Practical Strategies for Health Equity Researchers to Enhance Analytic Rigor and Generate Meaningful Insights From Qualitative Data

Jennifer K. Felner, PhD, MPH^1,2; Vida Henderson, PhD, PharmD, MPH, MFA^3,4 (View author affiliations)

Suggested citation for this article: Felner JK, Henderson V. Practical Strategies for Health Equity Researchers to Enhance Analytic Rigor and Generate Meaningful Insights From Qualitative Data. Prev Chronic Dis 2022;19:220134. DOI: http://dx.doi.org/10.5888/pcd19.220134.

PEER REVIEWED

Abstract

Researchers and public health practitioners increasingly need to leverage diverse methodologic approaches in health equity research that will lead to innovations in the assessment of health inequities and development of interventions to decrease health inequities. One well-suited approach is the use of robust qualitative methods (alone or in combination with quantitative methods). As more health equity researchers employ qualitative methods in their study designs, additional guidance is needed on how to conduct robust and rigorous qualitative data analyses. We share a 4-step analytic strategy for health equity researchers and practitioners — particularly those with limited training in qualitative data analysis — that can be used to effectively execute qualitative analysis to inform health equity–driven efforts. These strategies will guide those less experienced in qualitative methodology to employ a pragmatic approach to analysis that is sound, reasonable, and produces meaningful insight that can be used to inform efforts to advance health equity for communities with the greatest needs.

Top

Introduction

Qualitative methods use nonnumerical or nonstatistical processes to explore human behavior and experiences in context as well as complex social-level and structural-level phenomena, including the social production of health (1–3). Because health equity–driven research prioritizes eliminating socially unjust differences in health such that all have equitable access to resources, quality health care, and opportunities to be healthy, qualitative methods are an important tool in the health equity researcher’s or practitioner’s toolbox. Employing qualitative methods (alone or in combination with quantitative methods) offers opportunities to produce new insights into the sources of health inequities (4–6) and leads to innovations in multilevel intervention development to decrease health inequities (1,5,7).

Major public health funding bodies encourage researchers to propose study designs that integrate qualitative and quantitative data (8). Integrating qualitative methods in study designs allows researchers to develop a more nuanced and holistic understanding of relationships and contexts that influence health than quantitative methods alone can (1,8). As more researchers and practitioners employ qualitative methods, there is a need for accessible and straightforward guidance on how to analyze and identify meaning within qualitative data, particularly among those without formal training in qualitative methods. Understanding qualitative data is especially needed within the context of health equity research, in which qualitative methods may be a primary source of information about how and why inequities exist and what people think should be done to advance health equity for their communities.

Considerable time and effort are required to develop expertise in qualitative analysis; however, time and resources may be limited for those working in health equity–focused research and public health practice. Therefore, we share an overview of a systematic, yet pragmatic, qualitative analysis approach to explore phenomena in context, elevate voices of those affected by health inequities, and inform health equity–focused interventions and related efforts. We will not delve into additional details on the use of qualitative methods for health equity research, assessment, and evaluation (for a recent review, see Shelton et al [5]). Our pragmatic process follows some of the analytic strategies of applied thematic analysis (9) and other approaches popular in the health sciences, such as constructivist grounded theory (10) and phenomenology (11). Definitions of key terms are provided in Table 1.

Top

Analytic Strategies and Health Equity Implications

We provide a set of analytic steps that we have each applied to multiple qualitative data sources, including data from semistructured and unstructured interviews and focus groups (eg, data in the form of audio files and verbatim transcripts), participant-observation and ethnography (eg, data in the form of field notes), narratives (eg, data in the form of written or published text), and photovoice (eg, data in the form of photos and oral or written analysis of photos). For simplicity, we will focus on analysis of verbatim transcripts herein. These steps can be applied using computer-assisted qualitative data analysis software (CAQDAS; eg, NVivo [QSR International], MAXQDA [VERBI GmbH], Atlas.ti [Atlas.ti Scientific Software Development GmbH], Dedoose [SocioCultural Research Consultants, LLC, 12]); or basic word-processing software and spreadsheets (13,14). A summary of each analysis step and the estimated timeline for completion are provided in Table 2.

The analytic steps outlined herein are a team-based process. We firmly believe in involving diverse research teams in health equity research broadly and in analysis specifically. This means diversity in terms of methodologic or practical training, social identity or position (eg, race, gender, class), or research profession (ie, when possible and germane to the study goals, both professional researchers or public health practitioners and community partners are involved in analysis).

