As outlined in the previous section, gatekeepers in CRTs have fulfilled a variety of roles involving the protection of individual, cluster, and organizational interests. Unexpectedly, perhaps, the question of whether gatekeepers possess the legitimate authority to fulfill these roles remains unexamined. We believe that answers to this question are essential, because gatekeepers make decisions that have consequences for others. Gatekeeper proxy consent for cluster members may allow a study to proceed, thereby exposing people to research risks without their individual informed consent or in other cases, gatekeeper refusal to provide cluster or organizational permission may bar access to a potentially beneficial study intervention. Below we undertake an ethical analysis of the authority of gatekeepers to play the roles described in the previous section.
Gatekeeper authority to protect individual interests
Previous work by our group has significantly restricted the need for gatekeeper proxy consent for cluster members and permission to randomize cluster members. First, gatekeeper proxy consent for cluster members is not required when cluster members are not human research subjects. In our paper in this series addressing the identification of research subjects, we present a novel definition of a human research subject as ‘an individual whose interests may be compromised as a result of interventions in a research study’ [
5]. We suggest that when people are affected only indirectly by cluster-level interventions (and they are not otherwise subject to intervention or interacted with, and their private health information is not collected), they are not human research subjects, and thus their informed consent is not required. For instance, knowledge-translation studies commonly intervene in health-professionals practices to align practitioner behavior with evidence-based treatment guidelines. In these CRTs, patients may be affected only indirectly by the study intervention and, insofar as the health professional’s practice is brought into line with evidence-based standards, there is no risk of it affecting their interests adversely. When patients are merely affected indirectly by the study intervention, they are not human research subjects, and neither individual informed consent nor the proxy consent of a gatekeeper is required.
Second, gatekeeper proxy consent on behalf of cluster members is not required when a research ethics committee approves a waiver of consent. In our paper on informed consent, we considered circumstances in which it may be difficult to obtain the informed consent of cluster members because of cluster-level interventions or large cluster size [
6]. We suggest that if the research would not otherwise be feasible and the study involvement poses no more than minimal risk to the individual cluster members, a research ethics committee may reasonably approve a waiver of consent, meaning that individual informed consent is not required. When the requirement of informed consent has been waived, it is unnecessary for a gatekeeper to provide proxy consent for cluster members.
Third, gatekeeper permission to randomize clusters is not required so long as cluster members are approached for consent as soon as possible and before any study interventions take place. In our paper on informed consent, we pointed out that the purpose of informed consent is to allow research subjects to adopt the ends of the study as their own, thereby (partially) justifying exposing subjects to risk for the benefit of others [
6]. In individual-cluster trials in which clusters are randomized before the individual participants can be approached for their informed consent, it has commonly been assumed that gatekeeper permission for randomization is required. We suggest that, on the contrary, so long as cluster members are approached for informed consent as soon as possible, and before study or data-collection interventions have taken place, the moral purpose of informed consent may be fulfilled. Under these conditions, cluster members have the opportunity to adopt the ends of the study as their own and, crucially, they may decline study participation before they are exposed to any risk for the benefit of others. Thus, in these cases, gatekeeper permission to randomize clusters is not required.
The question remains whether gatekeepers may legitimately provide proxy consent for cluster members in studies that do not qualify for a waiver of consent or in which individuals are not approached for consent as soon as possible after randomization of the cluster and before study or data-collection interventions have begun. Generally, in the ethics of research, a proxy decision-maker is called upon to provide informed consent on behalf of someone else when the prospective research subject is incapable, that is, when the subject lacks the requisite cognitive capacities to provide consent themselves [
41]. In such instances, the next of kin usually serve as the proxy decision-maker, on the grounds that the next of kin are likely to know the prospective subject well and be naturally motivated to promote the subject’s welfare. A proxy decision-maker is responsible for making decisions that are in accordance with the subject’s previously expressed wishes and values, or that are consistent with the individual’s best interests [
41]. Further, proxy decision-makers must strive to avoid any conflict of interest.
