Study site

The study was conducted at the Kenya Medical Research Institute (KEMRI)-Wellcome Trust research programme (KWTP) in Kenya. The programme’s main centre is located in Kilifi County Hospital on the coast of Kenya, where a long term collaboration has been established with Ministry of Health managers and providers. Through this partnership, KWTP supports clinical services in Kilifi hospital and some peripheral clinics, including supplementing staff, supplies and equipment and providing a paediatric high dependency unit at the hospital [8]. Kilifi County’s population includes rural and semi-urban populations with the majority of the residents being from the Mijikenda ethnic group in Kenya. Statistics for the county show amongst the highest poverty levels, lowest literacy rates and highest indicators of gender inequity nationally [11, 12]. Amongst its activities, KWTP runs a Health and Demographic Surveillance System in the population of 260,000 people living within the hospital’s catchment area, in collaboration with County Health managers [13]. This is the population referred to as ‘the community’ in this paper.

Within the community, KWTP’s role in supporting clinical services is generally highly appreciated. It’s research role is less well and widely understood, contributing at times to concerns and rumours about common research practices such as taking blood and other biological samples, and the involvement of people who are well in research activities [12]. Given its close relationship with the community, KWTP has an active community engagement strategy to build mutual understanding and strengthen community input into research practice and policy [14]. Of relevance to this paper, these activities include regular meetings for information sharing and consultation with a network of around 200 voluntary KEMRI community representatives (KCRs), ‘typical’ community members selected by residents within smaller administrative geographic locations at public meetings to serve for a period of three years [15]. A team of community facilitators drawn from the community implement engagement activities; and other front line staff in the programme who support research activities (field workers) are also long term residents.

Study population, sampling and data collection

A total of 90 research stakeholders participated, including i) 33 KWTP staff (community facilitators and field workers), ii) nine area administrative leaders (assistant chiefs), iii) 22 opinion leaders (leaders or members of community-based organizations, including women groups, youth groups and village dispensary committees), iv) 22 KCRs and v) four mothers of current study participants. These types of participants were chosen to represent a range of perspectives on research activities while ensuring some level of familiarity with research to facilitate debate and discussion. Within groups, participants were purposively selected to maximize diversity in gender, location (urban and rural), religion, age and education. Staff participants had a high understanding of research, including research ethics, from routine in-house trainings. All non-staff participants had had some prior contact with the programme, through their roles as KCRs, as guardians of study participants, through regular meetings with research staff (assistant chiefs) or through attending a KWTP open day (opinion leaders). A summary of participant characteristics is included in Table 1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Summary information for participants.

https://doi.org/10.1371/journal.pone.0127842.t001

The structure of the consultation included a half day workshop followed by a set of approximately four hour small group discussions within a two week period. Five workshops were held, each including between nine and 31 participants from one of the main groups described. After each workshop, participants were split into two to five small groups (see Fig 1). Workshops acted as a platform to share basic information about research, KWTP and ethical review processes, to introduce two hypothetical scenarios based on current practices for benefits and payments at KWTP and to begin discussions about these. The research scenarios, a home-based interview on health-seeking behaviour for childhood fevers and a facility-based Malaria Vaccine Trial (See Table 2), were used to support detailed discussion of study benefits and payments for very different types of research throughout the consultation. The small groups allowed for more in-depth, individualised exploration of issues and time for reflection on new concepts introduced in the workshop. During discussions, medical benefits, cash and ‘in kind’ payments and community benefits were particularly explored, reflecting types of benefits in common use at the KWTP. Facilitation encouraged debate within groups, including through cross-checking opinions across groups and the use of non-judgmental probes around emerging ethical issues. Discussions were held at venues convenient to participants in the language of choice (English, Kiswahili or local language). Details of the sampling and data collection methods have been given in greater detail elsewhere [10].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Flow chart representation of the two-stage consultation process (participant numbers in brackets).

https://doi.org/10.1371/journal.pone.0127842.g001

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Study Scenarios.