Top

Step 1: Memoing, Annotating, Jotting in the Margins

In the first step of the qualitative analysis process, team members write and apply analytic memos to the data, known as memoing (also referred to as annotating or jotting in the margins). Memos are brief “notes to self,” capturing initial impressions of the data and salient ideas that may be analytic or reflexive (15). They are usually a few words or sentences and can be directly attached to the data by physically writing notes in margins on hard copy or by using electronic track changes features in software to identify important or salient ideas or thoughts. Writing is an important element of qualitative analysis; writing memos allows researchers to begin immersing themselves in the data from the outset by formulating initial ideas and impressions in narrative form, and it is an initial step in understanding the depth and range of participants’ thoughts, ideas, and expressions (15,16). Additionally, writing memos ensures that subsequent code development (step 2) is grounded in the data.

In this step, each team member is randomly assigned 1 to 2 transcripts to memo. When timelines are very tight, research teams may elect to memo only a subset (eg, 15%–20% of transcripts) of randomly selected or purposefully selected transcripts for maximum variation across data or participant types. Ideally, all transcripts will be memoed by a member of the research team.

Implications for health equity

Understanding lived experiences and root causes of health inequities requires deep exploration and inquiry into complex, multilevel factors that may affect multiple domains of a person’s or community’s health. Memoing helps the health equity researcher move beyond simply identifying and applying a priori or index codes (ie, predetermined codes) and enables researchers to be open to the direct lived experiences, thoughts, and ideas that are directly voiced or conveyed by participants. In addition, reflexive memoing can be used by health equity researchers as a process to reflect on their position regarding the research topic and communities of focus, relationship to participants, biases, and power balances that might affect the analysis process and findings generated.

Top

Step 2: Compile Annotations and Develop Codebook

After memoing, or a first pass of writing memos (some researchers memo throughout the analysis process, including during coding [step 3]), is complete, analytic memos can be compiled into a list to inform the identification of codes and development of the codebook. Word processing or CAQDAS can be helpful to easily output the memos into a single document. Reflexive memos may or may not be appropriate to include in this list, depending on the goals of the analysis. Once the memos are in a single document or list, a single researcher or multiple research team members read through the memos and reduce them to a few central words or short phrases that capture the essence of the memo. We recommend retaining a copy of the memos in their original form, which may be useful at later stages of analysis.

Research team members then read the reduced memos to identify key ideas and group them into “buckets” that are linked through a central idea. This process will inform the development of codes to organize and categorize segments of the data. Codes can be organized in a codebook, with each code represented by a descriptive word or phrase characterizing its meaning. Note that codes are not themes. Codes are simple, descriptive ideas. They are not higher-level concepts based on identified and interpreted patterns in the data. Codes are in service of identifying themes (16). Many researchers conflate these 2 concepts.

Depending on the goals and complexity of the analysis, codes may have a hierarchical structure in which codes are organized within code families of similar topics or ideas or into more fine-grained subcodes. The overall purpose of codes is to organize and categorize segments of data such that main ideas can be identified, interpreted, and shared (step 4b). A practical codebook will include names and definitions for each code and example quotes taken from the data that illustrate when codes should be applied. Often, details will include when to use and when not to use a code. The more detailed the codebook, the easier it will be for those applying codes to do so consistently and reliably.

Research team members should collaboratively develop codes and draft and refine the codebook (eg, clarify definitions, ensure codes are mutually exclusive, ensure code names are sufficiently descriptive). Codes directly informed by memos or reading of the data are referred to as inductive codes (ie, grounded in the data). However, researchers often have predetermined concepts they want to capture based on conceptual or theoretical frameworks, interview or focus group questions, prior research, or research questions or study aims. Codes based on these predetermined concepts are referred to as deductive codes, a priori codes, or index codes and tend to capture more general ideas than inductive codes. Most codebooks will include both inductive and deductive codes.

As with all analytic phases, openness to multiple iterations for refinement is needed. Another consideration among research teams is the level of coding that is needed for a given project. Although there is no predetermined number of codes appropriate for any given project, teams must decide if the analysis requires macro-level coding (codes that capture broader characterizations) or more detailed and specific codes or subcodes.

