Immersion–Crystallization: a valuable analytic tool for healthcare research

Abstract

This paper describes Immersion–Crystallization, a valuable qualitative analytic method, and the steps needed to undertake it. Qualitative and mixed methods research have significant applicability for studying healthcare issues and are widely utilized in a variety of disciplines. Qualitative data collection such as long interviews, focus groups, and participant observation are fairly straight forward and clinician-researchers may find these methods similar to their healthcare activities such as history taking, family meetings, and clinical observations. However, clinician-researchers commonly find that qualitative data analysis is complicated and hard to undertake. Making matters worse, many anthropological and sociological studies do not even mention how data was analysed and interpretations reached. As noted by Miller and Crabtree,1 “Unfortunately, the language used by qualitative researchers, especially with regards to analysis, is often obscure.” These barriers may stymie the broad use and application qualitative research in healthcare. The first known publication focussing exclusively on Immersion–Crystallization appeared in 1999, authored by Borkan, though in 1992 Miller and Crabtree2 had included Immersion–Crystallization as one of the “four idealized analytic styles.” In 2006, the Robert Wood Johnson Foundation sponsored the Qualitative Research Guidelines Project which included a website for people developing, engaging, and evaluating qualitative research projects in healthcare settings. The main authors of the Project, Cohen and Crabtree,3 provided particularly instructive definitions of immersion and crystallization that involve dual process cycles. Over the past 20 years, Immersion–Crystallization has been utilized widely among qualitative researchers in healthcare and elsewhere. A 2021 PubMed search revealed more than 215 articles that specifically identify the use of Immersion–Crystallization and the number is growing yearly (PubMed, NIH—National Library of Medicine4). Published articles utilizing this analytic method derive from a broad range of disciplines ranging from Family Medicine5 to Pediatrics and Internal Medicine,6 Pharmacy (Peters7), to Nursing8 and beyond. At its core, Immersion–Crystallization is an inductive, iterative process for identifying themes, categories, and patterns in the data. Inductive refers to reasoning that works from specific observations to generalizations, while iterative connotes a systematic, repetitive, and recursive process.9 As seen in Fig. 1, the qualitative research iterative cycle involves repeated rounds of data collection, data analysis, and data representation, with interpretation emerging from and influencing the other elements. Representation refers to the description, depiction, or portrayal of the data in a written, graphical, visual, audio, or other media/multimedia format. Immersion–Crystallization adheres to this iterative model, as opposed to the equivalent quantitative research method which is linear and involves hypothesis testing. Qualitative research iterative cycle. This figure portrays the iterative cycle that is the basis of all qualitative research. On the most basic level Immersion–Crystallization can be conceptualized as involving repeated exposure and probing into data leading to the emergence of meaningful themes and categories (see Fig. 2). Further analysis coalesces these themes and categories into patterns and interpretations. Iterative cycles continue until analytical saturation occurs and final, reportable interpretations are reached. Analytical saturation refers to achieving the state whereby no additional significant insights, patterns, or interpretations emerge from the data. Immersion–Crystallization process. This figure lays out the process of Immersion–Crystallization in an iterative manner with entry and exit points. Immersion–Crystallization, like other qualitative analytical techniques occurs, “before data collection, during study design and planning; during data collection as interim and early analyses are carried out; and after data collection as final products are approached and completed.”10 Immersion–Crystallization does not require the fastest computer or new software. It does demand patience, an ability to be contemplative and a facility to “listen deeply”11 to individuals, groups, organizations, and to the data itself. A broad range of data sources can be analysed using Immersion–Crystallization. These include the usual data types collected in qualitative research, such as interview and focus group transcripts, field notes from participant observation, and original texts. It can also be applied to audio and video recordings, cultural artefacts, and even quantitative survey and measurement databases. The only limitations are familiarity with the data source, its underlying context and theoretical assumptions and underpinnings. In describing the process of Immersion–Crystallization, although there are several distinct elements, it is best to consider it a flexible, iterative framework with guideposts and stages rather than a series of linear rungs. As noted by Creswell and Poth,12 “The processes of data collection, data analysis, and report writing are not distinct steps in the process—they are interrelated and often go on simultaneously in a research project.” Each stage within the Immersion–Crystallization process has multiple points of entry and exit, and they will vary with the researcher and the topic. Building on prior work by Moustakas,13 Miller and Crabtree,1,2 Huberman and Miles,10 and Borkan,14,15 the elements can be described as follows: Initial engagement: Analysis begins when a topic of study is chosen the focus of investigation and continues during study design. Practically speaking, this means that as researchers conceptualize and