Key Informants in Applied Qualitative Health Research

Abstract

Identifying and recruiting key informants is a widely used sampling strategy in applied qualitative health research. Key informants were first conceptualized within ethnography, but there is little methodological guidance about how to use this technique outside of that research tradition. The objective of this article is to offer practical suggestions about how existing methods for data collection with key informants could be translated to methodologies commonly used in applied qualitative health research. This article delineates how key informants could be conceptualized and sampled and how data sufficiency can be approached. The article prompts deeper consideration of the politics of representation and epistemic power that are inherent to the use of key informants in applied qualitative health research.

Introduction

Enrolling key informants is common in many applied qualitative health research studies. Although the “key informant technique” originated in ethnographic anthropology, key (Tremblay, 1957) informants are now commonly integrated into data collection strategies in diverse research methodologies. However, there is a dearth of methodological guidance for qualitative researchers working outside ethnographic traditions. As a result, many applied qualitative health researchers describe using “key informants” without explaining how they defined this role, how they conceptualized the value of information provided by these participants, and how they made decisions about identifying, sampling, and judging the sufficiency of data. This methodological gap also poses an important challenge to readers seeking to appraise the credibility and transferability of findings to other contexts or populations.

In this article, we synthesize existing methodological guidance about sampling and collecting data from key informants to offer practical suggestions about how to incorporate this technique into methodologies commonly used in applied qualitative health research. Our analysis is conceptual, looking to literature outside anthropology about the use of key informants in qualitative research. We synthesized this knowledge through our own scholarly lenses formed by conducting applied qualitative health research with key informants and non-key informants in the fields of medical education and health policy. We lean both on established methodologies (e.g., interpretive description, grounded theory, and case study) and methodological guidance established outside of formal methodologies (e.g., qualitative description, methodological guidance on sampling, and data sufficiency). Our intention is that the principles we highlight here will be adapted by researchers for congruency to the paradigm, methodology, topic, and content of their individual projects.

Using key informants within qualitative traditions outside of anthropology requires adaptation, as the philosophical assumptions and methodological traditions of anthropology do not directly translate to many methodologies used more commonly for applied qualitative health research. However, this work is worthwhile, as the incorporation of key informants can be a helpful addition to the methodological toolkit of the applied qualitative health researcher. A key informant sampling approach can be flexible and assist in building thoughtful linkages to existing knowledge and the experiential knowledge of the research team (Neergaard et al., 2009). Doing this work rigorously requires careful application to the study-specific theoretical foundation, methodology, and methods. The current article provides a foundation for researchers to develop a study-specific rigorous approach to incorporating key informants consistent with their chosen methodology.

Conceptualizing the Key Informant

Traditionally, key informants were engaged in anthropological research to provide “relatively complete ethnographical description of the social and cultural patterns” of their group (Tremblay, 1957). Key informants have been used for both qualitative and quantitative data collection, offering value through their ability to articulate observed social relationships to the researcher (Hughes & Preski, 1997; Seidler, 1974). Key informants are selected due to their ability to help the researcher understand cultural patterns, often providing background information that is inaccessible, implicit, or inefficient to identify through document reviews or other partial accounts (Marshall, 1996; Poggie Jr, 1972).

The term “key” evokes ideas both about the crucial importance of the informant and about their capacity to unlock or grant access to a hitherto inaccessible phenomenon. The term “informant” gestures toward the researcher’s naive position relative to research participants’ embedded knowledge or insider experience of an issue at hand (Morse, 1991); it also reflects an assumption that the subject of the research is complex or inaccessible enough to necessitate interpretation by a knowledgeable member of an in-group. Contrasted with research with participants who are not framed as key informants, insight and knowledge generated from key informant interviews is often framed as “more objective” or given more weight when interpreting results (Lokot, 2021). Research designs that integrate data collection with both “regular” study participants and key informants frequently reflect these tacit hierarchical assumptions.

