee [UHAS-REC/A.I Ul 17-18]. The team obtained written consent from study participants who participated in IDIs and verbal consent from those who participated in conversations.5 To protect informants’ identity, besides actual country and region names, individuals and facilities’ names used in this article are pseudonyms.