Development of the messages

As indicated above, communication about Ebola to the Sierra Leonean population during the early phases of the epidemic had been conducted on a top-down basis, and there had emerged a recognised need to take into account the concerns, views and experiences of ordinary people in order to ensure that the messages being disseminated were both relevant and understood as intended. On this basis, our 3-month message development project took an iterative approach, as illustrated in Fig 1: the top row represents the study team and the national and international level stakeholders, while the bottom row represents the community. The arrows linking the five boxes show how the process moved between the levels, whereby preliminary investigations were conducted into community perceptions of Ebola, Ebola messages, and the Ebola response. The findings from this formative research informed our development of draft messages aimed at promoting Ebola treatment-seeking behaviour. These draft messages were then field-tested, refined, and disseminated. The rest of this section presents this process in more detail.

The project was formally launched during a meeting in Freetown on January 17, 2015, attended by consortium members as well as representatives from a number of national and international organisations, including the Ministry of Health and Sanitation (MoHS) and the National Ebola Response Centre (NERC). The meeting was intended to raise awareness of the project among key stakeholders, and to seek their input into the process.

Seven MRC research assistants and transcribers, all experienced in qualitative research, were then trained, with sessions on the study methodology, safety and ethical issues (especially consent and confidentiality), and data security. Formative field work was undertaken during late January and February, in two study areas: urban Freetown, and rural Bombali (see Fig 2). Both urban and rural areas were included because we wanted to understand the different opportunities and challenges that these contrasting settings can present for messaging. The specific study sites were chosen because they were currently experiencing or had recently experienced active Ebola transmission.

[Source: World of Maps, licensed under the Creative Commons Attribution-Share Alike 3.0 license].

The formative research phase included 16 Focus Group Discussions (FGDs) comprised of people without any sort of leadership position in their respective communities. The FGDs were stratified by age (18±25 years, and >25 years) and sex, and were attended by 118 people (60 male, 58 female; mean = 7.4 participants per FGD). We also held a total of 24 individual, in-depth interviews with community leaders of various sorts (12 in each study area; 13 male, 11 female). These included religious leaders, traditional leaders, traditional healers, women’s and youth leaders, and medical staff engaged in the Ebola response. The reason for interviewing the latter group one-on-one was primarily for logistical reasons: such people are often quite time-constrained, so it can be challenging to bring several of them together in one place and at one time to take part in an FGD. However, this approach also permitted us to seek their individual professional perspectives and experiences, while the FGDs provided us with insights into the norms and social processes encountered in the wider community [20]. The educational level of our respondents was higher than the national average, in part because of our inclusion of the leaders and health workers: 17% had no education (as compared to 49% nationally [21]), 14% had at least some primary education, 38% had at least some secondary education, and 31% had at least some tertiary education.

The recruitment process for both FGDs and interviews followed a similar pattern in each participating community. First, we introduced the study to the respective village chief, who then called a meeting with key stakeholders, including traditional leaders, imams, pastors, women's leaders, youth leaders, health personnel, and teachers. During this meeting, our team explained the reason for the research, and that we wanted participants for both the FGDs and the interviews to include a mix of tribes, families, different geographical areas of the village, and people with different socioeconomic status and occupations. The key stakeholders then identified a list of possible participants, who we subsequently visited, accompanied by someone from the community to show us where they lived and to provide us with an introduction. We engaged individually with each suggested participant and confirmed that they met the selection criteria: willing and able to talk Krio, to talk about Ebola, and to take part. (A participant’s ability to communicate in Krio was necessary as it was spoken by all our interviewers, while not all of them spoke all of the other vernacular languages in the study areas. Some well-informed people who could not speak Krio may therefore have been excluded, but since as many as 95% of the Sierra Leonean population know the language [22], it is unlikely that this would have introduced any significant bias.) This process continued until we had the full quota for that community. The participant categories are presented in Table 1.

Questions in the FGDs and interviews were concerned with awareness of Ebola itself, Ebola messaging, and–based on knowledge from our in-country team members–issues to do with the Ebola response, such as treatment, ambulances and burial teams. Discussions and interviews were conducted primarily in Krio, but some discussions also included Mende, Tembe, Limba, or Loko. The data were recorded digitally, and transcribed directly into English; data quality was ensured by comparing audio recordings with the English language transcriptions, and correcting all transcription anomalies. Participants took part on a voluntary basis, having first provided written informed consent. As a token of gratitude for their contribution to the research, all participants received two bars of soap. Although some participants told of distressing experiences or stories, none were so upsetting that referral for counselling was deemed necessary.

A message development workshop was held in Freetown in mid-March, attended by eight consortium team members, the seven research assistants and their field supervisor, as well as representatives from the MoHS, the US Centers for Disease Control, and local NGOs such as Focus 1000. Four team members had been assigned to closely review all the transcripts prior to the meeting. Following the established principles of thematic analysis [23], these team members independently generated a list of recurring themes that emerged from the data. These were supplementary to the themes that had informed our questions as a starting point (awareness of Ebola itself, Ebola messaging, and issues to do with the Ebola response, such as treatment, ambulances and burial teams).

At the start of the workshop we collectively synthesised the four individual sets of themes into a single set of core themes for use in the message development process. These included issues such as trust and distrust, bribery and corruption, misperceptions about people being killed within the health system, and ‘seeing is believing’ (i.e. that people who had initially denied that Ebola was real changed their minds when they saw survivors). After reflecting on some basic health promotion and communications concepts (for example, gain- vs loss-framed messaging [24], cultural competence [25], and audience segmentation [26]), we then worked over five days in two teams of around eight people, each focusing on a different group of the core themes. Through this, a total of 26 draft messages was produced, each with an accompanying rationale, audience, dissemination channel/s, messengers, and list of associated operational issues. We define ‘Channel’ here as the means of distributing a given message, such as a poster, leaflet, radio jingle, or house-to-house meetings etc.; while the ‘Messenger’ is either an individual such as a traditional leader or an Ebola survivor who is depicted in or who articulates the message, or an institution such as the Ministry of Health and Sanitation which is seen to be responsible for it. Both English and Krio versions of the messages were produced, and two artists provided sketches for some of the main messages.

During late March and early April 2015, the 26 draft messages were field-tested in a series of eight FGDs in the same communities where the formative research had been conducted (four FGDs in each of the two study districts, with a total of 32 male and 32 female participants–these FGDs were not age-stratified). All of these people had participated in the initial FGDs, so they were already familiar with the project and its objectives. Each message was discussed in two different FGDs: we sought the perspectives of both women and men as well as both urban and rural participants on each message. For example, one message might be reviewed by rural men and urban women, while another could be discussed by rural women and urban men. In order to keep the discussions focused, each FGD covered no more than five different message topics.

Field testing sought to assess understanding, acceptability, perceived likely effectiveness of the messages, as well as appropriate distribution channels and messengers. Although the process identified a number of issues in the messages, all of them were seen as broadly acceptable and none were removed altogether. With this feedback, we reworked the messages into 14 topic areas and then formally presented the final document to the MoHS at a multi-stakeholder meeting, on April 15, 2015.