Key informants provide high level perspectives and comparative insights on issues or questions under study. This contrasts with the scope of insight attributed to other groups of qualitative interviewees, who are more often recruited to provide data rooted in their own lived experiences, or about their personal opinions, and beliefs. In some studies, this distinction is clear: the key informant may contribute knowledge of history, policy, and organizational interaction while the participant contributing lived experience speaks about their experiences embedded within that historical event or affected by the policy. For example, key informants may speak about the implementation, scale-up, and history of public funding for a new prenatal testing technology (Van Schendel et al., 2017), while participants with personal experience of that technology may speak about that experience to offer opinions, beliefs, and attitudes (Vanstone et al., 2015).

In some types of studies, the distinction is less clear. For example, in a study about patients’ post-operative experiences, an experienced nurse or surgeon might be sought as a key informant for their perceived ability to draw on and synthesize observations and insights gleaned from their professional practice. A hospital administrator might be engaged as a key informant to offer insight on typical lengths of stay and rehabilitation resources. A home care aide might offer key informant perspectives on how living spaces might be adapted and home services arranged to facilitate post-operative recovery. The knowledge each of these key informants contributes is based on their professional work (observations, data, and formal training) with many patients undergoing and recovering from the procedure in question. Accordingly, they may offer insights different from a patient or family caregiver, who would be able to speak more directly to the lived experience of anticipating, undergoing, and recovering from that surgery in their own social and medical context.

To further illustrate this difference between participants who share their own lived experience, each of the types of key informants mentioned above might be engaged to share their personal lived experience in response to a different question—perhaps about navigating complexity and uncertainty during a surgery for the surgeons (Cristancho et al., 2013), making decisions about implementing a new clinical program in a hospital unit they oversee for the administrators (Swinton et al., 2021), or grieving after client death for home care aides (Tsui et al., 2019). The difference here is whether the research question requires participants to draw on their personal experiences or offer comment based on accumulated expertise based on a synthesis of different forms of knowledge, experience, and observation of others’ experiences.

Qualities of “good” key informants have been posited to include:

• (1) Community role: “unique access” to, or experience of, the phenomena of research interest by virtue of a profession or social position (Hughes & Preski, 1997; Tremblay, 1957).

• (2) Knowledge: ability to “meaningfully” synthesize their experiences and to offer rich, reflexive insights (Marshall, 1996; Payne & Payne, 2004; Tremblay, 1957).

• (3) Willingness to engage with researchers: “relative ease” to be recruited to participate in research (Bernard, 2017; Tremblay, 1957).

• (4) Capacity to communicate “intelligibility” to the researcher (Pelto & Pelto, 1978; Tremblay, 1957).

• (5) Impartiality: biases should be “minimal” and well-known to researchers (Tremblay, 1957).

When selecting key informants and interpreting their data, it is important to remember that their perspectives are still the situated, partial perspectives inherent to qualitative data. Accordingly, drawbacks of key informant data relate to concerns about the partial and subjective view of phenomena under study that can be elicited from stakeholders, particularly as sampling strategies for key informant studies have yet to be subject to extensive methodological discussion. This form of subjectivity can be perceived as a source of bias that negatively impacts the rigor and trustworthiness of study results (Hughes & Preski, 1997; Lokot, 2021; Payne & Payne, 2004) but can be mitigated by transparent reporting in publication, alongside a frank and fair discussion of the above considerations that guide the identification and recruitment of participants.

What Knowledge Are They Representing?

Key informants are used in applied qualitative health research shaped by a diverse range of epistemological paradigms (Creswell, 2007). In positivist paradigms for inquiry, key informants are often presented as being able to reveal something about “the actual nature of the … domains on which they are reporting” [emphasis ours] provided they are asked the right questions (Poggie Jr, 1972), they are sufficiently unbiased (Tremblay, 1957), or their biases are accounted for by researchers (Bernard et al., 1984; Poggie Jr, 1972). In interpretive paradigms, key informants may be framed as co-constructing knowledge with researchers (Morse, 1991), where the subjectivity of individuals in both groups shapes the process and products of research inquiry.

