Menu
Menu
Menu
Menu
Background South Africa faces a complex burden of disease consisting of infectious and non-communicable conditions, injury and interpersonal violence, and maternal and child mortality. Inequalities in income and opportunity push disease burdens towards vulnerable populations, a situation to which the health system struggles to respond. There is an urgent need for health planning to account for the needs of marginalized groups in this context. The study objectives were to develop a process to elicit the perspectives of local communities in the established Agincourt health and socio-demographic surveillance site (HDSS) in rural north-east South Africa on two leading causes of death: HIV/AIDS and violent assault, and on health surveillance as a means to generate information on health in the locality. Methods Drawing on community-based participatory research (CBPR) methods, three village-based groups of eight participants were convened, with whom a series of discussions were held to identify and define the causes of, treatments for, and problems surrounding, deaths due to HIV/AIDS and violent assault. The surveillance system was also discussed and recommendations generated. The discussion narratives were the main data source, examined using framework analysis. Results The groups identified a range of social and health systems issues including risky sexual health behaviors, entrenched traditional practices, alcohol and substance abuse, unstable relationships, and debt as causative. Participants also explained how compromised patient confidentiality in clinics, insensitive staff, and a biased judicial system were problematic for the treatment and reporting of both conditions. Views on health surveillance were positive. Recommendations to strengthen an already well-functioning system related to maintaining confidentiality and sensitivity, and extending ancillary care obligations. Conclusion The discussions provided information not available from other sources on the social and health systems processes through which access to good quality health care is constrained in this setting. On this basis, further CBPR in routine HDSS to extend partnerships between researchers, communities and health authorities to connect evidence with the means for action is underway.
The study setting was the Medical Research Council (MRC)/Wits Rural Public Health and Health Transitions Research Unit in rural northeastern South Africa, which oversees the Agincourt Heath and Socio–Demographic Surveillance Site (HDSS). The Agincourt site was established in 1992 in response to an absence of vital information on rural populations in South Africa, and has conducted annual censuses since collecting information on vital events (births, deaths bind migrations) in a population of approximately 110 000 occupying 21 000 households across 31 villages (Figure 1) [29,30]. Agincourt established the Learning, Information, dissemination and Networking with Communities (LINC) group in 2004 to enable community participation in research and governance. LINC enhances research quality through the feedback of research results to community stakeholders [31]. Through these activities, the Agincourt HDSS tracks population health over time, measures the impact of interventions, supports community research and addresses gaps in population health data [32]. Agincourt Health and Socio-Demographic Surveillance Site (HDSS) in Bushbuckridge Municipality, Mpumalanga Province, South Africa. The research adopted a community–based participatory research (CBPR) approach [33,34]. This was based on the premises that deaths identified through routine health surveillance have social and health systems determinants, and the mechanisms through which these factors influence health outcomes can be reliably identified with local knowledge [35]. Given the time and resources available, communities participated in identifying and defining health problems only. Other than in terms of health surveillance, the communities did not formulate remedial actions, and these were not implemented or evaluated. The research was of a pilot nature exploring feasibility and providing formative data as a basis form which to develop fuller forms of participation in the study setting. Three villages were selected in the surveillance area in which to convene the discussion groups on the bases of demographic variation (Table 1) and feasibility within the time and resources available. LINC staff then approached women of reproductive age (WRA), family members, traditional healers, religious leaders, community health volunteers, health workers, village officials, and community leaders in villages to convene discussion groups that broadly represented the community. To mitigate against any potential bias as for result of social and power differentials in the groups, the group consisted of women only in one village (Table 2). Characteristics of selected villages* *Source: Household data collected by the MRC/Wits Rural Public Health and Health Transitions Research Unit (Agincourt), June 2013 [36]. Composition of discussion groups *All participants recruited were 18 y or older. Although participants typically had >1 role in the community, one primary role per individual was adopted for the purposes of convening the focus groups. Primary roles were also confirmed with participants.It was acknowledged that people with working arrangements, particularly village health workers and village officials’ availability for five consecutive weekly meetings could be compromised. We also acknowledged the ethical imperative of engaging participants who would otherwise be involved in earning income and/or in the provision of public services. Participant recruitment was based on the compositions above with a degree of pragmatism and flexibility towards those committing to the process, and with careful consideration of minimising disruption to local public services. † Group C was a womenonly group to mitigate against the power differentials arising from the heterogeneous constituency of the groups. ‡Close relative: parents, grandparents, siblings, children, in–laws, nieces, nephews and cousins. §Group C was a women only group to mitigate against the power differentials arising from the heterogeneous constituency of the groups. An introductory meeting was held where the purpose, planned activities and outputs of the study were described. Those willing to participate were enrolled, written consent was taken and written information on the study was provided. A total of 24 participants were recruited into three village–based groups of eight participants, which operated independently in a series of weekly discussions on four selected health conditions, plus the introductory meeting described above (Table 3). The conditions were selected on the bases of high prevalence (HIV prevalence is 45% among men and 46% among women aged 35–39 in the Agincourt HDSS [37] and mortality from violent deaths is “outstandingly high” [38]) and in terms of socio–cultural relevance [39–42]. Schedule of focus group discussions A qualitative approach to data collection and analysis was adopted to elicit the collective perspectives of the village–based groups on the relationships between medical problems and their social and health systems determinants. The focus group discussion (FGD) method was used to encourage participation, to capitalize on communication between participants and to explore people’s knowledge to gain an understanding of the collective norms and attitudes surrounding the two conditions [43,44]. A series of five weekly FGDs of 1.5–2 hours were held in each of the three villages, 15 FGDs were held in total. A senior qualitative investigator (SN) with detailed knowledge of the area facilitated the discussions. SN presented data gained via the annual census on the conditions to the groups and facilitated discussions on this basis. Topic guides were prepared for the meetings in which the conditions, their causes, treatments, and the means through which information on them was generated in the locality (ie, HDSS) were discussed. The discussions were audio recorded and translated from the local language xi–Tsonga into English and transcribed. Two investigators took observational field notes and provided generally assistance during the meetings (LD and KE). The narratives and field notes were the main data sources. Towards the end of the data collection, the groups were presented with and discussed a preliminary analysis to determine the plausibility and relevance of early interpretations of the discussions (Panel 1). Following completion of the data collection, a detailed analysis of the discussion transcripts was undertaken using framework analysis (NH). Framework analysis is a flexible tool to analyze qualitative data with the aim of creating a descriptive overview of an entire data set [45]. This method involved familiarization and coding of the data followed by preparation of summaries/charts to map the range of views on the phenomena of interest [46]. The steps of the framework analysis are summarized in Table 4. NVivo software (QSR International, London, UK) was used for data management and coding. An adapted framework analysis approach [45,46] Ethical considerations related to the research process and outcomes were integrated into the study design [47]. All participants gave informed consent that guaranteed anonymization of all identifiable data in study reports, and assured participants that they were free to leave the study at any time and for any reason. The study protocols were peer reviewed to determine local, methodological and substantive relevance. The Human Research Ethics Committee at the University of the Witwatersrand (clearance #M121039) and the Mpumalanga Province health authority also reviewed and approved the study protocol.
N/A