Implications for health equity

This step is an opportunity to leverage existing health equity–related frameworks, theories, or models to identify additional codes or code families and to guide the subsequent analytic processes. If researchers aim to understand how a certain health equity–related theory applies to or is aligned with the data, they might use constructs of that theory as codes (or to frame or categorize themes [step 4]). By using in vivo codes (a type of inductive code that use verbatim words or phrases from the data), however, the analysis is grounded in participants’ perspectives and retains their original words and language. Additionally, code development may be an initial step to inform new theory development or refinement when existing theories do not adequately capture relationships found in the data. For example, individual-level health behavior theories are often insufficient when examining a research question with an equity lens. Inductive coding can help researchers uncover multilevel factors that contribute to a person’s ability to enact behavioral change, resulting in theoretical frameworks that consider social and other external factors that affect equitable outcomes.

Top

Step 3: Coding Data

Coding is the process of organizing data by attaching codes to relevant segments of text We liken this to placing a sticky note on parts of the transcript to flag it for later retrieval. Transcripts (and other documents, such as photos) can be coded with CAQDAS or word processing software and spreadsheets (13,14). Once the data are coded, researcher(s) can then retrieve and review the coded text segments to identify the higher-level concepts across the data (themes). Research teams typically have at least 2 people, referred to as “coders,” who code transcripts, especially when there is a large amount of data.

Selecting text segments or codable units

An important consideration before coding is to determine what will be considered a codable unit. A codable unit is a discrete segment of text to which codes are applied. A common coding misstep is inappropriate determination of a codable unit (or a lack of training for the coders on what to code). When coding text, a codable unit must make sense when standing alone. It is often unhelpful for coders to select a few words or even a single sentence that does not encapsulate meaningful context as a codable unit because it will be difficult to interpret when reducing data (step 4a) and identifying themes (step 4b). For example, the research team may decide that a complete thought is considered a codable unit (which could be a few sentences or paragraphs) or that an entire response to each interview question is a codable unit.

Coding reliability

Before coding data independently, it is common for 2 (or more) coders to both code approximately 10% to 25% of the data to assess how similarly they are applying codes; this is referred to as intercoder agreement (ICA). If there is insufficient ICA, which means that coders are applying different codes to the same segments of text, there may be codebook issues to be addressed, such as unclear code definitions, missing codes, overlapping or redundant codes, or a need for more training. Once sufficient ICA is reached, coders may begin coding the remaining data independently. Best practice is for an experienced research team member to preselect codable units for the coders during ICA assessment (eg, by highlighting codable units before applying any codes). This will help coders learn what is considered a codable unit for the particular analysis and make it easier to assess coding reliability, because coders will each be working from the same point of reference (as opposed to potentially selecting and coding different segments of text).

Generally, there are 3 approaches to assessing ICA: subjective agreement, percent agreement, and statistical agreement, with debate about which, if any, is the most appropriate to use in qualitative data analysis (9,17,18) (for a useful overview of the debate, see O’Connor and Joffe [19]). Overall, the selected approach to assessing ICA will be driven by project goals, research team skill and access to analytic resources, philosophical underpinnings of the study, and feasibility — each of which may vary by study even if conducted by the same research team. We believe the goal of assessing ICA should be to generate research team dialogue and reflection that will inform codebook improvements and increase the coders’ confidence and effectiveness in coding important segments of the data. This assessment should be considered a helpful process, rather than an end goal to “prove” the reliability of an analysis and subsequent findings.

To assess subjective agreement, coders simply compare and contrast their code applications across segments of text and identify instances of differing code applications. Discussion is used to determine which, if any, code application is right, then coders make adjustments to the codebook or their coding as needed (sometimes referred to as consensus coding). Mathematical calculations are not conducted in this assessment of ICA.

To assess percent agreement, a research team member tallies the number of instances in which coders applied the same code(s) to preselected segments of text. That number is divided by the total number of instances in which coders applied the same code(s) to preselected segments of text plus the number of instances in which coders applied different code(s) to preselected segments of text.

Statistical agreement extends percent agreement by calculating a statistic of code agreement accounting for chance. Some suggest statistical agreement is superior to percent agreement because it accounts for chance and as such, should be prioritized to assess coding reliability (19,20). However, we do not ascribe to this notion for every study or research team. The most commonly used statistic is Cohen kappa (κ) and, more recently, Krippendorff alpha (α) (19). These statistics can be calculated by using multiple CAQDAS software programs as well as online calculators. For a free, detailed resource describing how to calculate and use these statistics, see Geisler and Swarts, chapter 5 (17). For a detailed application in applied public health research, see MacPhail et al (21).

