Irrespective of the final conclusions relating to REA, the study highlighted important considerations for improvements in the research ethics system in Ethiopia. Whilst there have been a number of efforts at institutional and national levels to build a competent ethics governance and review system, the views of stakeholders suggest that further work is required. Given the relatively narrow range of individuals surveyed (institutional researchers and regulators), it could be claimed that more comprehensive assessment of the research ethics governance system in Ethiopia is needed before making further remarks. The study also identified clues specifically related to the research consent process in Ethiopia both at information provision and decision-making stages. The findings suggested that REA tools could be considered relevant and potentially feasible in the Ethiopian context in order to address these gaps.
The on-line survey for the quantitative component, which we believe to be one of the first uses of this technique in Ethiopia for health research, was efficient in terms of generating information rapidly. However, researchers who were not accessing their e-mails during the data collection period were not included in the study. On the other hand, a significant proportion of the intended sample size was reached, as respondents were able to answer the survey questions irrespective of their physical availability or current location given internet connectivity. The fact that the respondents themselves were researchers who understood the importance of responding to surveys may have contributed to the high level of compliance [15, 16]. It was possible to send reminders and additional invitations on the basis of responses. This may have resulted in selection bias as busy people and those on vacation during the survey might not have responded to the survey. However, we do not know reasons for non-participation and non-response. The study was conducted at the beginning of the academic year, assuming most academics would have time to complete an online survey. The study has demonstrated the potential of using on-line survey tools in Ethiopia among groups such as academics and researchers. The other advantage of conducting the on-line survey was the use of the preliminary findings to shape the qualitative study.
The mix of study respondents is representative of the group targeted for the study; high level researchers and regulators of ethics review systems with considerable research experience. Their current roles and experience in research made them ideal to identify gaps in health research ethics, and to suggest possible ways of addressing them. However, since the study was limited to researchers and regulators, the perspectives of other important stakeholders in health research such as non-researcher health professionals, community members, senior officials and country level policy makers were not included in the study. Another point to note about the respondents is that, despite the fact that most were involved in conducting research, in ethical review or in the development of ethics review guidelines, a significant proportion of respondents had not had any formal training in research ethics. Possible explanations for this include the absence of structured research ethics training courses and the fact that Ethiopia is in the early stages of implementing a universal system of research ethics. One example of this is the recommendation given by the study participants about having a community representative on ethics review committees. This is already included in the national ethics guideline , but is not uniformly implemented. This lack of standardized and structured ethics training is a critical gap, and from the qualitative findings, there are suggestions that it may lead to systemic gaps in the ethical conduct of research. Other studies have documented knowledge gaps among academics and have suggested ways of including ethics training in mainstream curricula [18, 19]. In addition, the qualitative study did not include representatives of all academic and research institutions, which makes it difficult to generalize the findings to other settings. Yet we believe that most of the concerns are shared by similar institutions. The use of a mixed-methods design served to triangulate findings as the qualitative assessment provided deeper insight into the survey findings and bridged the gaps in the survey assessments. The online survey enquired, with closed options, into ethical pre-assessments and REA in reference to involvement of potential study participants, without further explanation of the REA concept. With the IDIs it was possible to discuss further, explain the essence of REA and enquire into the involvement of community members beyond study participants. Overall, the study was able to identify key issues relating to research appraisal and explore perceptions surrounding REA.
Respondents’ perceptions of existing consent processes were not found to be favourable. Most respondents thought that potential participants understood little about the consent process or the information provided to them. They also thought that participants’ best interests were rarely considered, reflecting gaps in communication and the decision-making process. The major gaps for study participants in relation to communication and comprehension included use of incomprehensible specialist terms to explain medical concepts, lack of health and research awareness of participants, undue expectations and manipulation by researchers. The same list of issues could influence decision making, which in turn is influenced by factors such as manipulation by researchers and the local dynamics in decision making. These factors vary from context to context. In Ethiopia ethno-cultural differences are very visible, and they make the use of generic consent forms and consent approaches for varied setting inappropriate.
Traditionally, consent forms have been developed by researchers based on the requirements of an information sheet and a decision page. This approach addresses consent from the perspective of the basic principles in ethics and the major international guidelines. The assumption is that the contents would be relevant fairly standardized irrespective of who the participants are. Sometimes consent forms have been developed by international investigators and ‘adapted’ purely by translation.
In Ethiopia, communities have varied levels of awareness of public health issues. According to the Ethiopian Demographic Health Surveys (EDHS), levels of understanding vary by characteristics such as area of residence, income, region, religion, ethnicity, and literacy status . Tekola et al. reported that words for ‘health’, ‘research’ and ‘medical treatment’ either do not exist in some languages or are used interchangeably in a confused and confusing way . This may give rise to expectations of medical treatment which confound the decision making processes. Some researchers might be tempted to take advantage of this vulnerability of participants to increase rates of recruitment by either promising unrealistic and unavailable benefits or at least by not challenging misconceptions. Gaps in the consent process are reported elsewhere related to information and communication and decision making and indicate the need for informed consent processes tailored to the context [1, 8–10, 21–26]. The use of qualitative assessment in the informed consent process for medical research in developing countries have also been reported .
Based on previous studies which have employed REA, consent processes have significantly improved in terms of both comprehension and decision making as a result of the knowledge and understanding gained and the steps taken to incorporate it. The key steps were identification of issues prior to and during the conduct of studies and guiding the consent process (including the provision of information) based on those qualitative findings. The REA tool has been recommended for use by researchers who have used them in their respective research [7–11]. The current qualitative findings support the need for pre-assessment to explore potential context-specific ethical issues. Pre-assessment in health research may have various objectives such as testing data collection tools and assessing study feasibility [28–32] and optimizing community engagement . However REA is distinct from these forms of pre-assessment being focused primarily on ethical issues.
