Results of literature screen

A total of 2027 candidate titles were identified through database searches, supplemented by reference list and website reviews. Of these, 1721 did not meet eligibility criteria, and 306 were included for abstract review. If the abstract was not available but full text was, the title was included for full-text review. After abstract review, 50 duplicates were found, 1 was not in English, 7 were not available (abstract nor full text) and 170 abstracts did not meet inclusion criteria. Seventy-eight articles were selected for full-text review, of which 24 subsequently did not meet inclusion criteria, 1 was in French, 1 was a duplicate and 2 did not have available full text. Therefore, 50 full-text articles were included in this review (figure 1, table 1–8). Of the articles excluded at the abstract and full-text review stages, the most common reasons for exclusion were: no mention of the informed consent process for research with paediatric or adolescent populations; research not taking place in an LMIC; and articles on paediatric research in LMICs that did not discuss the informed consent process.

• Download figure
• Open in new tab
• Download powerpoint

Figure 1

Results of literature screen. Flow diagram depicting results of the literature search and review procedure.

Summary of guidelines and commentaries

We identified seven guidelines that addressed issues of informed consent in international settings and in research involving children in our scoping review. Of these, we selected for detailed review five that were most comprehensive, summarising key recommendations in table 1. All guidelines emphasise the importance of obtaining individual, informed and voluntary consent for research.3 4 7–9 Importantly, however, the guidelines do not necessarily specify in detail how best to operationalise these core principles. For example, the Declaration of Helsinki comments only that informed consent requires that a subject be adequately informed of the ‘aims, methods, sources of funding, any possible conflicts of interest, institutional affiliations of the researcher, the anticipated benefits and potential risks of the study and the discomfort it may entail, post-study provisions and any other relevant aspects of the study’ (Article 26).7 Similarly, on voluntariness, the CIOMS guidelines note only that consent is voluntary if ‘an individual’s decision to participate is free of undue influence’ (p 35).3

Some of the guidelines do suggest modifications appropriate for lower resource settings, such as obtaining witnessed verbal consent when literacy is a barrier.7 9 The US National Bioethics Advisory Commission also acknowledges that oral consent might even be preferable in some circumstances.8 However, as other commentaries note, there is little specificity on how best to operationalise these suggestions, such as how to formally document verbal consent or characteristics of a qualified witness.10 11

Another important consideration of LMIC research addressed in guidelines is an emphasis on the need to at times obtain consent from community stakeholders and leaders, or other key local decision makers. Nevertheless, all guidelines unanimously assert that community-based consent can never replace individual consent. When local cultural practices around community-based consent contradict core principles of the international consensus on the informed consent process, such as the need for voluntary individual consent, researchers are advised to search for culturally sensitive ways of providing all information to potential participants without compromising the substantive ethical standard of informed consent, an adaptive process in which local research ethics committees are expected to place a substantial role.8 10–12

Finally, with respect to children or adolescents not capable of providing informed consent, in addition to obtaining consent from parents or legal representatives, most guidelines also reinforce the need to obtain assent from the child or adolescent in an age-appropriate way.3 4 7 9 The CIOMS guidelines note that assent is ‘a process…not merely the absence of dissent’ and requires ‘meaningful engage[ment] in the research discussion in accordance with…capacities’ (p 67).3 They also note that as adolescents reach the age of maturity, their agreement to participate may be ethically considered as informed consent. However, if they legally remain minors, researchers are cautioned that consent from a parent is still generally needed, but a list is provided of possible situations when parental consent might be waived, such as with legally emancipated adolescents, or under circumstances where obtaining parental consent is not desirable because of the research topic.3

Thematic summary of research on consent in LMIC paediatric research

Existing published work on informed consent in paediatric research in LMICs includes a number of review and opinion articles (table 2) as well as case studies describing the experience of individual research teams and discussing the challenges and solutions used when adapting consent processes to their local context. We summarise several major themes emerging from these studies here in narrative form and provide detailed key findings from the reviewed articles in the accompanying tables.

