Preparing for and Executing a Randomised Controlled Trial of Podoconiosis Treatment in Northern Ethiopia: The Utility of Rapid Ethical Assessment

BACKGROUND: Community-based randomized controlled trials are often complex pieces of research with significant challenges around the approach to the community, information provision, and decision-making, all of which are fundamental to the informed consent process. We conducted a rapid ethical assessment to guide the preparation for and conduct of a randomized controlled trial of podoconiosis treatment in northern Ethiopia. METHODS: A qualitative study was carried out in Aneded woreda, East Gojjam Zone, Amhara Regional State from August to September, 2013. A total of 14 In-depth Interviews (IDIs) with researchers, experts, and leaders, and 8 Focus Group Discussions (FGDs) involving 80 participants (people of both gender, with and without podoconiosis), were conducted. Interviews were carried out in Amharic. Data analysis was started alongside collection. Final data analysis used a thematic approach based on themes identified a priori and those that emerged during the analysis. RESULTS: Respondents made a range of specific suggestions, including that sensitisation meetings were called by woreda or kebele leaders or the police; that Health Extension Workers were asked to accompany the research team to patients’ houses; that detailed trial information was explained by someone with deep local knowledge; that analogies from agriculture and local social organisations be used to explain randomisation; that participants in the ‘delayed’ intervention arm be given small incentives to continue in the trial; and that key community members be asked to quell rumours arising in the course of the trial. CONCLUSION: Many of these recommendations were incorporated into the preparatory phases of the trial, or were used during the course of the trial itself. This demonstrates the utility of rapid ethical assessment preceding a complex piece of research in a relatively research-naive setting.