It is interesting that a significant proportion of the respondents mentioned they were aware of the role of some form of pre-assessment and stakeholder participation in exploring ethical issues. However, the extent of actual use of pre-assessment was not measured, and neither was any quantitative assessment made of the techniques that might already be used by researchers. The qualitative responses revealed that ethical pre-assessment are not widely used, and that the term may be used to refer to feasibility studies or piloting of tools. Genuine ethical pre-assessment is rare, and stakeholder participation usually refers to involvement of community representatives in identifying the problem to be investigated rather than the consent process per se, [34–36]. There are fewer experiences of pilot studies for ethical pre-assessment, and these were limited to the specific research projects under consideration and did not investigate the openness of researchers to their wider scale use [37–39].
Currently in Ethiopia, there is no established mechanism in place to assess the community risks, vulnerabilities and benefits beyond what is written in the application submitted for approval. Most IRB reviews are based on a simple Risk-Benefit assessment of the submitted proposal and an assessment of whether the consent form meets an acceptable standard and is written well. However, one important parameter in IRB formats is ‘Involvement of the community in the research’. Again, in practice this assessment of community involvement is based solely on what has been included in the proposal, and refers mainly to involvement in the development of the research question, which is understandable given that one of the implicit principles in community health research, is that ‘community concern’ is an important criterion.
A ‘pre-assessment’ which includes the community would reduce the reliance of the IRB on making a judgment based on the written scientific proposal alone. Community engagement approaches, which aim to create awareness and a sense of ownership by the community, are becoming more popular. These approaches often work with and through community groups such as Community Advisory Boards. Community engagement is also used as a process of influencing change in the community through provision of information, negotiation, local capacity building and empowerment [33, 40]. Community engagement might also be used to address ethical issues around research [26, 41–43]. Community engagement and REA share some overlaps, since they both address community issues and share qualitative methodological approaches. However, REA is primarily conducted by a REA team (a multi disciplinary team of researchers) and employs rapid ethnographic research methods, while community engagement will more commonly seek to involve community members and to engage over a longer period than suggested for REA. The two are not mutually exclusive.
The study documented considerable interest in REA as a tool to improve the research consent process. However, respondents were concerned about the potential burden REA would put on the researcher in terms of time and other resources. There is a concern that REA may unnecessarily delay small, cross-sectional research projects into less sensitive topics. One suggestion made was to conduct REA as part of the traditional pre-test during community based studies. However, the objectives and duration of such pre-tests vary considerably. Some are rapid and done in a day or two and would not permit REA, which required about 4–6 weeks in the earlier studies [8, 10]. Other types of pilot studies such as feasibility studies for randomized trials would be ideal in terms of integrating REA. The studies so far employing the REA method have documented its significant contribution in identifying important ethical issues in the research context [10, 11]. However, such studies have not assessed the usability of the tool by the research community. As REA is an additional tool which requires time and expertise, research teams need to take this into account when planning their project. It is then important to be clear about what resources are required so that researchers can take this into consideration. In addition, REA findings might possibly raise issues relevant to study design and therefore might best be done in advance of most pre-tests.
According to the findings of the current study, the ‘rapid’ aspect of the REA tool was at times confused with the idea of expediting the review process. Some respondents were confused about the terminology as they tended to understand that the tool was ‘rapid’ and was meant for accelerating the ethical review process. Most of the researchers indicated that the current appraisal process takes too long and there is a need for finding a way to improve this. They were intrigued to learn that REA entails additional time and resources.
Whether REA should be applicable to all studies depends on a number of issues such as the nature of the study, the characteristics of study participants, the nature of the issue under investigation and the availability of resources. Whilst it is important to avoid crude generalisations and stereotyping when setting ethical standards, it is possible to learn from similar cases. Sometimes earlier studies of a similar nature will provide information on which consent processes can be designed. Clinical trials and studies that include vulnerable subjects or sensitive issues would often require REA. In clinical trials and classical longitudinal studies, there are a number of encounters with the study subjects. This would allow plenty of room for addressing consent issues. Such longitudinal studies require repeated and long term encounters between researchers and the community. As the consent process is repeated, it can be improved. Some communities are better informed about science and research than others and are well known to the researcher. In localities where health systems are well accessed and previous research has been conducted, the community generally has a better understanding of health research and the anticipated ethical issues will already be familiar to researchers. REA is therefore likely best reserved for communities that are not well researched and are less familiar to researchers.
The main objective of the current study was to assess the attitudes of the research community towards the REA tool and their views on its perceived relevance. In this paper we did not intend to explore the issue of practical feasibility beyond recording the perceived relevance and applicability of the tools as expressed by researchers and research ethics reviewers. Perceived relevance of the tool is different from its practicality and feasibility which need to be explored further. Expressed demand and acceptability of an approach are important components in assessing feasibility but are not sufficient . However this study has clearly demonstrated the openness of researchers to considering the introduction of such a tool and suggests that they might willingly collaborate in developing a feasible intervention.
Qualitative data analysis was done in English while conducting the analysis in the language of data collection would have been preferred. Community representatives were not included in the study and the assessment regarding the research ethics review system in Ethiopia was primarily based on perception of researchers and ethics committee members. We used post graduate training as a proxy indicator of research experience which may not always be correct as there are variations in the profiles of post graduate programs.