In research conducted in either positivist or interpretive paradigms, considerations related to how key informants are identified, and knowledge claims that researchers seek to make, ultimately reflect and can reify or disrupt social and political power relations (Lokot, 2021; Soucy, 2000). Since key informants are identified on the basis of expertise relevant to a subject of study, the cultural norms, attitudes, and values that shape perceptions of expertise and experience play a foundational role in shaping researcher considerations about who is authoritative or insightful (Soucy, 2000). This invokes questions about epistemic injustice and what Fricker describes as the implications of prejudice entrenched in social and political economies of credibility (Fricker, 2009). When individual people or communities of people are not recognized or regarded in their capacity as knowers, they are subject to testimonial and hermeneutical injustice that hampers the perception of their expressions as credible and denies them the “interpretive resources” that provide them with architecture to make sense of their experiences in and of the world.

For example, a feminist perspective offers insight on key informants useful to any qualitative researcher, encouraging researchers to critically inquire about why individuals are assumed to be “key” and to whom. Those choosing to engage with their research from a feminist or critical perspective may further consider the under recognition of vested interests that are sometimes, potentially often and invisibly, enmeshed with the power and privilege of key informants. If such interests are not recognnised, qualitaitve research can contribute to the problematic power constructs and relationships that give rise to the identification of key informants and the reproduction of epistemic injustice. These issues can be mitigated by approaches that give communities of “ordinary” people voices and by recognizing the limits of key informant interviews in qualitative research (Lokot, 2021).

How Have Key Informants Been Used in Applied Qualitative Health Research?

Key informants have been used widely in a range of applied qualitative health research methodologies including grounded theory, interpretive description, and qualitative description. Key informants have been referred to as informants, experts, stakeholders, and “key knowledgeables” (Patton, 2014). In applied qualitative health research, key informants nearly always refer to a group sampled by virtue of their professional role or expertise using conventional sampling strategies (Bailey et al., 2012; Hodges et al., 2011; Van Melle et al., 2014). Key informants in applied qualitative health research have included healthcare providers, policymakers, scientists, community leaders, and educators, among others, situated in clinical practice, government, academia, community organizations, or other places that are believed to contribute to their contextual, overarching, and comparative perspectives on an issue.

For example, key informants who occupy social and professional roles have been used in community-engaged research on health promotion (McKenna & Main, 2013); senior health ministry officials were identified as key informants for a study on health system priorities in Canada (Abelson et al., 2017); and individuals in the ministry of health and nongovernmental or civil society organizations were considered key informants about the use of routine health information systems in malaria, tuberculosis, and HIV programming in Senegal (Muhoza et al., 2021).

Typically, key informants are engaged because they provide high level insight or understanding on the phenomena of interest, often through participating in interviews, focus groups, or deliberative methods. Key informants have also been used to probe about how an issue is thought of or acted upon in clinical practice or policy, as well as for understanding how debates on a topic within the field are navigated by its members.

Working With Key Informants

In practical terms, designing a study that solicits and makes use of data gathered from key informants necessitates a clear understanding of how epistemological, methodological, and analytical considerations align. In the following sections, we highlight some of the practical concerns that researchers adapting key informant technique must address. We make a distinction between intrinsically and extrinsically bound projects. By “intrinsically bound” we mean projects about a specific group of people (e.g., policy-making staff within a specific organization and an identifiable healthcare delivery team). By “extrinsically bound” we mean projects about a phenomenon where the boundaries of who may have relevant data to contribute are not as clear and where participants will not necessarily have any interaction with each other, live or work in the same jurisdiction, or participate in the same culture or social groups. Table 1 summarizes these ideas.

Table 1.

How to Operationalize a Reflexive and Transparent Approach in the Building and Reporting of a Key Informant Strategy.