Pilot or first round assessment of ICA

As described above, codable units should be the same and be preselected for coders. For first round ICA, 2 or more coders code an entire transcript or only half of a transcript — this is largely dependent on the amount of data. We suggest that when there are fewer than 20 transcripts, the coders may code half of a single transcript during this first round. The research team should predetermine the acceptable minimum standard for reliable ICA; 80% for percent agreement (scale of 0%–100%) and 0.61 for statistical agreement (scale of 0–1) have been identified as common minimum standards, although there is a lack of consensus on these standards (19). On the basis of this minimum standard, the research team can determine if coding is insufficiently reliable and thus codebook updates are needed (they almost always are) or if additional coder training is needed.

Second round assessment of ICA

Once the codebook is refined based on pilot or first round assessment of ICA, coders should code another full transcript (or portion of a transcript), recalculate ICA, and again discuss and implement any needed changes to the codebook in partnership with the broader research team. This process is typically repeated until sufficient reliability is achieved. Notably, sufficient ICA may be more difficult to reach with codebooks that contain a large number of codes and subcodes. However, that is not a reason not to include all necessary codes in a codebook. Coders should predetermine the code level for which they will determine agreement (eg, code family, subcode). Once sufficient ICA is reached, the remaining transcripts can be divided among the coders.

Implications for health equity

Journal reviewers or researchers who are less familiar with qualitative methods tend to rely excessively on the utility of ICA as an attempt to lend quantitative credence to qualitative methodology. Although assessing coding reliability is a useful analytic process that offers the opportunity for refinement to ensure that coding processes and the meaning of codes are valid, it is more important as an opportunity to engage in additional dialogue and reflection that can ensure a health equity stance in the analysis process.

Coding requires deep and focused attention to the data, which enables thorough insights, facilitates validity and transparency in interpretation of findings, is a vehicle for understanding participants’ perspectives and identifying and discovering relationships, structures the data, and makes it accessible (22). All of these attributes are critical when seeking to understand the complex interplay of factors that affect health equity. Codes are important guideposts for team members as they discuss, distill, and seek to understand data during the analysis process.

Top

Step 4: Data Reduction and Theme Identification

Step 4a: data reduction

Data reduction is a purely descriptive phase of the analysis process. Data reduction is taking a large amount of data (all data excerpts categorized by code) and distilling it to key distinct points that were conveyed by participants. To achieve this, the next step in the analytic process is to organize or group all coded text segments (ie, excerpts) by each respective code. CAQDAS or other specialized software allows researchers to easily export all coded segments for each code in desired formats (eg, Word, Excel, PDFs). One strategy for reducing data is for 1 or 2 team members to write data summaries for each code or code family by reading excerpts for each code from exported documents and narratively summarizing what was expressed by participants for each code or code family. This will result in data reduction, not in themes. Once summaries are completed, all team members read code summaries and collectively contribute thoughts and ideas for salient themes derived from the data. For quantitative data-oriented researchers, codes may be thought of as variables, excerpts as raw data, and summaries as descriptive results.

We recommend that researchers do not attempt to identify themes during the data reduction phase, although of course, some ideas will begin to form. This phase is only about reducing data before developing themes. Team members should have a thorough understanding of what was expressed by participants, independent of any given team member’s thoughts about relationships and associations. This allows each team member to reflexively contribute their own thoughts and ideas related to concepts expressed and sets the stage for increased depth and range of ideas during the theme-generation phase. As with most phases of qualitative data analysis, summarizing results is iterative. For example, after examining initial written summaries, teams may decide that it is necessary to conduct additional data coding to get more granular details of a particular code or code family, or different research questions may require additional examination of a particular phenomenon.

Step 4b: theme generation and meaning-making

At this point, analysis moves from categorization to theme generation and meaning-making. Two key types of qualitative analysis goals should be considered in preparation for this step. The first is descriptive qualitative analysis, which aims to identify and detail the who, what, and where of events. In these analyses, researchers stay close to the data and do not aim to uncover processes or phenomena that are under the surface of the data or develop theoretical or conceptual models based on the data (23). The second is interpretive qualitative analysis, which aims to move beyond description of the data to uncover more complex processes or phenomena, often with the broader goals of developing or informing theoretical or conceptual models and answering research questions. Both descriptive and interpretive analytic goals are often applied to the same set of data; however, it is recommended that researchers identify the goals of their study well before analysis begins to determine whether goals of analysis are descriptive, interpretive, or both. Qualitative health equity research and analysis are often interpretive in nature, given the common goal of identifying root causes of health inequities.

