Scenario 1: A home-based interview on fever treatment-seeking behavior.

A field worker visits a home to request that a mother of three young children participate in a study involving a one hour interview on her beliefs about the causes and treatment of fever in children. The mother is offered an opportunity to choose the time for the interview, and the field worker approaches the home respectfully and explains the study carefully.

• Do you think there is a need for this mother to be given anything for her participation in this study? Why do you think this and what should that be?
• What if the interview were longer, say 3 hours?
• What if it included a finger prick sampling of blood?
• If more benefits are seen as important, what role for community vs individual benefits?

Scenario 2: Benefits and payments in a Malaria Vaccine Trial.

A mother has consented to her child taking part in a Malaria Vaccine trial. This means that it is not known whether this vaccine will work against malaria or not, but the vaccine has been tested in earlier trials and found to be sate and likely to be effective. A blood sample will be taken to test the health of the child before the vaccine is administered; the child will receive a health check after the vaccine is administered; follow up at home for 6 months by a KEMRI Field worker, including 6 visits in total, some of which will include further blood tests; if the child is sick at any time during the study they will receive free treatment; and when needed by the study team, transport will be reimbursed. A study clinician is assigned to the dispensary to attend to all sick children and other community members as a way of giving back to the community.

• What do you think about these benefits and payments?
• Do you think there is a need for this mother to be anything else for her participation in this study? Why do you think this and what should that be?
• What if the trial included adults, not children?
• What differences are there between this situation and Scenario 1?

Subsequent small group discussions supported more in-depth discussion, checking and extending workshop scenarios, ensuring contributions from all participants, and allowing time for reflection between meetings. The method drew on an approach used previously in Kilifi to undertake community consultations on disclosing genetic findings [19]. In contrast to a more typical focus group discussion [20], we built in greater involvement of facilitators in directing the discussion and using probes to support individual and group reflection and debate. Facilitators aimed to explore the views of all participants as far as possible, use non- judgmental probes to explore reasoning and promote reflection, particularly for morally relevant issues, and avoid consensus building, in keeping with substantive forms of deliberative ethics [21], [22]. Discussions were held at venues convenient to participants, in languages of choice (English, Kiswahili or local language) and took approximately 4 hours, including a break for refreshments. Following the usual practices for community engagement activities, non-staff participants were compensated for time spent in these discussions, at a rate of approximately $3.5, along with reimbursement of transport costs.

All authors supported aspects of facilitation in this consultation, with non-staff community group discussions facilitated by MN, FK and SMw. FK and SMw are experienced community facilitators from this community; VM and SMo have been resident in Kilifi for more than 18 years, and MN for 6 years.

i. Too much influence on individual decisions.

While the inclusion of payments and benefits were seen as a reasonable influence on people's decisions to participate, there was widespread – but not universal - concern about the potential for high levels of benefits to act as a ‘problematic’ form of encouragement. In this situation, the ‘problem’ was mainly related to an unwanted effect on understanding and free choice, often linked to perceptions of the complexity and unfamiliarity of research in this community. People might not think enough about what was involved in participation, but make premature decisions based on an ‘offer’. Alternatively, people who had thought about the study but not fully understood might decide to participate anyway, again based on the ‘offer’.

…when the benefits are too much …it gets to a point where it's like it's a buy off, you don't have to decide, the price is too good to reject… (Field worker, male, 29 y)

These concerns were voiced in all groups, but most strongly amongst staff community members. The reason that ‘understanding’ was seen as important often seemed to be related to an intrinsic value or a right to make free decisions:

So even though there is not really you know taking advantage of the people per se but the fact that it makes people not to think, you know, to consider all the pros and cons of participating in the research, already that's infringing someone's right… (Community facilitator, male)

Some staff community members also saw understanding as important in supporting cooperation in studies, and - less often – in protecting the safety of participants, who could choose not to participate if they felt risks were too high. In general, issues about safety were not prominent in discussions, a feature likely linked to information given in preliminary workshops about the role of ethical review in checking that risks to participants were reasonable.