	Key considerations	Things to avoid
Choosing a methodology	Strive for clear congruence of the epistemological, methodological, and interpretive aspects of research.	Selecting a methodology which is philosophically opposed to the use of key informants as you conceptualize them; importing methodological concepts from ethnography into a disparate methodological tradition.
Defining relevant key informants	First, conceptualize your key informants: Be clear about why you are choosing a key informant strategy and how you differentiate these participants from other types of participants. Ask who is considered to be a key informant and why?	Avoid listing roles or attributes of potential key informants without justification.
	Second, define the boundaries of phenomenon under study: Who is in, who is out, and why?	Ensure you have boundaries for the study and eligible KI participants, which are not overly wide or permeable.
	If combining with other types of participants, consider what different weight might be given to the comments made by different KIs by virtue of their perspective?
Identifying and recruiting key informants	Choose KIs strategically, considering their ability to provide insight about both the group and phenomenon of study (Tremblay, 1957).	Relying solely on snowball sampling, which may recruit only those with aligned perspectives on the issue.
	Use an accessible and transparent way to recruit key informants to your study.
	Ensure that you are employing strategies to identify and recruit “less visible” potential KIs who are not already known to you or existing participants.
Collecting data	It will be important to acknowledge the expertise of the key informant and not be overly restrictive in the questions asked of this person. Instead, define the scope of the inquiry in order to give the informant a sense of the boundaries of the study. Within these boundaries, the key informant should be encouraged to speak freely about what they see as relevant and important, with the interviewer drawing the conversation back if it veers outside of the boundaries of the scope of the inquiry. This unstructured approach acknowledges the expertise of the key informant and the likelihood that they may identify points of importance or relevance unrecognized by the researcher (Tremblay, 1957).	Use data collection techniques which encourage the KI to understand your desire that they speak from beyond their personal experience, to help identify overarching patterns or features of the phenomenon under study.
Analyzing KI data	Follow the analytic guidance of your methodology, paying careful attention to the epistemic assumptions you hold about what kind of information KIs are providing.	Avoid treating data from KIs and other types of participants in the same way, if both types are included in the project.
Determining data sufficiency	Identify your strategy for determining data sufficiency. Will this be driven by convenience, access, or connections?	Traditional ways of considering data sufficiency when participants are speaking only about personal experience are unlikely to be rigorous when used with key informants.
	Consider how you will approach the challenge of identifying participants who are unwilling or unable to participate and the implications this has for data sufficiency.
Returning findings/credibility	Incorporating efforts to return the findings and elicit KI response on your interpretation may yield particularly rich insights that will increase the credibility of your study.	Be alert to different opinions in the field and have a plan to work with conflicting opinions or accounts within your participating KIs.

What Sampling Strategies Could Be Used?

As with any sampling strategy, it is important to begin by defining participant eligibility criteria. First, researchers should ask themselves how they will differentiate key informants from other types of participants. Is employment in a particular organization sufficient or do key informants need to have a particular level of experience or role within that organization? What characteristics will indicate the likelihood that a key informant can speak about the broader organization of a phenomenon, rather than just their personal experience interacting with the issue? These characteristics will be operationalized into eligibility criteria. When using a key informant strategy, eligibility criteria commonly describe particular roles, experience levels, organizations, or jurisdictions. They may also describe involvement with the particular policy, program, or intervention under study. For example, in their study of Canadian policy makers’ perspectives on setting objectives for health systems, Abelson and colleagues defined the eligibility of their key informants as those with particular forms of knowledge as exemplified through professional role and duration of tenure in that role (Abelson et al., 2017).

After defining eligibility criteria, researchers must identify a sampling strategy to assist them in choosing who to speak to from among the available pool of key informants. Key informant strategies fit with non-probabilistic, non-representative approaches to purposive sampling, and there are a variety of sampling options that researchers may choose from within this approach.