Understanding social norms around decision-making and protecting individual autonomy

An important principle highlighted in international guidelines on informed consent in LMICs is appropriate and early engagement with existing local leadership structures (such as a council of elders) balanced against respect for the autonomy of individual children or their caregivers.3 8 In practice, this can be a delicate balance to maintain (table 3). Kongsholm and colleagues, for example, describe consent processes in rural Pakistan, where family structures are patriarchal and hierarchical. In this setting, consent procedures involved first seeking consent from an elder, who provided initial consent for the entire family. However, under this approach, the voluntariness of individual participants may be undermined, and it is unclear how best to ensure that individuals still retain an ‘opt out’ mechanism or, conversely, the right to participate in research if they wish to do so but the elder declines.13

Another important consideration explored by some studies is understanding how not all potential consenting caregivers may feel empowered to decline participating in research. Consent procedures administered by local research personnel or by individuals with high social status, such as physicians, may inspire trust.13 14 However, it may also make them reluctant to decline participation, or to resist active participation. For example, in one study in Kenya, explicit refusals to participate were often considered to be impolite. Here researchers found that caregivers expressed their unwillingness to participate by delaying the consent process, or by participating inconsistently in research procedures even after initially having consented to the study.15

Adapting consent procedures to low-literate settings

There is strong consensus in international ethics guidelines that written, informed consent is preferred when conducting research (table 4). In the case of paediatric research, this typically involves obtaining written consent from one or both primary caregivers.4 9 16 However, in many LMIC settings, literacy may be low or a high value may be placed on oral interactions, and lack of alternative consent procedures may violate another core ethics principle, namely the equitable distribution of research benefits and burdens across populations.3 14 17 Some of the studies we reviewed described these procedures, with verbal consent commonly being obtained, most often in the presence of a literate witness who is able to read available consent documents.13 14 17 18 In one very thoughtful piece, Kalabuanga and colleagues note, however, that witnesses may often impose their views on the consenting caregiver and their child, rather than encouraging dialogue and acting as a safeguard, especially since they are often recruited in an ad hoc fashion (eg, other literate patients or ancillary hospital staff).18 Kalabuanga et al go on to suggest that these challenges may be mitigated by careful vetting and training of independent witnesses or, alternatively, by allowing potential consenting caregivers to use a trusted relative or friend as their witness.18

Another issue identified in the review is that of emerging mandates in some LMICs to document consent procedures. For example, in India, audiovisual documentation of obtaining informed consent is now required for most clinical trials if participants are low literate. This introduced significant new logistical challenges and costs related to obtaining and archiving recordings, and it may also pose a barrier to potential research subjects who may distrust or refuse to be recorded.19

Working in indigenous or less commonly spoken languages

International ethics guidelines emphasise that research information should be provided to consenting caregivers in a local language understandable to the individual (table 4).7 8 16 However, this is most commonly understood to be a working lingua franca, and the issue of provisioning consent processes in an indigenous language is largely unaddressed in LMICs.20 This is an important consideration, given that a substantial proportion of the potential paediatric research population in LMICs are from populations that speak indigenous or less commonly spoken languages.21 In an interesting review of lessons learnt in a paediatric vaccine trial in West Africa, Martellet and colleagues noted challenges in preparing consent procedures in some of the less common language groups included in the trial, where use of the written form was uncommon, where substantial need to rely on metaphor and paraphrase made back-translation difficult and where written documents were perceived as not being dynamic enough in cultures which valued interactivity and person-to-person exchange. They describe alternative procedures, such as the preparation of recordings of consent scripts in local languages and extensive practice sessions with research staff obtaining consent in local languages.17 Similarly, another vaccine trial in The Gambia described the successful use of Speaking Book audiovisual tools in local less common languages to consent caregivers.22