In common with many challenges discussed in this consultation, those seen as most at risk of being too easily influenced were the most-poor, thought likely to be most attracted by benefits and payments. At the same time, highlighting debates in the literature, for those who felt that a loss of ‘free choice’ was not particularly important, their reasoning almost always concerned the greater importance for the most-poor of getting high levels of benefits, given the importance of their unmet needs.

ii. Undermining household relations.

High levels of payments and benefits were also seen as potentially undermining household relations. In this traditionally patrilineal setting [24], male household heads generally are key decision makers in how family resources are used. While there is much diversity in the way household dynamics work in practice, through individual socioeconomic, cultural and temporal influences [25], in many households married women gain access to significant levels of financial resources through their husbands. Giving benefits and payments to mothers of child participants was seen to risk creating family disharmony if levels were ‘too high’. These risks included undermining fathers' traditional control of household resources; generating mistrust between parents; and that fathers taking control of cash benefits might not prioritise the participant child's interests. Concerns about marital disharmony emerged particularly strongly in specific non-staff community groups, but were recognised as a challenge in many others:

In the norms and traditions of the Mijikendas…men believe that a woman cannot provide things for herself. He believes that he is the sole provider for the woman… when she [mother] gets home the husband might think…it's not… an issue of research but he [KEMRI field worker] had his hidden agenda (Rural assistant chief, female).

iii. Challenges for community relations and values.

A common and strongly felt concern about high levels of benefits was the risk of creating conflict within communities, based on jealousy emerging between participants and non-participants [26]. A particular risk was that research selection processes would be seen as biased, leading to rumours of clan preferences:

But if we look at malaria research, none of my children have suffered from malaria but there is a neighbour whose 2 or 3 children often suffer from that illness… if… its only me who was selected… the person with sick children will feel bad because…hers are always sick, always going to the hospital so why can't she be selected? So then there will be a sort of hatred in that homestead (Rural assistant chief, female).

In addition, the problem of undermining voluntariness in research was linked by some staff and non-staff community members to the likelihood that people would lose commitment to a spirit of voluntarism, or social cooperation, fundamental to many aspects of community life. The concern was that giving payments and benefits at too high a level to study participants could ‘kill the spirit of volunteering’ by generating an expectation of reward for contributions to community development projects. In this case, it was often felt that any benefits given should act as tokens of appreciation, rather than ‘payment’:

What I feel is not right is when you get someone used to receiving money. It kills the volunteering spirit and that spirit is very important…When the chief calls us for a meeting….once they find out maize will be distributed a lot of people show up, but if they are told it's [just] a meeting no one comes. So I don't support constant giving of tokens except when one volunteers on their own. [Later] Because of the money [payments to study participants] progress will be hindered (Youth leader, male, 48 y).

As a further adverse impact on the community, high levels of payments and benefits were viewed as generating the risk of creating a sense of dependency amongst study participants which could, if continued over long periods of time, disrupt economic livelihoods of families and even the wider community (for very large studies). Adding to this risk was the potential for those in receipt of benefits to be looked down upon within the community as people who ‘live on aid’.

iv. Impact on the researcher-community relationship.

In situations where the level of payments and benefits were seen as ‘too high’, there was a widely perceived risk of creating suspicions about the motivation of researchers and research institutions, and risking breakdown in trust. Where there were pre-existing concerns and rumours about research activities, for example, based on misconceptions about the purpose of research [27] or hidden commercial interests of researchers, giving high levels of benefits and payments was seen as particularly risky for the institution's reputation:

…giving something like …a plastic mug…depending on what the research has planned… that sounds better to me than giving money because giving money, I tell you, to an African, its directly linked to buying (Field worker, male, 50 y).

These effects of commercialising the researcher-participant relationship were particularly marked for - but not exclusive to - cash payments. A particularly problematic effect of this ‘transactional’ view of the relationship was seen for research involving children:

They will not be doing it out of their willingness to participate…they will almost force themselves to participate because they know if their child participates they will benefit. So that will be like selling this child to receive monetary gain and that will be like a business investment (Mother of study child, 40 y, 12 y education).

