Menu
Menu
Menu
Menu
Background: Somalia has been ravaged by more than two decades of armed conflict causing immense damage to the country’s infrastructure and mass displacement and suffering among its people. An influx of humanitarian actors has sought to provide basic services, including health services for women and children, throughout the conflict. This study aimed to better understand the humanitarian health response for women and children in Somalia since 2000. Methods: The study utilized a mixed-methods design. We collated intervention coverage data from publically available large-scale household surveys and we conducted 32 interviews with representatives from government, UN agencies, NGOs, and health facility staff. Qualitative data were analyzed using latent content analysis. Results: The available quantitative data on intervention coverage in Somalia are extremely limited, making it difficult to discern patterns or trends over time or by region. Underlying sociocultural and other contextual factors most strongly affecting the humanitarian health response for women and children included clan dynamics and female disempowerment. The most salient operational influences included the assessment of population needs, donors’ priorities, and insufficient and inflexible funding. Key barriers to service delivery included chronic commodity and human resource shortages, poor infrastructure, and limited access to highly vulnerable populations, all against the backdrop of ongoing insecurity. Various strategies to mitigate these barriers were discussed. In-country coordination of humanitarian health actors and their activities has improved over time, with federal and state-level ministries of health playing increasingly active roles. Conclusions: Emerging recommendations include further exploration of government partnerships with private-sector service providers to make services available throughout Somalia free of charge, with further research on innovative uses of technology to help reaches remote and inaccessible areas. To mitigate chronic skilled health worker shortages, more operational research is needed on the expanded use of community health workers. Persistent gaps in service provision across the continuum must be addressed, including for adolescents, for example. The is also a clear need for longer term development focus to enable the provision of health and nutrition services for women and children beyond those included in recurrent emergency response.
The study presented here is part of a multi-country study coordinated by the BRANCH Consortium [10] and focused on the delivery of reproductive, maternal, newborn, child and adolescent health (RMNCAH) and nutrition interventions in ten conflict-affected countries: Afghanistan, Colombia, Democratic Republic of the Congo, Mali, Nigeria, Pakistan, Somalia, South Sudan, Syria, Yemen [10]. This study employed a mixed-methods design, including both secondary analyses of existing quantitative data and primary qualitative data collection and analysis. The quantitative component aimed to examine trends in the coverage of RMNCAH & nutrition interventions in Somalia since 2000, as well as conflict-related displacement and conflict-attributable mortality. The qualitative component aimed to better understand the coordination and decision-making processes of humanitarian actors delivering health services during this period, particularly in Bay region and in the capital Mogadishu in Benadir region, in the Central-South Zone of Somalia. Estimates of intervention coverage of RMNCAH and nutrition interventions in Somalia over time were extracted from reports of large-scale household surveys conducted since 2000 and available in the public domain: the 2006 and 2011 Multiple Indicators Cluster Surveys (MICS), the 2009 National Micronutrient and Anthropometric Nutrition Survey (NMANS), the 2016 Service Availability and Readiness Assessment Survey (SARA), and the 2017–2018 Somali High Frequency Survey (SHFS). Where relevant estimates were not published but the dataset was available, we generated estimates from the data, accounting for survey design. Estimates of deaths directly attributable to conflict in Somalia over time were derived from publicly available data from the Uppsala Conflict Data Program [2, 3]. Regional estimates of the cumulative number of internally displaced people were extracted from the Somalia 2019 Humanitarian Response Plan [1]. Primary qualitative data collection was conducted between August and October 2018 in Mogadishu and Baidoa in Somalia and in Nairobi, Kenya. Interview guides were informed by shared templates created by the BRANCH consortium. Through a collaborative process, the research team adapted these guides to reflect the local context [10]. A team of five local data collectors (four men and one woman) was utilised to conduct in-depth interviews. All the data collectors held advanced degrees in public health or the social sciences and had previous experience collecting qualitative data. They also had an in-depth understanding of public health issues in Somalia. They were bilingual in Somali and English which allowed them to conduct interviews in either language, depending on the respondents’ preferences. Interviews were conducted with local NGO staff, international NGO staff, MOH staff, and private health facility staff. Respondents included a range of individuals such as country directors, program managers, health officers, and project coordinators, and were purposively sampled through three strategies. Target organizations for key informant interviews were identified by the study team’s review of “Who does What Where (3W)” matrices produced by the UN Office for the Coordination of Humanitarian Affairs (OCHA), and specific respondents were identified through discussion with organizational representatives about which individuals would best be able to speak to our research objectives. Nominations were also solicited from organizational and government representatives attending a project inception meeting convened by the study team in Nairobi in 2017. Snowball sampling was also used, whereby participating respondents suggested other potential key informants to the study team. Initial contact was made via phone or email depending on the contact information available for the respondent. A total of 27 interviews were conducted face-to-face in the respondent’s offices while 6 interviews were conducted via Skype as the respondents were in geographies the study team could not access. For interviews conducted via Skype, informed consent was provided by participants verbally at the start of the interview, while those participating in in-person interviews provided informed consent through a signed consent form. No invited participants refused to give consent and all agreed to have the interviews recorded. Interviews lasted approximately 60–90 min (Table 1). Respondent information We tabulated quantitative estimates of achieved coverage of key health and nutrition interventions for women and children, stratified by zone (the lowest level of aggregation for which representative estimates could be derived from national survey sampling approaches): Northwest Zone/Somaliland; Northeast Zone/Puntland; and Central-South Zone. Time trends in the frequency of battle-related deaths and lethal violence events were generated nationally. Key informant interview transcripts were imported into NVivo 12 software. Latent content analysis, an inductive analysis approach, was used to analyze the data. First, ZA and AA read the interview transcripts, and then coded passages according to the speaker’s meaning. ZA then merged the codes into categories while ensuring that they had both internal homogeneity and external heterogeneity [11]. These categories were later merged into themes shared in this manuscript. Coding and analysis was undertaken in an iterative manner, so that new ideas that emerged could be probed in additional interviews. Peer debriefing with the analysis team also assisted in identifying areas to further investigate during interviews. Member checking was conducted through several consultations with stakeholders for their input and feedback.
N/A