https://doi.org/10.1371/journal.pone.0127842.t002

Data management and analysis

At workshops, a note-taker made detailed notes of discussions. Small group discussions were audio-recorded, transcribed and translated into English by experienced staff fluent in local languages who had been present at discussions. Facilitators kept notes to capture the flow and dynamics of discussions. Following workshops and small group discussions, facilitators held debrief meetings to cross check perceptions and discuss emerging issues. Data were managed using NVIVO 10 and Microsoft Word applications, using coded identities. Data were analysed using a modified framework approach [16], involving in-depth reading of transcripts, making detailed summaries of discussions, developing a coding framework from major themes in the question guide and emerging issues in the transcripts, and coding data across all transcripts. Two set of analysis charts were developed to compare information across and within groups; and to chart changes of opinion and dynamics across debates. Analysis was primarily conducted by VM and MN with support from other authors, using an iterative process. The main inductive themes were perceived challenges for ‘too many’ and ‘too few’ benefits, the role of time and study procedures in perceptions of appropriate benefits and the implications of different types of benefit. Interpretation drew on assessments of the strength, nature and level of support for opinions within the consultation overall, and the literature on benefits and payments and research ethics in general.

Ethical review

This study was reviewed and approved by the Kenya Medical Research Institute (KEMRI) Scientific Steering committee and the Kenya Medical Research Institute (KEMRI) Ethical Review committee. Verbal consent was sought from workshop attendees, documented in minutes. Written consent was given for participation and audio-recording of discussions in the small groups. The use of both verbal and written consent procedures for this study were reviewed and approved by the Kenya Medical Research Institute (KEMRI) Scientific Steering committee and the Kenya Medical Research Institute (KEMRI) Ethical Review committee.

Provision of cash versus goods

There were mixed views on the relative appropriateness of cash and non-cash benefits, with the advantages and disadvantages of these often working as mirror images of each other.

Provision of medical services

Community-wide benefits

The main form of community-wide benefit currently provided by KWTP is strategic support to routine government health services in the county through a close collaboration with the Ministry of Health in Kilifi. As described in the introduction, these community-wide medical benefits include extra staff, supplies, equipment and infrastructure, and aim to ensure that good medical care is widely available regardless of involvement in research [8]. Most of these services are provided long term, but additional community-wide medical services are included in many studies during the lifetime of those studies. During our discussions, we provided information on community-wide medical services because their integration into routine care makes them relatively invisible to many in the community.

Community-wide medical benefits were viewed as an important way of ‘giving back’ to the community in return for their support to research, building and cementing the relationship between the community and the research organisation. At times, this was seen as a responsibility for researchers:

Even though research is voluntary, we should remember that this research will benefit the community and KEMRI staff…without the community, no work will be happening. So why shouldn’t you give benefits to people participating in your research? (P3F9: KCR Member, Male, 55y, 12y education)

In previous consultations with staff at KWTP, community-wide benefits had been particularly strongly promoted as a means of providing extra medical support in a largely low-income community without risking individual forms of undue inducement (Molyneux 2012). This was seen as a way in which relatively highly resourced researchers could respond ethically to global structural inequities reflected in low access to good health care in settings such as Kilifi. These views were underlined in some cases, but not all, by perceptions of the relative wealth of the research institution. These views were echoed by some in this consultation, with community-wide medical benefits seen as fair compensation in some instances, given the challenges described for cash and goods, but this was limited to situations where inconvenience was relatively minor, for example, short periods of time spent in an interview. Importantly, community-wide benefits—of any type—were not seen as reasonable individual compensation for longer periods of time, given the economic implications of time in this community. While study participants might often have sacrificed important economic activities, they would gain the same benefits as non-participants. This was seen as unfair and as a disincentive to participate in research:

…within that 1 hour [of research participation] you could have done something big…and the research you will be doing will not benefit this mother alone but many people. So the other many people will be using that 1 hour doing their own duties and then they benefit from that research. So I think the mother should be given something… (P1F13: Village Dispensary Committee Member, Male 58yrs, 12yrs education)

Information and communication as cross cutting issues

Across discussions, a frequently emerging issue was the importance of information and communication about study benefits, where the latter was seen as a centrally important feature of good practice.