plan their research, initial meanings, themes, and categories emerge. The adroit researcher would be wise to jot these down, with the caveat that they are initial impressions and interpretations and subject to change as further iterative cycles occur. Reflexivity: Reflexivity refers to the examination of the researcher’s own beliefs, biases, judgements, and practices and how these may influence the research process and the findings. Qualitative studies require openness and an acceptance that the researcher is part of the research16 and that the researcher’s subjectivity influences data collection and analysis. Researchers, both individually and as a team, should explicitly consider the influence of these elements. They can explicitly describe and record them starting early in the research cycle and return to them as the research progresses. Howard Stein, a noted medical anthropologist who worked many decades in Family Medicine departments, encourages researchers to process one’s own biases and responses, using one’s being as both recorder and filter: The way I interpret Slovak American steelworkers’ life or Oklahoma wheat farming family life is that I look at myself as I look at them—not through some kind of “anthroscope.” I let them enter into my imagination. I don’t remove myself; I engage myself. We [qualitative researchers including anyone doing qualitative research] record raw data; we filter and record through our selves… The only way to understand a person is with your whole self.11 The reflexive nature of Immersion–Crystallization can even be an intersubjective experience which connects the researched with the researcher through the medium of the data. Immersion: This is the part of process wherein researchers immerse themselves in the data they’ve collected through reading, listening to, watching or examining.3,14 Immersions may be performed individually or in teams. Though early insights may arise during topic selection and research design, the first immersion into the data is likely to occur at the time of data collection. Insights from these early immersions are valuable in themselves though they can assist researchers with further focussing questions and refining the data collection plan. For example, at the end of any focus group, the moderator and observer can record themes and patterns that they felt emerged and use these insights to improve the moderator’s question guide for the next group. Similar insights can help refine the open-ended questionnaire in long interview studies. Even participant observation inquiry can be improved by initial immersion cycles. For example, in a study of occupational health and safety, key informants’ reports about which industrial processes they thought were the most dangerous can help the researcher’s shift their focus to these areas. Further immersion occurs at each stage of the research and during each iterative cycle. Crystallization: Crystallization is the process of temporarily suspending the process of examining the data (immersion) in order to reflect on the analysis experience and attempt to identify and articulate patterns or themes noticed during the immersion process.3 As noted above, the insights from crystallization can facilitate improvements in data collection. They also help with articulating ever clearer and more insightful interpretations. Crystallization can be performed individually or in teams. As demonstrated in the example in the Appendix, there are significant advantages of working in teams, not the least of which is the potential for dynamic interaction in the elucidation of themes, patterns, and interpretations. Creative synthesis: Although researchers may be excited by their ability to crystallize out meaningful themes, there are further creative synthesis efforts required to move from these “building blocks” to higher order patterns, interpretations, and theory. Pattern recognition is key to interpretation and involves making sense of the data or the “lessons learned.”17 In qualitative research, “Interpretation involves abstracting out beyond the codes and themes to the larger meaning of the data.”12 Creative synthesis may also involve contrasting themes and patterns to findings from other studies and consideration of where the findings stand in regard to particular theories. Pattern recognition can be daunting to the inexperienced. Though Immersion–Crystallization clearly involves left brain rationalistic analytical inquiry, it also involves “experiencing” the data in a more holistic, right brain approach, not unlike what is encountered in close study and appreciation of music and art. Moving from themes to patterns and interpretations is akin to not just knowing the notes but appreciating how they come together in a musical arrangement. Practically speaking, this synthesis can be facilitated by grouping themes into like categories. Redundancy can be eliminated through collapsing themes with similar meanings. Patterns may easily emerge at this point, or will upon further examination, reflection, and immersion. Corroboration and the search for alternative interpretations: After patterns and interpretations are reached further Immersion–Crystallization is required in order to corroborate the findings and to search for alternative interpretations. The search for alternative interpretations is of paramount importance and increases rigor. There is no single method to achieve this. For example, it can be done by examining more data using Immersion–Crystallization or other employing other analytical tools. If a research team is involved, it may be very valuable to include group Immersion–Crystallization followed by discussion. Member checking is also advised. This involves providing feedback to research participants regarding insights, patterns, and interpretations from Immersion–Crystallization. Such approaches may facilitate expanded interpretations, in