Understanding the factors enabling and blocking sustained implementation of cholera interventions in a fragile region of Nigeria: a multi-phase group model building study protocol

Study design and theoretical framework

This study will utilise a mixed-methods design, conducted in two phases. The first phase will use a cross-sectional descriptive design to address objective 1, and a qualitative study design using key informant interviews, underpinned by the social constructionism philosophical worldview, to address objective 2. The second phase (for objectives 2 and 3) will utilise a GMB approach, underpinned by community-based participatory theoretical framework²⁶. This will identify the dynamic interactions between health system capacity to support cholera progammes and local and national barriers and facilitators for implementing multi-sectoral cholera interventions; and identify leverage points (potential opportunities) for interventions and collaboration across stakeholder groups. GMB introduces social dynamics, which can affect the quality of model, buy-in from stakeholders, and ultimately the likelihood that recommendations from the model will be accepted and implemented by stakehoders. It also provides the opportunity for stakeholders to share their mental models on cholera and adjust their mental models as they learn from other stakeholders through moderated participatory engagement.

Study settings

Nigeria is made up of 36 states and the Federal Capital Territory (Abuja), with each state further disaggregated into several LGAs (there are 774 LGAs in Nigeria). The study will be conducted in Adamawa and Bauchi states, in north-eastern Nigeria (Figure 1). These states were selected because they were among the most affected states during the cholera outbreak in Nigeria in 2018¹⁶. Adamawa has some LGAs that have been directly affected by Boko Haram insurgency while Bauchi has some of its LGAs serving as host communities to persons displaced from Boko Haram affected states of Adamawa, Borno and Yobe. Adamawa state has its capital city in Yola and has an estimated population of 4.7 million people across 24 LGAs²⁷. Bauchi state has its capital in Bauchi and has an estimated population of 7.5 million people across 20 LGAs²⁷. The estimated number of primary, secondary and tertiary health facilities per 100,000 persons in Adamawa is 24.70, 0.69 and 0.03 respectively; 16.40, 0.36 and 0.03 in Bauchi²⁸. In addition to Adamawa and Bauchi, we will also conduct both key informant interviews and GMB in Abuja in order to capture the perspectives of national cholera stakeholders.

Figure 1. Nigerian map showing the study locations.

States in north-east Nigeria are coloured yellow, while the FCT (part of north-central Nigeria) is coloured blue. Source: StraightNews Nigeria, with permission.

Data collection, management and analyses

Phase 1: Quantitative cross-sectional study

Preparations for data collection

A minimum of three research assistants per state will be recruited and trained for data collection, as well as taking informed consent and adhering to ethical practices. Data collection tools will be updated following piloting with healthcare workers in a purposely selected health facility in Abuja to determine completeness, clarity and accuracy. Data collection will be done using Open Data Kit (ODK) Collect installed on password-protected mobile devices.

Sampling

Based on pragmatic considerations (e.g. cost and time), we will adopt a purposive stratified sampling approach to recruit 20 rural and 30 urban health facilities in each state (n=100), with a history of confirmed cholera cases during the 2018 outbreak. Only health facilities (government- and private-owned primary, secondary or tertiary) considered functional by the state health authority will be eligible for selection into the study. The selection of contrasting LGAs (urban vs rural) is due to the dependence of cholera attack and case fatality rates on setting type in Nigeria, with higher values in urban than in rural areas¹⁶. Convenience sampling of healthcare workers of various professional cadres present on the day of the survey will be done.