The circumstances in which gatekeepers might provide proxy consent in CRTs are dissimilar in important ways from the circumstances in which proxy decision-making typically occurs. In contrast to the conventional situations in which proxy consent is required, cluster members are often competent and capable of freely choosing whether a study is in accordance with their interests and values. Insofar as cluster members are competent, gatekeepers would only have the authority to provide proxy consent if the cluster members had autonomously authorized them to do so, and we take it that this is rarely the case. Further, gatekeepers typically have neither a close personal relationship with cluster members, nor a detailed knowledge of their individual wishes, values, or interests. This suggests that gatekeeper proxy consent for cluster members is doubly problematic. The provision of proxy consent on behalf of a cluster member who is competent, and who has not autonomously authorized the gatekeeper to make such decisions, violates that person’s autonomy. Further, the conditions that confer legitimacy on a proxy decision-maker, such as understanding the interests of cluster members, generally do not apply. Thus, gatekeepers are not legitimate proxy decision-makers for cluster members.
We conclude, therefore, that gatekeepers generally do not have the authority to provide proxy consent on behalf of individual cluster members. As a result, gatekeepers should not provide permission to randomize cluster members nor should they provide proxy consent for cluster members (Table
2). Studies that do not qualify for a waiver of consent or in which individuals are not approached for consent as soon as possible, and before study or data-collection interventions begin, should not proceed on the basis of proxy consent from gatekeepers.
Table 2
Summary of recommendations for the appropriate use of gatekeepers in cluster randomized trials (CRTs)
Gatekeepers should not provide proxy consent on behalf of individuals in CRTs | The fact that cluster members are typically competent and gatekeepers do not have detailed knowledge of cluster members’ decision-making history, interests, and values undermines the legitimacy of gatekeepers as a proxy decision-makers |
Gatekeepers should not provide permission to randomize or proxy consent on behalf of cluster members, and CRTs should not proceed on the basis of such permission or proxy consent |
When a fiduciary relationship exists between the gatekeeper and cluster members, as in a physician–patient or teacher–student relationship, the gatekeeper may provide permission to approach cluster member | Gatekeepers who are fiduciaries may deny permission to approach cluster members whose interests are likely to be unduly compromised by study participation |
Gatekeeper permission to approach cluster members is not appropriate where no fiduciary relationship exists between the gatekeeper and cluster members |
When a CRT may substantially affect group-based interests, and a gatekeeper possesses the legitimate authority to make decisions on behalf of the cluster, gatekeeper permission to enroll the cluster in the trial should be sought | When a gatekeeper possesses legitimate authority with respect to the individuals involved and the decision at hand, the gatekeeper’s permission to enroll the cluster in the study should be sought |
Ambiguity about the authority of a gatekeeper may be reason for consultation with cluster members |
When a gatekeeper does not have the requisite authority, researchers should not approach the gatekeeper for permission to enroll the cluster in research, and a CRT ought not proceed on the basis of such permission |
Cluster permission does not supplant the need for individual informed consent from cluster members |
When a CRT may substantially affect group-based interests, researchers should seek to protect these interests through cluster consultation to inform study design, conduct and reporting | Cluster consultation may be used to seek input on how the CRT ought to be conducted so as to enhance study protections and benefits for clusters |
Mechanisms may include open public forums, meetings with opinion leaders, presentations at religious or civic organizations, and the use of radio, television, or the internet |
Recommendations from cluster consultation are not binding and, where there are good reasons to do so, researchers may decline to make suggested changes to a study |
When a CRT may substantially affect organizational interests, and a gatekeeper possesses the authority to make decisions on behalf of the organization, organizational permission should be sought from the gatekeeper. | Organizational interests may be separable from cluster interests in a CRT |
The gatekeeper will consider the effect on the organization, including availability of staff, financial implications of participation, and compliance with organizational policies |
| Organizational permission does not supplant the need for individual informed consent from cluster members |
Is there any role for gatekeepers in legitimately protecting the interests of research subjects in CRTs? We believe so. In certain circumstances, the gatekeeper may legitimately provide permission to approach research subjects, and this may protect individual interests in CRTs (Table
2). When the gatekeeper has fiduciary obligations to individual cluster members, as in a physician–patient or teacher–student relationship, the gatekeeper may be viewed as having an obligation not to allow researchers to approach a cluster member whose interests are likely to be unduly compromised. Thus, gatekeeper refusal from physicians or teachers for researchers to approach individual patients or students is an instance of the legitimate protection of individual interests.