In intrinsically bound research studies, a complete target population approach may be appropriate (Patton, 2014). This approach seeks to sample everyone within the group of interest, for instance, a particular workplace in a case study or individuals involved with the origination or implementation of a particular policy, where there are only a small number of individuals with relevant expertise.

Quota sampling strategies define different categories of theoretical or analytic interest for purposive sampling, and participants are selected to fill categories (Patton, 2014). In qualitative research, quota sampling is not used for representativeness nor to achieve a specific quantity of participants, but rather to ensure that theoretically relevant perspectives are included in the data. A key informant quota sampling strategy would emphasize breadth, purpose, and variety in key informant participants. This could be done by identifying multiple factors of interest related to, for example, role, organization, experience, geography, domain, or sector of work and then seeking key informants who fit in each of these domains.

Snowball sampling, sometimes known as chain referral sampling, is a purposive sampling strategy that is useful when individuals in the group of interest are not necessarily known to the researcher and may be hard to reach (Parker et al., 2019). Snowball sampling asks existing participants to identify new potential participants. This can be very useful with key informants, who may be knowledgeable about others who have similar roles or interests in their field. As colleagues or people known to each other professionally, a snowball approach may also help with purposive sampling. For example, an initial participant could be asked to identify another key informant in a similar role but with a different form of expertise, conflicting perspective, or working in a specific jurisdiction or organization, as warranted by the purpose of the study. A snowball sampling strategy may also help with recruitment, as the contact is more likely to be known to the target participant than the researcher.

Key informants may also be identified with a theoretical sampling approach, such as that associated with grounded theory (Glaser & Strauss, 2017). This is typically an approach to sampling that incorporates insights from analysis of preliminary data. These emerging insights may indicate particular features of relevance or gap, which the researcher uses to select additional participants.

When using a methodology that encourages iteration between data collection and analysis, multiple sampling strategies may be combined. For instance, the categories in a quota sampling strategy may evolve theoretically. Snowball sampling may assist in filling those quotas. A maximum variation or convenience sampling approach might be used to elicit the initial set of data that will inform a subsequent theoretical sample (Emmel, 2013). In Pahwa’s qualitative description study about ethical and social values toward lung cancer screening with low dose computed tomography in Canada, maximum variation sampling was initially used to sample key informants who were scientists, clinicians, and policymakers currently engaged in Canadian lung cancer screening activities (Pahwa, 2023). These individuals were identified through review of published scientific papers and reports, lung cancer screening study websites, and personal knowledge of key experts and influential figures in the field. As initial key informant interviews were conducted and analytic categories developed, Pahwa used theoretical and snowball sampling to identify subsequent key informants who could be invited to the study to speak about themes or ideas in need of further perspectives, explanation, or articulation. Snowball sampling was particularly useful for identifying policymakers whose identities were not as evidently available in the public sphere on scientific publications or policy reports.

Identifying and Recruiting Key Informants

In an intrinsically bound study, the sampling strategy is likely to point to specific individuals who can then be recruited into the study as key informants. In extrinsically bound studies, identifying eligibility criteria for key informant participants may still leave a need to identify particular individuals from among those who meet the criteria. Additional layers of methodological planning are required to choose who to recruit from among eligible individuals.

For example, in Cavanagh’s extrinsically bound study of physician roles in the provision of care to those who have experienced intimate partner violence, she wished to sample key informants who were stakeholders in the provision of this care (Cavanagh, 2022). She used a quota sampling strategy to identify that she was interested in, for example, speaking with someone who made macro-level policy about intimate partner violence at a provincial or national level. To fill this category, she needed to identify particular individuals and choose who to approach. She did this by employing a variety of tactics to identify potential individuals, looking at publicly available information through published white paper or policy reports, talking to experts in the field, and searching LinkedIn profiles to identify who may currently or previously have held a relevant position. Snowball sampling was also very useful here, as initial participants were often aware of other experts in their field and could suggest and sometimes help recruit other informants.

