Data Analysis

Data analysis was based on three segments (Figure 1):

1. Phenomenographic research was used to draw a picture of how QIE/IPL is perceived. If there are moderate differences among groups, I will reflect on them. If those differences are significant, it may be possible to create separate outcome spaces for each group.
2. A case study of each professional organization involved in the research was focused on QIE/IPL-related practices and technology and the official policy used in the organization. That can help us better interpret data from phenomenographic research and gather insight into what is possible in reality. For example, in my recent research (Hlede, 2015), I found that in the ASA, all research participants indicated that IPL is the preferred way to go; however, the official policy of the organization in 2015 did not reflect that.
3. The interaction among groups and existing and potential QIE/IPL projects was analyzed through activity theory so we can get a better picture of interprofessional activities that, at this moment, shape perceptions of each profession.

Figure. Data analysis

Phenomenographic analysis. During the interviews and the analysis of the transcripts, the focus was on what the informants said about the phenomenon and how they talked about it (Larsson & Holmström, 2007) and how they perceive the relationships among them. The first step was to become familiar with each transcript and all concepts mentioned in the transcript (Figure 5).

Image. Visual display in NVivo—combining audio, text, and graphical displays of categories

Following that, during the compilation phase, passages that provided comments about QIE/IPL were tagged with short descriptions. These descriptions were grouped into categories based on concepts to which they were referring. To optimize the validity of phenomenographic research, the categories were created as logically separate and exclusive, and they correspond to a significant degree with the data from the literature on IPL/QIE and health-care reform. Therefore, as suggested by Ornek (2008), the probability of categories being considered by other researchers is high.

A unique color was assigned to each category, and an NVivo graphical display was used to track categories and sets of categories. This system helped with cross-referencing categories and estimating the theme, thematic field, and margin of each category (Sjöström & Dahlgren, 2002).

Figure 2. Object and process of phenomenographic research based on Sjöström and Dahlgren (2002) and Bowden (2005)

The outcome space is created as an image illustrating interrelations among categories. All categories are tightly connected. Their themes and thematic fields (Sjöström & Dahlgren, 2002) are of varying size in individual transcripts; however, after summarizing the transcripts, that difference was not very noticeable, even inside each interviewee group. Therefore, the outcome space was presented as a summary of all groups. The outcome spaces specific for each profession may be created in follow-up research.

Tools Used for Data Analysis

NVivo

Qualitative data analysis software (QDAS). NVivo 11 Pro for Windows, 64-bit (QSR, 2016), was used in the research. The QDAS can significantly enhance qualitative data analysis (Yuen & Richards, 1994). Although analysis and theory construction is a task for the researcher and not the software (Zamawe, 2015), the software creates an additional layer over established research methods and can alter outcomes (Paulus, Woods, Atkins, & Macklin, 2015); therefore, how the software was used is worth mentioning.

Selection. I selected NVivo because it is with ATLAS.ti, one of the two QDAS options supported by Lancaster University. The university has provided valuable lessons on NVivo best practices, and a wealth of materials is available online.

Utilization. NVivo was used for two groups of tasks. The first group covers data management and support for data analysis. The process of coding and analysis was, in many ways, identical as if it was done on Google Docs or paper. The benefit was that the process was easier and faster, and it is easier to track progress. Another group of tasks covers activities that are hardly possible with traditional pen and paper toolset. A quantitative analysis of content and data visualization is the most important example.

Data visualization. In this case, the role of NVivo is enhanced compared to traditional QDAS usage. Knowing that Paulus et al. (2015) stated that 87.5% of 763 articles they analyzed reported only the software name, it is fair to believe that in most cases, QDAS is used only to ease the process of standard phenomenographic practice and that a significant number of researchers did not master advanced features of the software. In this research, I will showcase some of the unique features NVivo provides, primarily data visualization and the utilization of dynamic connections between data from interviews and external resources. There are five main reasons for that approach:

1. Graphics can enhance the understanding of connections and differences between various elements of the complex health-care system, which are the focus of this research.
2. Data visualization is emerging as an important tool in online research (Kennedy & Allen, 2016).
3. Advanced NVivo functionality is underreported in research (Zamawe, 2015).
4. Majority of CME/CPD readers are not familiar with QDAS.
5. Contemporary visual culture expects well-visualized materials. Addressing that expectation may help bridge the gap between the methodology used in this research and a still-strong preference toward a positivist worldview among health-care professionals.

References

1. Bowden, J. A. (2005). Reflections on the phenomenographic research process. In J. A. Bowden & P. Green (Eds.), Doing Developmental Phenomenography. Melbourne, Victoria: RMIT University Press.
2. Chan, Z. C., Fung, Y.-l., & Chien, W.-t. (2013). Bracketing in phenomenology: only undertaken in the data collection and analysis process? The Qualitative Report, 18(30).
3. Hlede, V. (2015). Interprofessional Learning: Anesthesiologists’ Perspectives. Assignment, Doctoral Programme in E-Research and Technology Enhanced Learning. Department of Educational Research. Lancaster University.
4. Kennedy, H., & Allen, W. (2016). Data Visualisation as an Emerging Tool for Online Research. In N. G. Fielding, R. M. Lee, & G. Blank (Eds.), The SAGE Handbook of Online Research Methods (pp. 307-326). London, UK: SAGE Publications.
5. Larsson, J., & Holmström, I. (2007). Phenomenographic or phenomenological analysis: Does it matter? Examples from a study on anaesthesiologists’ work. International Journal On Qualitative Studies On Health And Well-being, 2(1), 55-64. doi:10.1080/17482620601068105
6. Ornek, F. (2008). An overview of a theoretical framework of phenomenography in qualitative education research: An example from physics education research. Asia-Pacific Forum on Science Learning and Teaching, 2(11).
7. Paulus, T., Woods, M., Atkins, D. P., & Macklin, R. (2015). The discourse of QDAS: reporting practices of ATLAS. ti and NVivo users with implications for best practices. International Journal of Social Research Methodology, 1-13.
8. QSR, I. (2016). NVivo 11 Pro for Windows.
9. Sandbergh, J. (1997). Are phenomenographic results reliable? Higher Education Research & Development, 16(2), 203-212.
10. Sjöström, B., & Dahlgren, L. O. (2002). Applying phenomenography in nursing research. Journal of Advanced Nursing, 40(3), 339-345. doi:10.1046/j.1365-2648.2002.02375.x
11. Yuen, H. K., & Richards, T. J. (1994). Knowledge representation for grounded theory construction in qualitative data analysis. Journal of Mathematical Sociology, 19(4), 279-298.
12. Zamawe, F. C. (2015). The Implication of Using NVivo Software in Qualitative Data Analysis: Evidence-Based Reflections. Malawi Medical Journal, 27(1), 13-15.