Dilemma of providing cash or goods

In the literature, cash benefits have been argued as important in maximising choice for participants [20]. They have also been associated with increased risks of commercialization of the researcher-participant relationship [1, 2, 6], and—for research involving children—use in ways that do not benefit study participants [3]. Those who argue the importance of choice, as an autonomy issue, see removal of this option as an example of over-paternalism on behalf of researchers [2]. Arguments that cash payments should be avoided for the most-poor given increased risks of undue influence are also countered by recognition that withholding cash payments for this group would be unfairly discriminative [3].

Across this consultation, community stakeholders made similar arguments. Notably, in comparing cash and goods of the same value, cash payments were associated with an increased likelihood of many risks associated with ‘giving too many benefits’ [10]. The greater the financial value of the benefit, the more cash was seen to accentuate the risks involved. These included a higher risk of unacceptable degrees of influence on decisions to participate in research, and generating family conflicts and rumours within the community, particularly where cash payments in research were unfamiliar.

Amongst these risks, those of undue inducement and commercialisation, that is, that the relationship between researchers and participants would look like one based only on ‘buying and selling’, were particularly emphasised. These were also importantly countered by the importance of choice given by cash payments, particularly for the most-poor in this community. For research involving children, a strong fairness argument was made for the capacity of cash to allow mothers to share benefits between siblings, again prioritised for the most-poor. A challenge raised for cash for family dynamics was premised on gendered patterns seen as common within households in this traditionally patrilineal community; fathers ‘should’ traditionally have sole control over household resources, particularly cash. This view was controversial within these debates, on the basis of being unfairly discriminative towards mothers and unrepresentative of the community at large, particularly over more recent times. It also raises controversy in relation to wider principles of unfair discrimination.

A further challenge for cash payments was that the highly visible financial ‘face’ value of cash was seen as likely to prompt comparisons of benefits between studies and over time in ways that can create dissatisfaction with studies and the research programme. The actual amount to be paid as cash (unless for reimbursement of costs) was also seen as very hard to establish since appreciation would be closely influenced by levels of need. There would be a high risk that the same payment would act as an undue inducement in some families and seem insulting in others. Practical challenges for cash included security risks for staff carrying cash and risks of mismanagement by the same group.

Giving goods as study benefits was highly valued in this consultation, as long as these were useful. But goods were not devoid of the risks associated with cash, potentially working in exactly the same way to generate undue inducement, complaints related to comparisons made between studies and over time, rumours and family conflict. Further, and importantly, goods were seen as most appropriate as forms of appreciation or motivation for participating in studies, and less so as compensation for time where this has a significant economic value [10]. In this community, where poverty and subsistence forms of livelihood are relatively common, the economic value of time was associated with a capacity to earn enough money to feed families on a daily basis. Significant amounts of time taken in research should therefore be compensated with enough cash to buy food for the family that day, although some argued that equivalent amounts of foodstuffs (such as flour and dried beans) would be a reasonable alternative.

Comparing scenarios, a key influence on the risk of cash generating many of these problems was the site of the research interaction. Cash payments were seen as particularly sensitive if made during household visits, as compensation for extended periods of time (for example, lengthy interviews or group discussions), in which case goods were seen as more acceptable. Cash as compensation for time was less sensitive when given for attending research clinics outside the home, when transport costs were already being reimbursed. Transport reimbursement could clearly not be made using goods. To a large extent, these views reflected common understandings and expectations of study benefits. Cash payments are routinely made to reimburse travel costs, whereas compensation for time spent in research is a much less familiar concept. It seems that giving extra cash to compensate for time in addition to reimbursing travel costs is less visible than when time spent at home is compensated in this way. Skilled and consistent communication about the reasons for different forms of benefits and payments may lessen this risks, but giving cash payments at home should be seen—at least in this context—as potentially more sensitive.