Regardless of the analytic goal (descriptive or interpretive), moving from codes to themes is perhaps the most abstract and time-consuming phase of the analysis process. Sometimes researchers get bogged down with ensuring near perfect ICA when that energy and time is better spent on data interpretation and theme generation. Themes are high-level concepts based on patterns and linkages in the data — representing shared units of meaning connected by a central organizing concept or phenomenon (24,25). We conceptualize themes as the “a-ha,” “so-what,” or “big take-away” of the data. Clarke and Braun (24) explain that themes differ from basic topic summaries of the data in that “themes [are akin to] key characters in the story we are telling about the data (rather than collection pots into in which we place everything that was said about a particular data domain)” (p. 108). Even a descriptive qualitative analysis should strive to move beyond simply reducing the data and grouping data into buckets (step 4a) to identifying higher-level themes across the data.

So how do researchers identify and detail the themes of their data? Strategies have been described in prior publications (16,26–28). Some key strategies involve identifying 1) repetitions across the data, though repetitions alone are insufficient to signify a theme; 2) metaphors and analogies in the data (this could be found in both the textual, coded data as well as in the analytic memos developed during step 1); 3) transitions in the data (ie, natural or intentional shifts in participants’ comments or words that connect ideas or concepts such as “because,” “since,” “if,” or “then”); 4) similarities and differences across the data or multiple sources of data (ie, how a described experience or perspective is similar and different across transcripts or across data from various sources such as interviews, focus groups, or observations); 5) missing data or “silences in the data” (ie, considering what was left unsaid or not mentioned in and across the data may shed light on topics that participants wish to avoid or that researchers may have thought would be relevant but were in fact not relevant for the participants); and 6) elements of or connections to established theory, which may help place the findings in a broader conceptual or theoretical context (9,26). In addition, it may be helpful to develop thematic networks or maps to visually connect ideas between higher-level organizing themes and more concrete ideas related to the theme (for examples, see Attride-Stirling [29] and Richards et al [30]). Some CAQDAS produce visualizations of relationships between codes or patterns in the data; however, simply drawing these networks or maps by hand is effective. Contrary to some methodologic discourse, we suggest themes do not “emerge” during qualitative data analysis (although we have each been guilty of using this language in the past), but rather are “produced by the researcher through their systematic analytic engagement with the dataset, and all they bring to the data” (18, p. 9) on the basis of their own experiences, personal identities and social positions, and training.

In our experience, the most helpful theme-generation process involves some or all of the steps described here plus multiple rounds of research team dialogue based on the coded data and code summaries in the context of the study aim(s). In this approach, research team members apply their own theoretical lenses and knowledge to the reduced data to discuss and identify themes. Moving from summarizing the data to identifying themes takes time, intellectual work, and makes some team members uncomfortable because it requires conceptual leaps that transform lived experiences to higher order concepts. However, just as we make conceptual leaps in quantitative analysis, the same is true for qualitative analysis.

Salient ideas are not necessarily the most commonly occurring; therefore, avoid equating frequency with importance. Ideas expressed by only 1 participant may be as important as ideas expressed by multiple participants. Likewise, a few participants may have discussed a particular idea in depth, resulting in a high frequency of a specific code, but that frequency of code may not indicate a meaningful high-level pattern or phenomenon. Some researchers working with qualitative data may choose to use counting or numbers when relevant for their analytic goals and audience, or when frequency has theoretical or practical meaning (31), but we suggest this be used carefully and sparingly.

Implications for health equity

Perhaps the most important function of qualitative research for the health equity researcher is the opportunity to elucidate and contextualize lived experiences and social processes to inform intervention and program development, policy, evaluation, and theory. Those affected by health inequities are often prey to underrepresentation; a lack of understanding about their experiences; and the social structures, norms, and ideologies that perpetuate health inequities. Data derived from qualitative methods must accurately and appropriately describe conveyed experiences, and interpretations and implications of data must be thoroughly examined and considered among diverse research teams (eg, by discipline, social identity, training).