Data collection

In determining the health system’s capacity to support multi-sectoral cholera interventions in both the Adamawa and Bauchi states, we will adapt the WASH health care facility core indicators²⁹ and the WHO Service Availability and Readiness Assessment tool³⁰. The trained research assistants will spend 1–2 days at each selected health facility. They will administer a structured questionnaire with the manager or administrative head regarding the current health system’s capacity to support various cholera interventions (e.g. WASH, OCV, surveillance etc.). Additionally, the research assistants will administer a questionnaire on cholera case management and knowledge with a convenience sample of five healthcare workers who meet predefined eligibility criteria—i.e. willingness to participate in the study and to sign an informed consent form—on the day of health facility survey. Prior to the survey, information about the proposed study will be communicated to healthcare workers via a formal letter to the facility manager or administrator. When possible, the research assistants will also perform discrete observations to objectively ascertain the presence of cholera interventions. In terms of duration, data collection will be completed within six weeks of the study start date. The study materials and consent form can be found as extended data³¹.

Data analyses

The current health system capacity to support various cholera interventions will be determined using descriptive analyses. Continuous variables will be described using mean and standard deviation for normally distributed variables, median (IQR) for non-normal continuous variables, and frequency and percentages (%) for categorical or binary variables. The status of a cholera intervention will be described by calculating its available indicators divided by the total number of indicators. Analyses will be stratified by state, study setting (urban vs rural) and health facility type. The responses on knowledge will be scored as ‘1’ while other responses, such as ‘incorrect’ or ‘don’t know’, will be scored as ‘0’ (zero). The scores will be added to obtain a total score for each study participant and a median score will then be calculated. High knowledge score will be defined as a total score ≥median score and low knowledge score as a score <median score. Factors (e.g. cadre, state, training, recent case management experience etc.) potentially associated with knowledge will be explored using multivariable logistic regression. The findings will be presented as adjusted odds ratios with 95% confidence intervals. A p-value of <0.05 will be considered statistically significant. All analyses will be performed using Stata version 13 (StataCorp, College Texas).

Phase 1: Qualitative study

Study participants

Three groups of participants (see Table 1 for their distribution by location) will be purposely recruited for the key informant interviews (KIIs). Participants will include:

At the state level:

At the federal or national level:

All study participants must be willing to participate in the study and to sign an informed consent form, and considered a cholera stakeholder (community or healthcare professional). All KIIs will be conducted by the lead researcher (KE) either in a health facility or an office at a time convenient for each study participant. Being a qualitative study without a fixed sample size, we will aim for at least 20 participants in each study location, but recruitment will stop once saturation (i.e. when study participants have provided a range of information or perceptions about the study until no additional information is being provided) is reached³². The semi-structured interview guides will be developed separately for each study participant type (i.e. community members, healthcare providers, national publichealth stakeholders) and piloted before data collection. Where possible, the data collection process will be aided using an audio-recorder. The collected will be transcribed verbatim and analysed using a thematic approach, aided by Nvivo software. Thematic analysis will follow the six-phased approaches recommended by Braun and Clarke³³, bearing in mind that these phases are not a linear process, but more of a recursive process where one moves back and forth as needed. The six phases include familiarisation with data, generating initial codes, searching for themes, review of initial themes, defining and naming themes, and producing the report.

Findings from this study component, alongside those from the cross-sectional study, will be triangulated to finalise the script development for GMB in phase 2.

Phase 2: Group model building

GMB workshop process

We will conduct a total of nine all-day GMB workshops, two with community participants and two with health professionals in both states (total of 8), and one with stakeholders in Abuja (i.e. federal government and partners). Table 2 provides an overview.

We will use locations with minimal distractions to the participants, such as an event centre within a hotel or government ministries. The rationale for participant-specific GMB workshops is to minimise the effect of power dynamics and recognise variation in understanding of cholera by various stakeholders. Between 7–10 participants (12–15 in Abuja given it’s a one-day event) with similar characteristics as those in the qualitative study component will be purposely selected for each GMB session. Invitation of potential participants will be facilitated by a combination of letter/email and phone call. We will ensure that public health preventive measures with regards to coronavirus disease 2019 (COVID-19) (e.g. physical distancing and use of face-masks) are in place during each workshop. The protocol and scripts for each workshop will be modified from those available from Scriptapedia³⁴ and based on preliminary findings from study phase 1. The research team members will be assigned roles as outlined in Table 3. After each GMB workshop, the research team will have a debrief session to reflect on processes and to make necessary adjustments as needed.

