DSRT-839: Advanced Research Methods : Week 1 - IRB/Research Proposal Draft - Part I
vutukuri
UC-Dissertation-Template--APA-7th-Ed.-QualitativeV5.30.24.docx
QUALITATIVE DISSERTATION 1
Your Approved Dissertation Title Here in Upper and Lowercase Letters
First and Last Name
Submitted to the Faculty of the Graduate School
in Partial Fulfillment of the
Requirements for the Degree of
[insert degree]
University of the Cumberlands
Month and Year of Graduation
Table of Contents Chapter Three 4 Procedures and Methodology 4 Introduction 4 References 13
List of Tables
Table 1: Name of the Table…………………………………………………………………1
List of Figures
Figure 1: Name of the Figure …………………………………………………………………1
Chapter Three
Procedures and Methodology
Introduction
This research focuses on explaining the methods and procedures used in studying the opinions of healthcare professionals regarding the use of machine learning (ML) to ensure the privacy of electronic health records (EHRs). This chapter describes the design of the research, the way samples were chosen, how data were gathered, how the data were analyzed, and important aspects of trustworthiness, reflexivity, and ethics. This type of research needs transparency and careful work to make sure its findings can be used elsewhere and are credible. The study is supported by various sources in healthcare cybersecurity and machine learning (She, et al., 2022). Also, it aims to explore the real-life and moral consequences of using ML to defend EHRs. This section begins by explaining the value of the qualitative approach and its paradigm, then provides a clear outline of the methodology used in the case study. Sections after this outline how the researcher influences the research, explain the approach to choosing participants, describe data gathering, analyze findings, and discuss the reliability of the results.
Research Method and Paradigmatic Perspective
The research design used in this study is qualitative, as it helps to study detailed, real-life experiences and perceptions. While quantitative research looks to measure facts and prove hypotheses, qualitative research is based on personal explanations and interprets how individuals think about their experiences (Vu, 2021). Since the main research question relates to subjective feelings and opinions, qualitative research is the best approach.
It follows a constructivist-interpretivist framework, believing that reality is shaped by social interactions and a person’s situation. This kind of research values what people involved experience and think, focusing on the meaning behind the use of ML in healthcare security. The current study uses She et al. (2022) as a starting point, which studied how well ML works in healthcare but did not consider views from professionals (She, et al., 2022).
Qualitative Research Approach: Case Study Methodology
This research chooses a qualitative case study approach since it aims to study real-world views in the healthcare sector. Case study research helps examine a whole system, namely the healthcare sector, involving ML for securing electronic health records (Seh, et al., 2022). It helps the researcher obtain valuable data related to the experience and beliefs various stakeholders have with ML tools.
This methodology fits well with the analysis because it highlights the subtle moral and workforce factors that numbers alone don’t show. Both phenomenology and grounded theory were discussed, yet phenomenology explores what is common in many experiences, while grounded theory is meant to develop new theories. Since the study is focused on current practices and perceptions in one place, the case study method is the best choice.
Trustworthiness
To make qualitative research trustworthy, one should confirm that it is credible, dependable, transferable, and confirmable (Ahmed, 2024).
Credibility: To make the research more credible, the study asked participants to look at the transcripts and initial findings and point out any errors. People from various roles in healthcare, including IT, clinical, machine learning, and data, were brought together to gain different insights. In addition, the interview process was examined by two expert professors to see whether it matched the aims of the research.
Dependability: The dependability of the research was maintained by having a detailed audit trail of everything from collecting data to making coding and analysis decisions. NVivo software helped in keeping the data organized and evenly handled (Allsop, Chelladurai, Kimball, Marks, & Hendricks, 2022).
Transferability: Even though statistical generalization is not the goal of qualitative research, thick descriptions of participants’ experiences at work and within the organization were still mentioned. Survey respondents were chosen to include people from a variety of backgrounds and positions.
Confirmability: The researcher used a reflexive journal to record all their thoughts, decisions, and views during the study. It allowed researchers to rely on participants’ views, not their own, when making sense of the data.
Role of the Researcher
In this qualitative case study, the researcher controlled each stage, including setting up the study, recruiting study subjects, leading the interviews, and interpreting the results. The researcher collected data using semi-structured interviews to explore how participants felt about ML helping to protect EHR data. Tasks involved creating the interview plan, making sure the IRB approved it, transcribing the interviews, and performing a thematic analysis.