An important opportunity to apply the analytic processes we have outlined is within the context of community-based participatory research (CBPR) projects. CBPR has the potential to link research and action to advance health equity by authentically and equitably involving community partners in all aspects of the research process (32,33) (for examples of participatory qualitative data analyses, see Dill [34], Hebert-Beirne et al [35], and Switzer and Flicker [36]). Care should be taken to determine the extent to which community partners wish to engage in each step of the analytic process. Such involvement of community partners has the potential to ensure that findings are sufficiently grounded in the needs, ideas, and experiences of those affected by health inequities and that recommendations adequately reflect community priorities. At a minimum, if the analysis process itself is not participatory within a CBPR project, it should be done with “accountability to the community” (37, p. 851), such as sharing preliminary findings (often referred to as “member-checking” [38]) with community partners or other stakeholders to validate and offer additional considerations regarding researchers’ interpretations and recommendations to advance health equity through intervention development or policy making.

Top

Beyond Analysis: Reporting Findings

Qualitative data analysis is iterative in nature, and the multiple steps involved, even if nonlinear, should be thoroughly described in publications and presentations of findings (including processes such as memoing, codebook development, testing and refinement, and approaches to theme generation) (39). Typically, researchers report findings by theme, including description and interpretation of the theme, and use verbatim excerpts (quotes) from the data to provide evidence for the theme and honor participants’ voices. Quotes should be edited only for clarity (it should be clearly noted when an excerpt has been edited) and must stay close to participants’ original words or phrases, because it is inappropriate to correct grammar or change a participant’s words. We caution against using too many quotes to support a theme, as a high volume of verbatim text can be cumbersome for a reader to digest — it is the researchers’ job to explain the theme for the reader, not the reader’s job to discern the underlying meaning of multiple quotes. For a resource on how to report findings for dissemination to various audiences, see Guest et al, chapter 10 (9); for a how-to on writing thematic statements to enhance presentation and translation of findings for public health and health sciences audiences, see Sandelowski and Leeman (27).

Top

Applications of Analytic Process in Health Equity Research

The steps we have laid out are a foundation for a meaningful yet pragmatic analytic process, rather than a strict recipe for how to analyze qualitative data within the context of health equity research. Indeed, every project has different goals; thus, the application of these steps may vary considerably between projects, even those led by the same team of researchers or practitioners. In the Box, we provide brief examples of how this broad analysis process was applied to 2 studies focused on elucidating the determinants of and identifying solutions to health inequities affecting 2 different communities.

Top

Conclusion

We have shared strategies that can be used to effectively conduct qualitative analysis and generate meaningful results to inform health equity–related efforts. These strategies may be particularly useful for less-experienced health equity researchers and practitioners. Participants in health equity–focused qualitative and mixed methods studies give of their time and energy, often sharing intimate details of their needs, perceptions, experiences, and even fears. It is up to us as health equity researchers to honor these precious data by analyzing them thoroughly and with care.

As a final note, we invite readers to consider that qualitative methods in and of themselves are not aligned with the goals of health equity research. The research worldview and approach to knowledge generation of the researcher(s) and the practical goals of the research are more important than the methods used when it comes to advancing health equity through research (46). Thus, a health equity–focused research project should begin with a goal aligned with a health equity stance, such as identifying the roots of health inequities, facilitating the voices of communities affected by health inequities, or intervening on the socio-structural determinants of inequities. The selection of methods (qualitative or otherwise) and analytic strategies can then flow from said goal.

Top

Acknowledgments

The authors have no conflicts of interest to disclose.

Dr Felner is supported by the Tobacco-Related Disease Research Program of California (TRDRP) under award no. T29FT0265 and the National Institute on Minority Health and Health Disparities, National Institutes of Health (NIH), under award no. U54MD012397. Dr Henderson is supported by the National Cancer Institute under award no. K01CA248852 and award no. P30CA015704, the American Cancer Society (ACS) under award no. 00707222, and Andy Hill Cancer Research Endowment Fund. The funders or sponsors did not participate in the work. The content is solely the responsibility of the authors and does not necessarily represent the official views of TRDRP, NIH, ACS, or the Andy Hill Cancer Research Endowment.

No financial support was received for this work, and no copyrighted material was used in the conduct of this research or the writing of this article without permission.

Top

Author Information

Corresponding author: Vida Henderson, PhD, PharmD, MPH, MFA, Fred Hutchinson Cancer Center — Public Health Sciences, 1100 Fairview Ave N, Mail Stop M3-B232, Seattle, WA 98109-9024. Telephone: 206-667-6355. Email: vahender@fredhutch.org.

Author Affiliations: ¹San Diego State University School of Public Health, San Diego, California. ²Institute for Behavioral and Community Health, San Diego State University Research Foundation, San Diego, California. ³Fred Hutchinson Cancer Center, Public Health Sciences Division, Seattle, Washington. ⁴University of Illinois at Chicago, School of Public Health, Division of Community Health Sciences, Chicago, Illinois.