When the relationship between gatekeeper and cluster members is not fiduciary in nature, the gatekeeper does not have the authority to be relied upon to protect the interests of the cluster members. In these situations, the gatekeeper may identify potential research subjects, but may not grant permission to approach individuals. The gatekeeper acting in this role should not be relied upon to protect the interests of the subject For instance, consider the role of the index member in the aforementioned study of HIV prevention among Roma men. Identification of cluster members may be an important role pragmatically in the conduct of the study, but the researchers and research ethics committees should be clear that it does not involve the protection of individual interests.
Gatekeeper authority to protect cluster interests
The wide variety of groups studied in CRTs presents a challenge to those trying to determine who has legitimate authority to represent and protect the interests of a cluster. CRTs may involve a variety of clusters, with varying degrees of cohesiveness; clusters may be sports teams, classrooms, nursing homes, work sites, primary-care practices, geographical areas, villages, or other communities (Table
1). Cluster interests may include the preservation of the identity of the group and the maintenance of the integrity of social structures. CRTs vary in the degree to which group-based interests may be affected by cluster participation. In some cases, such as knowledge-translation studies seeking to promote adoption by physicians of evidence-based guidelines, few group-based interests may be implicated; in other cases, implications for the cluster may be substantial. For instance, studies investigating the genetic determinants of breast cancer in Ashkenazi Jews may protect research subjects by maintaining their anonymity, yet the naming of the community in publications may lead to the perception that Ashkenazi Jews as a group are more susceptible to breast cancer and discrimination [
39].
In certain cases, CRT participation may substantially affect group-based interests yet protecting these interests may be difficult. Clusters often do not have organized structures or legitimate authorities capable of speaking on their behalf. Further, where organized structures for representative decision-making are present, these structures may not have been established with the intention of making decisions about research participation [
2]. Finally, as Hutton points out, a representative may refuse to adopt the functions of a gatekeeper for a cluster [
4]. These situations therefore raise the question, as stated by Hutton: [
4] ‘to whom should responsibility for the decision to enter the cluster be passed’?
The debate in the community-based research literature over a representative’s authority to provide permission to enroll a community in a study may provide some insight into a gatekeeper’s authority with respect to clusters. As discussed in the introductory paper in our series, recognition of community-based interests and the moral status of communities led to the formulation of the principle of respect for communities [
1]. According to Weijer and Emanuel, community permission is appropriate if ‘the community has a legitimate political authority, which could be a legislative assembly, mayor, or tribal council, that has the authority to make binding decisions on behalf of its members’ [
39]. The principle of respect for communities relies on identification of community or group characteristics to help determine when community consent is required and when consultation alone is appropriate. Importantly, community permission is not a substitute for individual informed consent.
When a CRT involves a well-defined community, and when the CRT may substantially affect group-based interests, the protections required by the principle of respect for communities may be applied directly. In such cases, researchers may be required to obtain permission from gatekeepers to enroll the cluster, prior to seeking individual informed consent from cluster members (Table
2). A gatekeeper may give permission for the cluster to participate in the study if they have legitimate authority with respect to the individuals involved and if their authority extends to the decision at hand. Whether the gatekeeper has legitimate political authority depends on whether, among the individuals who are significantly affected by the gatekeeper’s decisions, there is widespread satisfaction with the gatekeeper’s ability to make such decisions.
A variety of mechanisms may be used to determine whether the gatekeeper’s authority is legitimate. For instance, researchers may consult with the leaders and people in the group to understand the social dynamics of the group, and members’ satisfaction with the gatekeeper’s role. Consider the example of a community-based CRT investigating a breastfeeding education program in India, which sought consent from community leaders and health-system authorities to include their communities in the study [
20]. Whether those who acted as gatekeepers in this trial had the authority to provide cluster consent depends on two conditions: whether the members of the community understood the gatekeepers’ roles as including the authority to make these decisions, and whether they were largely satisfied with the institutions involved, that is, the political system used to select community leaders and the local health system. Individual satisfaction with these institutions will be based not only on past decisions that have been made by the institutions involved, but also on the decisions at hand.
