Qualitative case studies: Boundaries and data sources

Studies that define case scope and context together

A case study is a research design that explores programs, organizations, individuals, events, or processes in depth as a single unit of analysis in real-life contexts. What constitutes the case determines which participants are included, the period of investigation, the activities observed, and the documents collected. Defining the boundaries of a case involves specifying which experiences and contexts will be investigated together to answer the research question. Interviews, observations, documents, and quantitative data reveal different aspects of a case.

Methodological traditions and common design elements of case studies

A literature review of the history and methodological orientations of case studies compares approaches developed in the social sciences, education, business, law, and healthcare (Harrison et al., 2017). Major antecedents include early 19th-century biographical research, early 20th-century anthropological fieldwork, and observational studies by the Chicago School of Sociology from the 1920s to the 1950s. Experimental and statistical research became dominant from the late 1940s onward. Interest in systematic qualitative analysis and demand for educational program evaluation supported the development of case studies. Evaluations of curricula and innovative programs in the 1970s created a need to address participants’ perspectives alongside sociopolitical contexts. The review compares definitions, philosophical stances, case selection, and analytical procedures across three major methodological approaches and synthesizes their common elements.

The approach with postpositivist characteristics examines theoretical propositions and rival explanations. When selecting multiple cases, it distinguishes between literal replication, which predicts similar results, and theoretical replication, which predicts contrasting results. The approach combining constructivist understanding with pragmatic research procedures defines a case as a bounded system. Its particularistic, descriptive, and heuristic characteristics emphasize description and understanding of a specific subject. This approach uses inductive reasoning, thematic and content analysis, triangulation, and a chain of evidence leading from data to interpretation. The interpretive approach values the uniqueness and complexity of cases, as well as meaning co-constructed by researchers and participants. It uses direct interpretation, categorical aggregation, vignettes, and thick description. Intrinsic studies are driven by an interest in the case itself; instrumental studies aim at understanding a broader issue; and collective studies involve multiple cases. The quantitative–qualitative continuum illustrates these differences. Understanding designs requires examining each approach’s philosophical premises and analytical procedures together.

Common design elements include bounded cases, real-life contexts, in-depth exploration, purposeful case selection, and multiple sources of evidence. Cases are bounded by time, place, and activity. Designs may involve single or multiple cases and may be holistic, covering the entire case, or embedded, including sub-units of analysis. Interviews, observations, focus groups, documents and artifacts, and questionnaires and surveys are combined according to the research question. Connecting the philosophical stance, research questions, case selection, and data collection and analysis helps explain the rationale for adopting individual methods. Methods are procedures and techniques, while methodology is the researcher’s perspective and basis for decision-making. Researchers examine their involvement through memos and research journals that document how the collected data lead to interpretation.

Structuring boundaries and data sources in healthcare case studies

A methodological paper examining design and implementation issues across four healthcare studies describes case study research as an approach to understanding policy, service changes, and learning experiences in real-life contexts (Crowe et al., 2011). The first study explored why rates of minority ethnic participation in asthma research varied in the UK and the US. In-depth interviews and supplementary questionnaires with researchers from both countries were analyzed using the Framework approach. Support for ethnically diverse participation in the UK coexisted with practical concerns. Researchers also expressed stereotypes about language barriers and a lack of altruism. In the US, the National Institutes of Health policy mandating the inclusion of minorities and women shaped researchers’ attitudes and experiences. The second study compared workforce reconfiguration in respiratory services across four Primary Care Organisations (PCOs) in England and Wales. During screening, telephone interviews were conducted with representatives from 30 PCOs. Data sources for the selected cases included interviews, email discussions, documents, patient diaries, website news, and national workshop materials. Financial deficits, organizational uncertainty, a lack of clinician engagement, and conflicting policies complicated the development of sustainable services. The turbulence of the 2006 organizational restructuring produced outcomes that diverged from the original aim of strengthening purchasing and commissioning functions. Negotiations over the ownership of work and income, professional interests, and relationships established through past collaborations also influenced implementation.