Top

References

1. Jeffries N, Zaslavsky AM, Diez Roux AV, Creswell JW, Palmer RC, Gregorich SE, et al. Methodological approaches to understanding causes of health disparities. Am J Public Health 2019;109(S1):S28–S33. CrossRef PubMed
2. Creswell JW. Research design: qualitative, quantitative, and mixed methods approaches. 4th ed. Sage Publications; 2014.
3. Denzin NK, Lincoln YS. Introduction: the discipline and practice of qualitative research. In: Denzin NK, Lincoln YS, editors. The Sage handbook of qualitative research. 5th ed. Sage Publications; 2018.
4. Hebert-Beirne J, Felner JK, Castañeda Y, Cohen S. Enhancing themes and strengths assessment: leveraging academic-led qualitative inquiry in community health assessment to uncover roots of community health inequities. J Public Health Manag Pract 2017;23(4):370–79. CrossRef PubMed
5. Shelton RC, Philbin MM, Ramanadhan S. Qualitative research methods in chronic disease: introduction and opportunities to promote health equity. Annu Rev Public Health 2021;43:37–57. CrossRef PubMed
6. Griffith DM, Shelton RC, Kegler M. Advancing the science of qualitative research to promote health equity. Health Educ Behav 2017;44(5):673–6. CrossRef PubMed
7. Agurs-Collins T, Persky S, Paskett ED, Barkin SL, Meissner HI, Nansel TR, et al. Designing and assessing multilevel interventions to improve minority health and reduce health disparities. Am J Public Health 2019;109(S1):S86–S93. CrossRef PubMed
8. Creswell J, Klassen A, Plano Clark V, Smith K. Best practices for mixed methods research in the health sciences. National Institutes of Health, Office of Behavioral and Social Sciences Research; 2011.
9. Guest G, MacQueen KM, Namey EE. Applied thematic analysis. Sage Publications; 2012.
10. Charmaz K. Constructing grounded theory. 2nd ed. Sage Publications; 2014.
11. Creswell JW, Poth CN. Qualitative inquiry and research design: choosing among five approaches. Sage Publications; 2016.
12. Wickham M, Woods M. Reflecting on the strategic use of CAQDAS to manage and report on the qualitative research process. Qual Rep 2005;10:687–702. CrossRef
13. Ose SO. Using Excel and Word to structure qualitative data. J Appl Soc Sci (Boulder) 2016;10(2):147–62. CrossRef
14. Meyer DZ, Avery LM. Excel as a qualitative data analysis tool. Field Methods 2009;21(1):91–112. CrossRef
15. Birks M, Chapman Y, Francis K. Memoing in qualitative research: probing data and processes. J Res Nurs 2008;13(1):68–75. CrossRef
16. Saldaña J. The coding manual for qualitative researchers. 4th ed. Sage Publications; 2021.
17. Geisler C, Swarts J. Coding streams of language: techniques for the systematic coding of text, talk, and other verbal data. The WAC Clearinghouse; University Press of Colorado; 2020.
18. Braun V, Clarke V. Conceptual and design thinking for thematic analysis. Qual Psychol 2022;9(1):3–26. CrossRef
19. O’Connor C, Joffe H. Intercoder reliability in qualitative research: debates and practical guidelines. Int J Qual Methods 2020;19:1609406919899220. CrossRef
20. Hruschka DJ, Schwartz D, St. John DC, Picone-Decaro E, Jenkins RA, Carey JW. Reliability in coding open-ended data: lessons learned from HIV behavioral research. Field Methods 2004;16(3):307–31. CrossRef
21. MacPhail C, Khoza N, Abler L, Ranganathan M. Process guidelines for establishing intercoder reliability in qualitative studies. Qual Res 2015;16(2):198–212. CrossRef
22. Skjott Linneberg M, Korsgaard S. Coding qualitative data: a synthesis guiding the novice. Qual Res J 2019;19(3):259–70. CrossRef
23. Sandelowski M. Whatever happened to qualitative description? Res Nurs Health 2000;23(4):334–40. CrossRef PubMed
24. Clarke V, Braun V. Using thematic analysis in counselling and psychotherapy research: a critical reflection. Couns Psychother Res 2018;18(2):107–10. CrossRef
25. Braun V, Clarke V. Reflecting on reflexive thematic analysis. Qual Res Sport Exerc Health 2019;11(4):589–97. CrossRef
26. Ryan GW, Bernard HR. Techniques to identify themes. Field Methods 2003;15(1):85–109. CrossRef