addition to checks on corroboration, authenticity accuracy, and credibility. Arriving at final interpretations: The determination that one is at the end of the Immersion–Crystallization process for a particular study can be tricky to recognize, though there are some parameters that may be helpful. As Cohen and Crabtree3 note, “These dual processes continue until all the data have been examined and patterns and claims emerge from the data that are meaningful and can be well articulated and substantiated.” Similar to data saturation, there is interpretation saturation, noted above. after which no further significant interpretations arise. Immersion–Crystallization may be used as the sole analytic method in a study, it may be the first of multiple analytic steps, or it may be done simultaneously with other analytic methods. For example, some qualitative researchers have used Immersion–Crystallization to develop codebooks for further data analysis while mixed method researchers may combine it with their quantitative work. Reporting: Once the researcher or research team feel they have reached final interpretations, reporting can occur via a wide range of media—e.g. oral presentations, posters, videos, blogs, manuscripts, etc. Immersion–Crystallization is not a dogmatic process and significant variations occur both within the approach and between it and other qualitative analytic methods. As with much of qualitative research, the specifics of any particular analytic approach are less important than a dedication to rigor, subjectivity, and contextual grounding. This type of analysis may be daunting to the novice, so mentors or guides and prior knowledge of the route may be reassuring, though certainly not required. Qualitative research methods, including Immersion–Crystallization, are sometimes critiqued regarding validity and reliability. Although a full discussion is beyond the scope of this manuscript, suffice it say that these concepts arise from quantitative research with its grounding in statistics and are not easily applicable to qualitative methods. Even the terms may be different—such as “trustworthiness”17 in place of reliability. Further discussion is available from Creswell and Poth12 and Whittemore et al.18 Qualitative research is popular in healthcare; however investigators often find the analysis of qualitative data to be a daunting enterprise. Immersion–Crystallization is a valuable analytic approach for healthcare research. It provides a strategy to move from a research question and raw qualitative data to insights in the form of themes, patterns, and interpretations that can be succinctly reported in presentations and publications. I encourage its further adoption and wide utilization for current and future investigations. None declared. The author would like to thank Professor Roberta Goldman of the Department of Family Medicine at Brown University for her insightful comments, edits, and inspiration. He would also like to thank Professor William Miller and Professor Benjamin Crabtree for their lifetime of dedication to qualitative research in Family Medicine. In addition, thanks are due to the Committee on Advancing the Science of Family Medicine (CASFM) and Professor Jodi Holtrop for initiating this Research Briefs project. None declared. The data underlying this article are available in the article. This appendix provides an actual example that demonstrates the principles of Immersion–Crystallization. The diagnosis and treatment of low back pain (LBP) in primary care have been complicated, expensive, and often unsatisfactory to both patients and physicians. Until the 1990s most knowledge on LBP stemmed from orthopaedic surgery studies which were not applicable to primary care settings. To help understand LBP in primary care, a focus group study was conducted that recruited patients from Family Medicine practices. The groups were run by a moderator with an observer in attendance and sessions were audio-recorded, transcribed, and both recordings and transcriptions were distributed to the research team. In addition, the moderator and observer recorded their initial impressions at the end of each focus group, and these were also distributed to the research team. Immersion–Crystallization was chosen as the analytic method. To begin, each member of the research team was asked to individually immerse themselves in the data over the course of 6 weeks—listen to the recording, review the transcripts and record themes. Next, the research team met on 3 occasions to interpret the material together. To get the members of the team into the proper mindset for Immersion–Crystallization analysis, each meeting started with listening to sample recordings from focus groups that had particularly rich dialogue. They were then asked to review their list of themes and to modify, add or subtract prior to discussion. Once ready, all themes were recorded on flip charts and as each page was completed, they were taped onto a wall. After listing, themes were grouped together by category and any redundancies were eliminated. The next step was to search for patterns and overarching interpretations in the themes. These were also recorded on flip charts and displayed. The group then sought alternative interpretations, asking whether the themes and patterns might be interpreted in a different manner. This process continued until no further interpretations were generated. The published study14 produced a patient-centred primary care classification system of backache symptoms based on typical pain intensity, dysfunction, duration, and treatment. It also challenged the dominant typology of “acute and chronic” pain and suggested a more dynamic and variable model and emphasized the important distinction between level of pain and level of function.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 126 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.