One perceived consequence of a shift in the researcher-community relationship towards commercialisation was the setting up of expectations which might be difficult for researchers to meet in future, particularly for organisations with fewer resources. Countering this view, others felt all researchers should be obliged to plan their research with the same reasonable levels of payments and benefits. As a last - and contested - point, a few expressed concerns that commercialising the relationship would take a sense of ownership of the research away from participants and the wider community.

v. Negative impacts on research.

Three main issues were raised about risks to research through setting benefits and payments at too high a level. Firstly, staff and non-staff community members perceived that people considering participation might falsify information in order to become eligible for studies:

If people come to know that there are too many benefits…people who do not meet the study criteria will look for any means so as to join… and if she fails to join, she will envy the others who joined (Urban assistant chief, male).

Secondly, an issue - particularly raised by staff community members - was that over time staff responsible for informed consent would become de-skilled, since a pressure to communicate well would be removed by many people's keenness to participate. Conversely, reasonable levels and forms of benefit to study participants were seen as an appropriate means of supporting the work of this staff group through encouraging recruitment.

Finally, some community members reflected on a risk that giving ‘too many’ benefits would eventually be reflected in the cost and sustainability of research, and therefore the products of research, including that ultimately the vaccines and drugs produced might be unaffordable in low-income countries.

i. Unfair economic and financial burdens to families.

The risk that ‘too few’ benefits would generate unfair burdens for families was the most serious challenge described in these consultations, and the most strongly and widely expressed concern. Burdens included inadequately taking account of money spent (as a direct cost) and of lost opportunities to earn income or undertake other essential unpaid tasks (indirect costs), including child care, house repairs and subsistence farming [28]. Socioeconomic status was a major influence on the risk of unfairly burdening participants, with potential risks seen as more likely and more serious for the most-poor, but which category included a high proportion of the population overall.

While direct costs, for example transport refunds in the Malaria Vaccine Trial scenario, were clearly noticed and more easily accounted for, discussions on the importance and ways of compensating for indirect costs involved much more controversy, particularly in accounting for time. There was a common concern about the risk of fundamental forms of unfairness when time was not adequately accounted for on the basis of having unpaid or informal livelihoods:

… if we're basically saying we are not compensating this person because they… are losing nothing anyway, are we not already taking advantage of their vulnerability in the sense that we know they are not… losing anything and therefore we have no responsibility to give them anything, so we can just take that opportunity to get…that information? (Community facilitator, male)

Views on the role of time were explored by increasing the length of an interview from 1 to 3 or more hours in research scenario A, and considering time spent at and travelling to a study clinic in Scenario B. The central emerging issue concerned the difficulty in both noticing and accounting for costs associated with essential unpaid work and lost income, and the uncertain role of time as a parameter in assessing these, particularly given the prevalence of poverty and the nature of subsistence livelihoods:

[In Kilifi] poverty is the same everywhere. So for that 1 hour…she will think it's better if she had gone to fetch firewood… She might not have anything to cook but she could use firewood to boil water for bathing…a normal person cannot fail to have something to do at home (Christian youth group leader, 29 y, 8 y education)

Common livelihoods were described as essential unpaid domestic work, subsistence farming and paid casual labour, including piece-meal work (‘vipande’ in Kiswahili). Time taken from paid work generates a clear indirect cost, but payment in cash or in-kind would also often be involved where a family member or neighbour was asked to take over essential unpaid work. The critical point about compensating for time in relation to these livelihoods was that a daily income or food production was often essential to a family's wellbeing, through the ability to ‘put food on the table’ at the end of the day. Indicative of this perception, estimates of the levels of payments and benefits were often focused around resources needed to provide a family meal. Duration of time was generally seen as important; shorter periods of time were seen as less in need of ‘compensation’ because they would be likely to incur less economic costs. For periods of about an hour, many considered that ensuring flexibility (for example, by booking appointments for interviews), giving small tokens of appreciation or simply showing respectful attitudes and communicating carefully would often be enough, assuming participation was truly voluntary. The inclusion of some learning in the interaction would also limit the need for additional ‘benefits’. For periods of time of 3 hours or more, most felt some form of compensation was important, as economic costs would very likely be incurred, with particularly serious implications for the most-poor:

In my view, it's important that the benefits are increased [with time] because there will still be difficulties that day…even if the transport expenses were refunded, and if they were meant to do some casual work in order to get some food after the visit to the centre, they would probably sleep hungry that day because there is nothing that fills the gap for the day spent at the centre (Christian youth group leader, 29 y, 8 y education)

Some non-staff community members disagreed, arguing that shorter and longer periods of time have similar requirements for payment or other benefits, particularly for the most-poor where even small sums of money could buy valued goods like a container of water. There were also challenges in using amounts of time (for example hourly rates) as the only parameter of indirect costs of participation. For casual labour, hiring is often linked to availability at key points in time such as the start of the day. Missing the hiring time would exclude the possibility of earning cash that day:

…especially for the poor families…once you use their time, on humanitarian grounds…even the child may be hungry and then you leave without offering the mother anything. After that the mother goes to look for casual jobs only to be told that there are no vipande remaining and then she goes back home while you still want to participate the child in research, it will be of no help (Village Dispensary Committee member, male 58 y, 12 y education).

Comparisons were made with social norms in relation to visitors, being described as norms of visitors both giving and receiving. An underlying understanding in relation to these norms was that (within reason) exact amounts of time spent would not be seen as particularly important for visitors, but rather who came and what was discussed:

OK, with people in our communities… they normally don't weigh the time spent with them. What they'd say is, they came (all laugh) yeah they came…So ok, it doesn't make any difference whether you go there and spend like two, twenty minutes or an hour with them…they don't consider the time…it's your presence there. (Field workers, male, 21 y)

ii. Low motivation to participate & research failure.

Since most agreed that payments and benefits were an important motivation to people's decisions to join studies, giving too few benefits was seen as risking research failure. A potential ‘community-wide effect’ was described where word spread that ‘very few benefits’ were associated with a particular study. Staff and non-staff community members were concerned about the knock-on risks of low participation in studies, including failure to develop new treatments and health services and to sustain local employment opportunities:

… many people will drop out from the study and the study will not succeed… probably you will be laid off, it will be chaos, diseases like malaria will still be there… on the ground people will suffer, they will fall sick…(Urban assistant chief, male)

More subtly, low levels of motivation might not be very visible to research staff but still have important effects on research and on the researcher-community relationship:

They will think their time is not used appropriately…you will find you are interviewing her while she continues cleaning utensils, you talk to her but she keeps quiet cleaning her utensils and she bids you goodbye cleaning her utensils!(Laughter) (Rural assistant chief, male)

A specific concern in relation to risks of low motivation was that while ‘low benefits’ are likely to discourage participation in general, this effect might occur inequitably across a community. This issue was also seen as a form of ‘double unfairness’ in relation to free choice for the most-poor, who might be influenced to participate in studies at levels of payments and benefits that others in the community might even consider exploitative or insulting (that is, being too poor to refuse even a bad offer). Community facilitators described a resulting unfairness in who carries the burden of participation and related risks of bias in research findings that could impact validity.

iii. Undermining researcher-community and family relations.

As above, where benefits and payments set at ‘too low’ a level appeared insulting, there was a risk of undermining the relationship between researchers and the community, leading to suggestions by staff members that a minimum value should be set:

… if we were to visit someone and we give him 100 shillings - would they appreciate that? Or you would even create more problems for yourself? Because I think there is a level where you do things in the community and they feel you have actually…undermined them, they wouldn't appreciate it (Community facilitator, male).

A final point was made about the way that setting benefits and payments at too low a level could have a negative effect on relationships within families, through disagreements on how limited family finances should be prioritised between research and other activities.