27. Sandelowski M, Leeman J. Writing usable qualitative health research findings. Qual Health Res 2012;22(10):1404–13. CrossRef PubMed
28. Miles MB, Huberman AM, Saldaña J. Qualitative data analysis: a methods sourcebook. Sage publications; 2018.
29. Attride-Stirling J. Thematic networks: an analytic tool for qualitative research. Qual Res 2001;1(3):385–405. CrossRef
30. Richards JE, Hohl SD, Whiteside U, Ludman EJ, Grossman DC, Simon GE, et al. If you listen, I will talk: the experience of being asked about suicidality during routine primary care. J Gen Intern Med 2019;34(10):2075–82. CrossRef PubMed
31. Sandelowski M. Real qualitative researchers do not count: the use of numbers in qualitative research. Res Nurs Health 2001;24(3):230–40. CrossRef PubMed
32. Wallerstein N, Duran B. Community-based participatory research contributions to intervention research: the intersection of science and practice to improve health equity. Am J Public Health 2010;100(suppl 1):S40–6. CrossRef PubMed
33. Minkler M. Linking science and policy through community-based participatory research to study and address health disparities. Am J Public Health 2010;100(suppl 1):S81–7. CrossRef PubMed
34. Dill LJ. Poetic justice: engaging in participatory narrative analysis to find solace in the “killer corridor”. Am J Community Psychol 2015;55(1-2):128–35. CrossRef PubMed
35. Hebert-Beirne J, Hernandez SG, Felner J, Schwiesow J, Mayer A, Rak K, et al. Using community-driven, participatory qualitative inquiry to discern nuanced community health needs and assets of Chicago’s La Villita, a Mexican immigrant neighborhood. J Community Health 2018;43(4):775–86. CrossRef PubMed
36. Switzer S, Flicker S. Visualizing DEPICT: a multistep model for participatory analysis in photovoice research for social change. Health Promot Pract 2021;22(2_suppl):50S–65S. CrossRef PubMed
37. Stoecker R. Are academics irrelevant?: Roles for scholars in participatory research. Am Behav Sci 1999;42(5):840–54. CrossRef
38. Creswell JW, Miller DL. Determining validity in qualitative inquiry. Theory Pract 2000;39(3):124–30. CrossRef
39. Raskind IG, Shelton RC, Comeau DL, Cooper HLF, Griffith DM, Kegler MC. A review of qualitative data analysis practices in health education and health behavior research. Health Educ Behav 2019;46(1):32–9. CrossRef PubMed
40. Felner JK, Dudley TD, Ramirez-Valles J. “Anywhere but here”: querying spatial stigma as a social determinant of health among youth of color accessing LGBTQ services in Chicago’s Boystown. Soc Sci Med 2018;213:181–89. CrossRef PubMed
41. Felner JK, Dyette O, Dudley T, Farr A, Horn S. Participatory action research to address aging out of LGBTQ-supportive youth programs in Chicago. J LGBT Youth 2022;19:109–34. CrossRef
42. Cragun D, Weidner A, Lewis C, Bonner D, Kim J, Vadaparampil ST, et al. Racial disparities in Racial disparities in BRCA testing and cancer risk management across a population-based sample of young breast cancer survivors. Cancer 2017;123(13):2497–505. CrossRef PubMed
43. American Cancer Society. Cancer facts and figures for African American/Black people 2022–2024. American Cancer Society; 2022.
44. Henderson V, Chukwudozie IB, Comer-Hagans D, Coffey V, Grumbach G, Spencer S, et al. Development of a culturally sensitive narrative intervention to promote genetic counseling among African American women at risk for hereditary breast cancer. Cancer 2021;127(14):2535–44. CrossRef PubMed
45. Yzer M. The integrative model of behavioral prediction as a tool for designing health messages. In: Cho H, editor. Health communication message design: theory and practice. Sage Publications; 2012:21–40.
46. Bowleg L. Towards a critical health equity research stance: why epistemology and methodology matter more than qualitative methods. Health Educ Behav 2017;44(5):677–84. CrossRef PubMed

Top

Tables

^a Estimated timeline based on total of 20 hours of work per week. May vary depending on how much time is dedicated to each step and how many team members are working on certain aspects of analysis.

Top