Institutional review board

For the study, we obtained IRB approval from the Duke Institutional Review Board, Kilimanjaro Christian Medical Center Ethics Committee and National Institute of Medical Research. There was a minimal level of risk for this study. Patients completed a written informed consent prior to participation but to ensure privacy, names or identifying information of participants were not collected.

Study design

The semi-structured focus group was designed based on the grounded theory and assessed using thematic analysis. Grounded theory is a research method that generates theories grounded in data [14]. Constant comparative analysis, a process in grounded theory used for category development [15], led to focus group questions (appendix A) about community drinking behavior and what is appropriate and acceptable behavior when deciding to or seeing someone drink and drive. Focus groups were conducted between August 2016 and January 2017.

Focus groups were conducted by two trained female research nurses at Kilimanjaro Christian Medical Center (KCMC). The research nurses both have over ten years of experience conducting focus groups among similar patient populations. A relationship was established with the participants at the onset of the focus group sessions. Focus group participants were informed of the aims of the study as well as the qualifications of the research team and focus group facilitators. They were informed that the focus group is a component of a greater research collaboration aimed at reducing alcohol related injury in northern Tanzania. Facilitators reported their interest in the research as based on an investment in the livelihood and betterment of the community.

Study setting

This project took place in Moshi, a city in the Kilimanjaro region of Northern Tanzania with a population of 143,799. Moshi is home to Kilimanjaro Christian Medical Center (KCMC), the third largest hospital in the country and the referral hospital for northwestern Tanzania [16]. KCMC serves the urban and rural population of Moshi and was therefore selected as a central location to assess general perceptions of the region. Assessing the qualities and factors that lead to drink-driving is important for this region as road traffic crash rates are climbing and current data from the KCMC Casualty (Emergency) Department suggests that 28% of all the patients who arrive to the Casualty Department for treatment of an injury consumed alcohol prior to their injury and therefore are ‘Hazardous drinkers’ [17].

Although Tanzania does have a national blood alcohol limit of 0.08g/dl for the general population and 0g/dl for commercial drivers [8], policies including random breath testing, police checkpoints, and penalty point systems are currently not in place [8]. High alcohol consumption is more of a problem in Moshi than other regions of Tanzania because it is less religious than other Muslim predominated regions [18]. The strong Chagga culture of the region views alcohol as an integral part of traditional ceremonies, thus making it a more acceptable habit [19]. Another factor contributing to higher alcohol consumption is the large tourist population attracted by being situated at the base of Mount Kilimanjaro. Tourism brings in significant revenue that contributes to higher disposable income, which is associated with higher alcohol consumption rates [18]. Demand for bars created by tourists could contribute to the wide availability and exposure to advertisements of alcohol experienced by northern Tanzanians [18].

Study population

Focus groups participants were a convenience sample of injury patients, their families, and community advisory board members. Patients and their family members were identified face-to-face in the KCMC ED waiting or treatment areas, after treatment or stabilization, and were offered participation in the focus group. If patients or family members accepted, they were invited to return to the hospital for patient and family focus group days, respectively.

Injury patients (both patients and drivers) were included in the focus groups if they were ≥18 years of age, seeking care at KCMC for any acute (<6 hours) injury, clinically sober at the time of enrollment, medically stable, able to communicate in Swahili or English and consent to participate. Patients were excluded from enrollment if they were medically unstable or had a deteriorating condition, were too critically ill to participate, did not speak English or Swahili, were <18 years of age, presented for non-injury related complaints or follow-up care, or did not consent to be enrolled. Family member focus group participants were family members of a patient who was able to be enrolled in the study, who also agreed to participate, and speak English or Swahili.

Research nurse facilitators attended Community Advisory Board (CAB) meeting to conduct focus groups in order to utilize this group's perspective on drinking behavior within the community. The KCMC CAB is comprised of 30–40 adult community members who understand research, have advised investigators on pertinent research questions, cultural norms, and cultural acceptability of interventions, treatments and research protocols. All interested CAB participants enrolled were ≥18 years of age and were present at the monthly CAB meeting when we conducted the focus group discussion.

Data collection

Patient and family focus groups were scheduled once 5–10 eligible interested participants were recruited from the emergency department and provided informed consent. CAB focus groups took place prior to the monthly CAB meeting and consisted of 5 to 10 members who provided informed consent. No participants dropped out. Each focus group, led by trained research nurses, lasted between 45 and 60 minutes. They were audiotaped and transcribed for formal qualitative analysis utilizing thematic analyses. Transcriptions occurred within days following the focus group and research nurse notes were included in the transcriptions about the content.

Focus groups among each population group, patients, families, and CAB members, were conducted using an iterative process until thematic saturation was reached. This was done for four patient, four family member, and two CAB focus groups. The original focus group questions were designed to understand the general perceptions of drink-driving in Moshi Tanzania. These questions included “What is acceptable drinking behavior?” and “What is thought of a person who drinks and drives?”. These scripted questions were piloted by the research nurses conducting focus groups. After thematic saturation was reached, the focus group script was altered to expose reasons for preliminary data from patient surveys. Focus groups were conducted until thematic saturation was again reached.

After the focus groups were recorded and transcribed, each script was translated from Swahili to English. English scripts were then assessed for potential cultural misinterpretations, back translated and annotated for English, American comprehension. Transcripts were assessed by bilingual American and Tanzanian researchers and not returned to participants as they were difficult to contact after completion of the study. All transcripts, audio tapes, and related data will be kept for six years after study completion.

Data analysis

Analysis was iterative throughout the study, which allowed emerging themes, derived from the data to be explored in later focus groups. The focus groups among patients, family members, and CAB members were coded separately and then analyzed. Comparing and contrasting across and within these datasets highlighted emerging themes and divergence of perspectives. Thematic saturation occurred when no new themes developed from focus group analysis and marked the end of the qualitative study for individual population subsets.

All transcripts were coded by DE and BM, using a thematic narrative approach, reflecting the research questions and themes raised by the participants [20]. The researchers (DE and BM) separately completed coding with primary and secondary level coding classifications. DE and BM then compared coding with advisors specializing in qualitative research. The Tanzanian research group reviewed the evolving thematic codes and resulting narratives and gave input based on their experience with the focus group populations and cultural knowledge. Representative quotes for each theme were then selected based on comparative analysis of DE and BM coding with input from research team members.

Demographics and characteristics of participants

A total of ten focus groups were conducted (4 patient, 4 family, and 2 CAB) with a total of 104 participants (37 females and 67 males). Demographic information is not presented here in the interest of maintaining respondent confidentiality. While the study was not intended to focus on commercial drivers, most participants spoke about boda boda or dala dala drivers. Focus group discussions between all three population types highlighted major themes of ‘drinking is ingrained in boda boda driver culture’, ‘individuals have a personal responsibility to address drink-driving’, and a ‘police enforcement on drink-driving is necessary’.

Emerging themes and related quotes

Limitations

One limitation of the study is the potential for social desirability bias, especially on a stigmatized topic such as drink-driving. To limit the effect of this potential bias, we used research personnel who were comfortable talking about stigmatizing topics and included questions to talk about the general society and not the participants individual actions or behaviors. As inclusion criteria required participants to be injury patients who were medically stable to enroll in the study, we lack the perspective of those who were critically ill or in deteriorating condition. The fact that we enrolled injury patients and their family allows us to understand the motivations for continued drink-driving to plan for interventions, but might not fully represent the Tanzanian general society’s perspective about drinking or drink-driving.