Gender dynamics in caregiver consent

Local gender dynamics and decision-making procedures when consenting male and female caregivers for research are an important consideration (table 5). For example, when consenting with caregiving couples or within an extended family unit, instances are discussed where a female caregiver wishes to allow her child to participate, but is unable to do so because her husband or another male authority figure refuses.13 The opposite may also occur, if a research study is consented by a male figure, but requires significant participatory effort from the primary female for study-related activities, leading the woman to express their refusal through procedural delay or inconsistent participation.15 Given concerns about gender power imbalance and potential repercussions for consenting female caregivers, some studies discussed working to routinely involve fathers or male authority figures in the consent process for more complex or higher risk research interventions.15 23 In one interesting study based in India, Rajaraman and colleagues found that caregivers were more likely to actively participate in the consent process when both were present. They also observed, however, that this factor may have been due to the fact that most study staff obtaining consent were male, and they call for more research on how the gender of research staff impacts the consent process.24

It is important to note that most discussions of gender dynamics that we reviewed were limited in nuance, tending to focus on instances of overt over-riding of female decision-making by male authorities. A broader consideration of the range of ways in which female caregivers communicate, influence and negotiate decision-making with male family members and other community authorities is an obvious point for future investigation.

Disclosing potential benefits and risks of participation in research

Participation in some research studies, particularly those with a randomised controlled design or those with differing intervention arms, may not result in direct benefit to all participants. Several studies report difficulties explaining to caregivers that medical research procedures may not result in direct benefit to their children, and in verifying that caregivers comprehended randomisation or control procedures (table 6).25–28 Others noted the need to address issues of information recall and retention, particularly with complex study procedures or consent forms, and to emphasise the right of study withdrawal and the ongoing reaffirmation of consent throughout a study.26–29 Furthermore, other reports discussed how therapeutic misconception—the perception by research subjects that participation in any component of a multiple-arm, controlled trial will result in therapeutic benefits—might be hard to avoid in certain contexts, as it might be affected by factors like educational level and cultural and religious beliefs about disease.13 18

At the same time, care must be given to a culturally appropriate degree of information disclosure. For example, in several studies, caregivers—especially those of higher socioeconomic or educational status—were more likely to participate when provided with detailed and in-depth information about the study processes and given opportunities to ask questions.12 23 24 30 At the same time, other case studies point out how overdetailed discussion of study procedures or scientific rationale may provoke unneeded reserve or suspicion where such detailed disclosures by health professionals are not culturally customary.13

Finally, in settings where access to healthcare and other important social goods may be limited, even basic diagnostic or ancillary procedures that occur as part of research studies may be better than the local standard of care, leading to an undue inducement or highly compelling incentives for caregivers to enrol their children in research, even after being informed about the experimental nature of studies and the risk-benefit balance.11 13 18 These considerations highlight the importance of considering the socioeconomic and cultural background of study settings well before beginning research and making plans to incorporate appropriate early, equitable benefit-sharing measures when possible, such as using study resources to improve community-level care, not just care for eligible trial participants.18

Adolescents

Adolescents constitute a special population with vulnerabilities different from those of adults and younger children, and they should be included in research that addresses their specific needs (table 7). However, as legal minors they often cannot give informed consent for research.16 In research in LMICs, regulations vary significantly from country to country regarding when adolescents can provide legal consent for research.31 For example, even when legal frameworks allow adolescents to seek  contraception services without parental permission, they cannot necessarily provide consent for research on that theme.32 33 In a scoping review of postabortion care research, Zulu and coauthors discuss how the need to balance adolescents’ privacy needs and the demand for parental consent poses difficulties for researchers in this field.34 Woollett and colleagues describe an interesting case study where they sought consent from a High Court in South Africa for research involving orphaned HIV-positive adolescents. In that study, they provide detailed recommendations for consent involving adolescents, including training staff about confidentiality requirements; recognising immature decision-making by adolescents and developing appropriate methods for probing comprehension and consent; and using methods that promote active participation in research, such as mobile phones.33

Assent

Paediatric research guidelines are unanimous on the need to obtain age-appropriate assent from children and adolescents who do not provide their own informed consent (table 1). However, we found little explicit discussion or description of procedures for obtaining assent in the research reports we reviewed (table 8).35 36 One interesting qualitative study on parental perceptions of assent in Jordan revealed considerable variability in caregivers’ perspectives about at what age assent should be solicited or if assent should even be obtained and dissent respected in all cases.23