If it is unclear whether a gatekeeper’s authority encompasses decisions about participation in health research, researchers should determine which method of proceeding would be most likely to be be conducive to the cluster members’ satisfaction with the institutions involved. It has been reported that the views of community leaders are generally poor substitutes for the views of individual community members [
42]. This suggests that ambiguity about whether a gatekeeper’s role includes making a decision about participation in a particular CRT might be a reason for consultation with cluster members. Cluster consultation is necessary because it increases the likelihood that cluster members will be satisfied with the institutions involved, and the continued legitimacy of the institutions involved depends on the satisfaction of the people whom they purport to represent.
When CRTs involve clusters other than well-defined communities, gatekeepers typically will not have legitimate authority to speak on behalf of the cluster. In such cases, researchers should not approach gatekeepers for permission to enroll the cluster in research in order to protect group-based interests, and a CRT should not proceed on the basis of such permission (Table
2).
Cluster consultation is a second mechanism for the protection of group-based interests in CRTs. As we have seen, cluster consultation may usefully augment cluster permission, particularly in cases in which there is uncertainty about the gatekeeper authority. When a CRT may substantially affect group-based interests, cluster consultation may usefully and legitimately protect these interests even in the absence of a legitimate political authority (Table
2). Thus, compared with cluster permission, we understand cluster consultation to have a broader application to CRTs as a protection for group-based interests. Cluster consultation involves a partnership between researchers and community members, from research design to publication [
43]. The degree to which a cluster can participate will depend on community characteristics and cohesiveness [
39]. Cluster consultation may be sought when the cluster has a common history, common culture, or other characteristics that provide cohesiveness to the group. In these cases there are several aspects of the research endeavor in which cluster members may take part, including consultation over protocol development, involvement in the conduct of research, dissemination of information, and publication of results [
39].
The diversity of social groups in CRTs poses a practical challenge to effective cluster consultation. Dickert and Sugarman have usefully described the goals of consultation as enhanced protection, enhanced benefits, legitimacy, and shared responsibility [
44]. Possible means of soliciting feedback on a study include open public forums, meetings with opinion leaders, presentations at religious or civic organizations, and the use of radio, television or the internet. As previously stated, cluster consultation differs from cluster permission in that researchers are seeking input on how the study should be conducted, not whether it ought to be conducted. According to Dickert and Sugarman ‘it would be disingenuous to enter into a consulting arrangement where the consulting party does not intend,
ex ante, to take the consultants’ advice. If relevant consultants have strong negative reactions or endorse particular modifications, those reactions or modifications have significant moral force and warrant respect and careful consideration’ [
44]. Although they are morally weighty, recommendations from cluster consultation are not binding, meaning that ‘investigators may sometimes justifiably act contrary to such opinions’ [
44].
The gatekeeper function of protocol approval relates to the practice of cluster consultation, insofar as cluster members may participate in all stages of the research process, including providing feedback on protocol development. The partnership model requires cluster feedback to ensure the appropriateness of the study for the clusters involved. To this end, research ethics committees, whose mandate is to review study design and protocols, include community representatives among their members. Cluster approval of the study protocol may be appropriate when the research ethics committee is not representative of the community in which research is taking place.
Gatekeeper authority to protect organizational interests
Another important role of gatekeepers is the protection of organizational interests. Gatekeepers may protect or promote organizational interests by providing permission in the name of an entire organization, such as a hospital, nursing home, or school, to participate in a study (Table
2). A gatekeeper’s agreement to allow an organization to participate in a CRT may provide opportunities for individuals within that organization to participate in and benefit from research, but a gatekeeper’s refusal of permission may mean that individuals within the organization may be denied participation in potentially beneficial research. Thus, the interests of the organization may conflict with the interests of individuals within it. An organization’s administrators and managers will be guided by their legal and professional responsibilities to act in ways that promote the safety and privacy of their members, and promote the proper functioning of the organization itself.
Although it may be useful to conceive of these two sets of obligations as distinct, it should be noted that they are not entirely independent of one another. Insofar as being part of the organization serves the interests of its members, then presumably, serving the interests of the organization also goes some way towards serving the interests of its members. Gatekeepers acting on behalf of organizations may legitimately make the decision about whether the organization will participate in a CRT because they are considered to be well situated to judge organizational interests. This may simultaneously serve the interests of those within the organization. Nevertheless, although an organizational gatekeeper may decide to allow researchers to have access to individual employees, such permission is not a substitute for individual informed consent.