The third study was a longitudinal mixed-methods study of five National Health Service (NHS) acute hospitals and mental health Trusts during the early implementation of electronic health records. The NHS launched the initiative in 2002 to implement electronic health records in all hospitals in England by 2010. The research team gathered data through interviews, documents, field notes, observations, and quantitative measures. Analysis used a sociotechnical coding framework encompassing technology, people, and organizations, alongside data-driven themes. It showed that implementation progressed more slowly than anticipated. The analysis identified demands for greater choice for hospitals and negative consequences of centralized contracts that excluded hospitals. Stakeholders expressed frustration over the politically determined ten-year implementation timeline. Case selection considered hospital types among early adopters and the diversity of three centrally procured systems. Research at one institution where access was difficult was halted to focus on the others. The fourth study was a multi-site mixed-methods study of patient safety learning among medical, nursing, physiotherapy, and pharmacy students. The patient safety study selected eight cases in total, two per professional group, to investigate educational, organizational, and practice contexts. Phase 1 examined curriculum documents, educators’ and students’ views, and observational data. Phase 2a examined organizational policies and managers’ perspectives. Phase 2b examined the culture of daily practice. Students valued informal learning in practice, but opportunities to participate in risk management and incident reporting activities were limited. Patient safety was largely implicit in curricula. Resources, peer pressure, and the quality of supervision influenced safe practice.

Defining a case involves jointly determining when the research begins and ends, the population, organization, or region under study, and priorities for evidence collection and analysis. In research on the implementation of electronic health records, an NHS Trust can serve as the case when the focus is on technological implementation. A group of healthcare professionals can serve as the case when the focus is on social and organizational experiences. The selection of typical and atypical cases, and the choice between single-case and multiple-case designs, depend on the purpose of inquiry, theory, accessibility, and the burden of participation. Analysis begins with an understanding of individual cases before comparing similarities and differences across cases. The Framework approach organizes data in five stages: familiarization, identifying a thematic framework, indexing, charting, and mapping and interpretation. Triangulation is used to examine the same phenomenon across multiple data sources and to explore different explanations. In an evaluation of a cancer care program discussed in the paper, surveys conducted before the intervention and five months after it began showed no statistically significant results. Patient interviews confirmed perceptions of increased participation in care and potential benefits from the program process. Interpreting differences between cases requires considering how institutions vary in data access, the timing of data collection, and available analytical resources. The PCO study did not follow the organizational restructuring through to completion. The electronic health records study had limited access to details of contracts between the government and suppliers. The applicability of its results was assessed in relation to the characteristics of early adopting institutions. In the patient safety study, gatekeepers may have influenced access to documents. Difficulties recruiting focus groups led to a shift to interviews and paired discussions. Analytical time was limited relative to the volume of data. The paper suggests using theory while examining unanticipated themes and alternative explanations. It also recommends participant review, researcher reflexivity, anonymization that accounts for identification risks, and documentation of how conclusions were reached.

Reading notes

From the perspective of workplace learning in HRD, learners’ accounts of what they have learned can be examined in relation to the tasks they participated in and the roles they held. Whether the institution or the professional group serves as the case also changes what is compared and which data are required. If activities described in documents are rarely observed, follow-up questions can clarify when the activities take place, the conditions for participation, and the observation period.

References

• Crowe, S., Cresswell, K., Robertson, A., Huby, G., Avery, A., & Sheikh, A. (2011). The case study approach. BMC Medical Research Methodology, 11(1), 100. https://doi.org/10.1186/1471-2288-11-100
• Harrison, H., Birks, M., Franklin, R., & Mills, J. (2017). Case study research: Foundations and methodological orientations. Forum Qualitative Sozialforschung / Forum: Qualitative Social Research, 18(1), Art. 19. https://doi.org/10.17169/fqs-18.1.2655