GMB tools

The GMB process will utilise three interactive system mapping tools (see provisional scripts for these tools in Extended Data): graph over time; cognitive mapping; and causal loop diagrams.

Graphs over time allow the participants to share their understanding of cholera as well as perceived drivers of its recurrent transmission in their community over a specific time period. Using an empty graph with time on the x-axis (with a vertical line for the present time) and a variable on the y-axis, the participants will be guided by the researchers to fill in the graph. They will be prompted with questions such as: “What is the trend of cholera in your community since the 2015 presidential elelction in Nigeria?”, “What are the factors influencing the transmission of cholera in your community?”, “What are the interventions available for cholera control in your community?”. As well as drawing on the empty graphs the historical trends of cholera, the participants will be asked to identify two future pathways they predict would occur if current cholera trends continued or if intervention occurred.

Cognitive mapping is a visual tool that will introduce participants to systems thinking by exploring their understanding of the facilitators/barriers to the implementation of cholera interventions (e.g. surveillance and oral cholera vaccination) and consequences of implementing these interventions successfually or not²⁵. For this activity, a template will be developed and provided for the participants to complete.

Causal loop diagrams capture the dynamic nature of an issue and the presence of feedback in systems²⁵. This activity focuses on providing the participants with an understanding of why feedback is important in a system, using pictoral examples from relevant studies. Here, simple feedback loops which are a basic operating unit of systems³⁵, will be designed from participants’ previous activities (e.g. barriers to the implementation of cholera interventions and consequences of action and inaction). To further consolidate findings from the causal loop diagrams, a sub-set of of participants will be asked to critique the collective models by adding, deleting and modifying structures in the map. We will also seek to explore potential interventions to identified barriers to implementing cholera interventions in the both Adamawa and Bauchi states. The primary focus here would be to use the following questions to probe for possible actions which could be taken to address the identified barriers: “What variables could you increase or decrease?” “How could you impact connections: strengthen, or weaken a connection, speed it up or slow it down, add or delete connections?” This activity will require the participants to write potential actions on post-it notes and place them on the causal loop diagrams where they consider appropriate. Finally, all the participants will be required to select the top three actions for each intervention that their group would like to see progressed.

Post-GMB workshop

GMB data, alongside field notes and reflections from the research team, will be analysed thematically. Additionally, findings from the GMB workshops will be supported with quotes from thematic analysis of transcripts from the KIIs. Lastly, GMB findings will be digitised using VenSim software, version: 8.2 (an open-source software).

Ethics and dissemination strategy

The protocol for this study has been reviewed and approved by the Nigeria National Health Ethics Research Committee (NHREC Approval Number NHREC/01/01/2007-24/08/2020). Additionally, participants for the qualitative and GMB workshops will be required to provide informed consent after reading or listening to the study information sheet, and will be assured freedom to withdraw from the study at any stage. Autonomy and confidentiality will also be maintained throughout the conduct of this study; For example, we will delete all personal identifiers from the dataset prior to management and analyses, and only the research team members will have access to the dataset on a password-protected laptops. Lastly, while participants will be informed that there are no direct benefits from participating in the study, those in the GMB workshops will however be given stipends to compensate the cost of transportation and potential loss of income for the day—this is appropriate in the context of GMB workshops²⁵. Expected findings will be disseminated through peer-reviewed publications and local and international scientific meetings.

Underlying data

No data are associated with this article.

Extended data

Figshare: Understanding the factors enabling and blocking sustained implementation of cholera interventions in a fragile region of Nigeria: a multi-phase group model building study protocol. https://doi.org/10.6084/m9.figshare.13686073.v1³¹

The project contains the following extended data:

Data are available under the terms of the Creative Commons Attribution 4.0 International license (CC-BY 4.0).