Medical services as a benefit

In previous research on benefits and payments in Kilifi, many researchers and health managers placed a high value on providing medical services to study participants as a form of study benefit [21]. Providing medical services was seen as a way of balancing the benefits and burdens of participation that would have a lower risk of generating undue inducement or commercialisation than other forms of study benefit. Since the provision of medical services is a core activity in health research, additional advantages were seen in drawing on existing resources and skills.

Across this consultation, stakeholders saw the medical services provided by KWTP within studies as an important benefit for participants, particularly for research involving children for whom the acute illnesses often managed in this way (for example, malaria or pneumonia) were recognised as more common. KWTP services were generally seen as being of higher quality, more reliable and quicker to access than government health services [8, 22], although the latter were officially provided free of charge to children under the age of five years. Given these advantages, research-related medical benefits were widely seen as likely to influence many positively in deciding to participate in research, with highest risks for the most-poor [23].

At the same time, medical benefits—like goods—were seen as mainly being appropriate as a form of incentive or appreciation for participants, and not as a compensation for time spent in research, where the latter was economically significant. In addition, several pointed out that medical services would only act as a benefit if and when children became ill. Not only was this uncertain, but it was a circumstance that could not be hoped for by parents. Thus, as noted above, the time spent in participating in research should be separately assessed in considering appropriate study benefits overall.

A particular issue for medical benefits explored in this consultation was the issue of who within a family should have access to the services provided to young children participating in studies. This question had been generated during earlier research, when field workers at KWTP described perceived unfairness during interactions with families in which mothers or siblings of study children were unwell, but protocols allowed for medical services to be provided to a study child only [8, 24]. The same issue was recognised in this consultation, framed as a harm in generating additional burdens for mothers if medical services were limited in this way. If a sibling of a study child was unwell at the time of a scheduled research clinic appointment, mothers would have to find a way of seeking care for the sick sibling separate to keeping the research appointment. If the study child was also unwell, the dilemma would be greater since an option to skip the research clinic appointment (highly inconvenient to researchers) would be removed. The same situation would occur if the mother was unwell at this time. One strategy would be to take an older child out of school to assist. In any case, the mother would be forced into a position of prioritising the care of one child over another, widely seen as unreasonable.

The view that the interests of a study child should not be placed above those of other children in the family—described here for medical benefits and, earlier, for other forms of study benefit—emphasises that the research approach of placing a high importance on study children as individuals, separate from their families, does not fit well with local conceptions of good practice, reflecting similar positions in the literature [25]. It has been strongly argued and widely accepted that a ‘fair benefits approach’ in research should consider wider communities as potential beneficiaries as well as individual study participants [26]. From our findings, it can also be argued that researchers should consider the position of a study child’s family in planning fair benefits. Involving children in research inevitably engages other members of the family. At the same time, mothers, who are generally (in this setting) responsible for taking care of children, must be equally concerned about the welfare of all their children. On this basis, researchers have a responsibility to ensure that their activities do not compromise the position of families they engage with. The suggestion made in this consultation that researchers should at least offer urgent medical services to siblings or caretakers who are unwell at the time of scheduled clinic visits, would fit with meeting this responsibility.

Community wide benefits

The ‘fair benefits approach’ described earlier focuses on the importance of taking account of wider contextual issues that generate ethical challenges for researchers, rather than focusing narrowly on a balance of benefits and burdens to individual participants [26, 27]. By this account, a wide range of types of benefits and beneficiaries—including community-wide benefits—are important for researchers to consider. Particularly for international research in low-income settings, this argument identifies a need for researchers to act in ways that contribute to diminishing existing global inequities in access to health and other resources; community-wide benefits are one way in which this can be achieved. Community stakeholders in this consultation supported this view, seeing community-wide benefits as an important way of researchers’ ‘giving back’ to the community, acting as a form of appreciation and recognition of unmet health needs [8, 17]. This form of benefit was seen to build trust and good relations between researchers and the wider community. At the same time, community-wide benefits were not seen as a substitute for individual study benefits, including those needed to show appreciation, motivate or compensate participants for their time or for inconvenience. As described in the literature, individuals may not be willing to take up the risks and inconveniences of research participation for the sake of benefits that will be enjoyed by the wider community [27].