Once potential individuals are identified, the researcher must choose who to approach. There are many considerations here. Most methodological guidance encourages researchers to think strategically, considering the ability of individuals to provide insight about the group and phenomenon of study (Tremblay, 1957). At this point, it will be important to be very clear about relevance. What are the boundaries of your phenomenon? With key informant sampling, it can be tempting to move outside of typical temporal or geographical bounds. A policymaker in Ireland might have insight on the roles of Canadian physicians in caring for people who have experienced intimate partner violence, or a scientist in the United States might have perspectives about the ethical principles involved with lung cancer screening programs in Canada, it is true. But without a clear rationale for reaching that far, it will be difficult to make an argument for the comprehensiveness and sufficiency of your sample, so make judicious decisions about defining boundaries and reaching beyond them.

As the project continues, think about the evolving sample. What areas of affinity and conflict are present in the existing sample? What gaps may exist? Some of this work may be possible to do during the planning phases of the project, consulting with collaborators who have expert knowledge of the field. Other work in this area will come during data analysis, as initial data indicates potential theoretically relevant areas that need to be investigated to provide a comprehensive view on the phenomenon (Charmaz, 2014).

Reflexivity is important in this stage. Consider how your own identities, networks, and assumptions shaped the way you defined, identified, and recruited key informants. An audit trail can help document these important research decisions (White et al., 2012). This documentation can be a catalyst for reflexive conversation among the research team about who was not invited or who did not respond to invitations, what knowledge or perspectives may not be represented, and what information provided by key informants may be expressed or influenced by vested interests. Documenting and reporting these processes can assist in facilitating the confirmability and trustworthiness of the project, providing much needed transparency. It can also invite critical discussion of the limitations of how key informants were conceptualized, identified, and sampled, as well as of the information collected from key informants.

Accounting for Data Sufficiency

Accounting for data sufficiency is one of the most under-developed aspects of existing methodological guidance about key informant sampling. Feasibility constraints are particularly acute in key informant research. Unlike other types of qualitative research, where pools of potential participants are very large, key informant groups can be quite small. They may also be hard to reach, busy, and have professional, organizational, cultural, or legal constraints on information they can share with researchers. While feasibility limitations to recruitment are real, they do not diminish the need for rigorous ways to establish data sufficiency.

In an intrinsically bound project where the researcher has invited participation from all eligible individuals via a target sampling approach (Patton, 2014), it may be sufficient to describe that all were invited, and all volunteers were included (Khalid et al., 2019). The researcher could increase credibility by offering comment on any common characteristics noted about those who declined to participate, and on the influence that may have on the data represented in the research. Depending on the methodological approach used, there may be other ways the researcher could enhance the credibility of a limited sample, perhaps by including other forms of data (e.g., documents and observation) where it is methodologically appropriate to do so.

In intrinsically bound studies, regardless of whether or not a target sampling approach is used, we may borrow from Robert Yin’s conceptualization of data sufficiency within case study research. According to Yin, data sufficiency is related to the “completeness” of the data needed to understand the case (Yin, 2017). Yin suggests this can be done by establishing of “converging lines of inquiry” or identifying consistency across data from multiple sources. Others understand this as a form of triangulation (Flick, 2018).

In an extrinsically bound project, determining and justifying data sufficiency becomes more complex. Common qualitative data sufficiency constructs like saturation will often be insufficient, due to a mismatch between underlying assumptions of the study and the key informant technique. Saunders and colleagues describe four different ways that saturation is defined (Saunders et al., 2018). When defined as information redundancy, saturation is inconsistent with key informant sampling because it doesn’t recognize that key informants are speaking beyond their personal experience, providing an overview of the organization of a phenomenon. Accordingly, each key informant may contribute a very different perspective that is not intended to overlap. Redundancy may therefore not be possible and also may not be needed to offer assurance that relevant perspectives are incorporated. When incorporated during data analysis, achieving thematic saturation does not assure the researcher that relevant divergent perspectives have been adequately incorporated.

If methodologically congruent, researchers may choose to adopt Malterud and colleagues’ proposal of information power as a model for determining data sufficiency (Malterud et al., 2016). Information power is useful for planning what resources will be needed for a study related to the anticipated number of participants, which will have implications for transcription, participant honoraria, and researcher time. However, information power must be continuously evaluated during data analysis. This model posits that data sufficiency rests on whether there is specific information to answer the question and offers five items that may influence the level of information required: study aim, sample specificity, use of established theory, quality of dialogue, and analysis strategy (Malterud et al., 2016). These items are inter-related and influence each other. For example, a study of expert perspectives on which barriers and resources affect the ability of adolescent and young adult cancer survivors to return to education or work used the theory of information power to inform their judgment of data sufficiency, judging 15 key informants to be sufficient on the basis of strong dialogue, the use of an underlying theory, and the use of a cross-case analysis strategy (Pedersen et al., 2018). Key informant data may reach information power because this model of data sufficiency accounts for participants speaking beyond their personal (specific) experiences who are likely to offer strong and sophisticated dialogue with the researcher, although in analysis the researcher would need to explore the meaning and basis of this perspective.

Ethical Considerations

There are particular ethical considerations when using key informants, many of which are related to the unethical nature of conducting methodologically weak research. In relation to sampling, researchers may reflect upon who are they conceptualizing as a key informant and inviting to the research and why? How are they being invited to the study? Importantly, who is not being invited and what are the possible implications of the eligibility criteria for participation and the reproduction of certain forms of power, privilege, and knowledge?

Key informants who already have epistemic power are often readily identifiable to the researcher from published studies or existing knowledge of who is influential in the field. Key informants who have relatively less influence, perhaps due to a dissenting or critical stance on a topic, may be more challenging for an outsider researcher to identify. These critical voices risk being excluded from the research. Researchers may consider the potential implications of including and excluding key informants on reproducing epistemic injustice in the analysis and interpretation of study results, particularly with respect to whose ideational power is amplified, whose vested interests are being mobilized through the research and given credibility in the research, and why.

For topics where there is contention or debate, key informants are often engaged on issues that involve a struggle for epistemic power among themselves. The researcher is thus tasked with negotiating and producing meaning from ideational or knowledge conflict. They may notice that in the desire of key informants to give voice to their ideas, certain types of evidence or knowledge (of which they may have a stake in producing), and the mandates of institutions and interest groups to whom they are affiliated are promoted. It is also important for the researcher to acknowledge that epistemic communities and epistemic power are not monolithic and fixed in time and space; these are continuously changing and sometimes reinventing themselves as political priorities and knowledge paradigms shift, evolve, or emerge.

Researcher positionality may impact how key informants are sampled in a study. Researchers who are part of the in-group at the outset of the research may depend to some extent on sampling key informants with whom they share a personal connection, invariably invoking their personal reputations during participant recruitment. In these situations, it is vital that the researcher be transparent about their positionality and subjectivity and be reflexive and conscientious about their power, power dynamics, and the situated relatedness of the researcher and key informant in the field. Researchers who do not have an “in” at the outset of the research also need to be reflexive about power and relationships as these are equally relevant, albeit usually in the inverse direction (i.e., key informants holding power over researchers). The process of researchers possibly gaining an “in” as the research progresses, through gaining insider knowledge or even becoming part of the in-group, also needs to be reflected upon and articulated by the researcher through the overlapping lenses of positionality, power, and access.

The implications of these ethical considerations are many and dependent on the paradigm each researcher is working within. Feminist and critical researchers will be well familiar with these topics and likely already prepared to include an analysis of which voices are amplified, who is absent, and who may be silenced in a particular analysis. Researchers working in other paradigms may find these considerations less familiar and may choose to incorporate these considerations into their sampling and recruitment strategy, in order to ensure that less-resourced, less-established perspectives are all included.

Worked Examples: Key Informant Conceptualization and Sampling in Two Methodologies

Example 1: Pahwa engaged key informants via one-on-one semi-structured interviews in a qualitative description study about ethical and social values toward lung cancer screening with low dose computed tomography in Canadian jurisdictions (Pahwa, 2023). The aim of this research was to understand what key informants perceived as the ethical issues associated with lung cancer screening in Canada and to observe their processes of navigating these ethical issues from their distinct place in lung cancer screening research, clinical practice, or policy. The methodology, qualitative description (Sandelowski, 2000), was selected to stay close to the spoken words and ideas or meaning ascribed to them by key informants as elicited in interviews. The interpretation of results was subject to a low level of inference to provide a comprehensive description of subjective facts of the phenomenon, lung cancer screening ethics, in everyday terms. Epistemologically, this study was carried out from a pragmatic position, permitting a slight degree of interpretive inference within the descriptive aim required by the methodology. Key informants were conceptualized as scientists, clinicians, or policymakers who influence lung cancer screening in Canada. Key informants were expected to possess sufficient knowledge about the design, benefits, harms, and objectives of lung cancer screening, and the willingness and capacity to communicate with Pahwa about their perspectives on ethical aspects. Impartiality was relatively more challenging for Pahwa to estimate at the outset of the research, although biases were sometimes deduced as key informant interviews were conducted. This study was justified as a KI study because insights on the ethical aspects of lung cancer screening in Canada were not possible to attain from the literature or interviews with other types of research participants.

Example 2: Cavanagh conducted unstructured interviews with key informants for an extrinsically bounded interpretive description study exploring stakeholder priorities and perspectives about medical training related to intimate partner violence (IPV) (Cavanagh, 2022). The primary goal of this study was to draw on expertise of stakeholders outside of medicine to generate training recommendations to enhance physicians’ clinical practice related to IPV in the future. Interpretive description, a methodology developed by nursing scholars, was chosen in view of its alignment with the study goal of developing insights with practical applications to medical education (Thorne et al., 1997). Data analysis proceeded concurrently with data collection, incorporating constant comparison, member-checking, and other methodologically specific strategies for assessing credibility and validity of new findings. Epistemologically, interpretive description is rooted in constructivist and interpretivist paradigms, producing findings that claim to represent and interpret phenomena of clinical interest. Key informants were conceptualized as stakeholders doing professional and/or advocacy work related to IPV in fields outside of medicine that meant they would have uniquely informed perspectives on improving medical care for people affected by IPV. Participants were initially identified using quota, theoretical, and snowball sampling strategies; in the course of interviews, participants were asked directly to comment on aspects of their positionality that shaped their perspectives, and were also asked to suggest prospective participants whose informed perspectives were likely different from their own. The use of key informants in this study was justified as the goal of this study was to generate recommendations drawing on the unique access and insights of experts seldom included in medical education research.

Conclusion

Key informants are an oft used and powerful tool in applied qualitative health research. In this article, we have offered parameters that qualitative researchers may use to thoughtfully and deliberately incorporate key informants into individual research projects. This includes consideration of how key informants could be conceptualized, a critical perspective on what knowledge key informants represent, and methodological questions to guide decisions about how key informants could be sampled, identified, and recruited and how data sufficiency could be formulated. The worked examples from health policy and medical education research, two areas within applied qualitative health research, bring to vision the congruence between why key informants were engaged, how they were conceptualized, and what methods were used to collect data. This article addresses a methodological gap in the qualitative health research literature about key informants, providing flexible suggestions that can be adapted to many qualitative methodologies to increase comfort with designing a rigorous key informant strategy. Beyond these implications for designing research, we also highlight the need for additional critical research that elaborates on the specific epistemic and ethical complexities inherent to the use of key informants in qualitative health research.

ORCID iDs

Manisha Pahwa https://orcid.org/0000-0002-9131-8721

Meredith Vanstone https://orcid.org/0000-0002-7347-6259