Researcher Positionality
Healthcare technology and data ethics are important areas of both study and work for the researcher (Hensen, et al., 2021). The researcher points out that they have a previous interest in digital health, which could influence how the findings are interpreted. To reduce bias, the researcher used introspection and created a journal for self-reflection. The research team valued patient privacy and data security, yet they set these values aside to make sure participants were heard first.
Reflexivity and Bracketing
The researcher used reflexivity and bracketing to try to avoid influencing the research results. Reflexivity meant that the researcher constantly thought about how their actions might influence the research (Kimberley, 2021). Bracketing was achieved by noting preconceptions about ML effectiveness in a journal and intentionally setting them aside when analyzing interview data. Using this method ensured the main themes were based on what the participants shared instead of what the researcher imagined.
Sampling Procedures and Data Collection Sources
Sampling Strategy
Participants were intentionally chosen based on having experience working with or using ML for safeguarding EHRs. To be part of the study, individuals had to be healthcare IT professionals, clinicians, or data security officers working with ML. Only those who did not have a direct role in EHR security or ML work were excluded. By using this strategy, the data collected was useful and informative.
Instrument Development
The semi-structured type of interview was developed for this study (Ruslin, Mashuri, Rasak, Alhabsyi, & Syam, 2022). The research questions, theories, and previous research guided the development of the protocol. The research examined the roles individuals played, their grasp of machine learning, visible results, and any ethical issues. Open-ended questions were included in the protocol to make it more flexible and thorough.
Table 1: Mapping of Interview Questions to Research Questions and Existing Literature
|
Interview Question |
Research Question |
Literature Base |
||
|
1. Can you tell me about your current role and experience with ML in EHR security? 2. Have you had any direct experience with machine learning technologies in healthcare settings? |
RQ1: What types of machine learning algorithms are used to ensure data confidentiality in EHRs? |
Machine learning methods such as SVM, KNN, and Decision Trees have been applied to secure EHRs, offering advantages over traditional security mechanisms (She et al., 2022). Interviews enable exploration of technical roles in real-world settings (Gill et al., 2008). |
||
|
3. What is your understanding of how machine learning is used to secure electronic health records? 4. In your opinion, what are the advantages of using machine learning over traditional methods (e.g., encryption, access control)? |
RQ2: How effective are these algorithms when tested on real or synthetic healthcare data? |
Machine learning has proven effective in identifying patterns of cyber threats, especially when trained with synthetic datasets (Masood & Sonntag, 2020). Interviews allow professionals to express insights beyond measurable data (Tracy, 2024). |
||
|
RQ3: What performance metrics best measure the success of ML algorithms in identifying and preventing threats? |
Effectiveness is often assessed using metrics like accuracy, precision, recall, and F1 score (She et al., 2022). Qualitative insights help understand why certain metrics are favored in operational environments (Karatsareas, 2022). |
||
|
7. What concerns, if any, do you have about using machine learning to protect patient data? 8. How is patient confidentiality maintained while using these technologies? |
RQ4: What concerns exist regarding ML and patient data confidentiality? |
Trustworthiness in qualitative research requires consideration of ethical risks, including confidentiality and data misuse (Lincoln & Guba, 1985). Participants provide nuanced perspectives on ethical implications (Tracy, 2024). |
||
|
9. What improvements would you suggest for the current use of machine learning in your organization? 10. Do you see these technologies becoming a standard practice across the healthcare industry in the next five years? |
RQ4: What concerns exist regarding ML and patient data confidentiality? |
Healthcare professionals may hesitate to fully adopt ML without ethical clarity and governance frameworks (Karatsareas, 2022). Semi-structured interviews facilitate the discussion of future readiness and ethical comfort (Tracy, 2024). |
Expert Review and Field Test
One professor from the University of Cumberlands had expertise in qualitative methods, and the other had expertise in healthcare cybersecurity. Both went over the interview guide, came up with suggestions to improve wording, and advised reshuffling some of the interview questions. The suggestions from reviewers were included in the final protocol.
The study included two participants who did meet the criteria but were excluded during the field test. The field test showed that the questions were clear and met the expectations (Fajeriadi & Irhasyuarna, 2021). A few changes were made, and the interviewer learned how to conduct follow-up questions and move between subjects. The field test went ahead only after receiving IRB approval.
Participant Recruitment
Professional healthcare and IT organizations, along with LinkedIn, were used to invite people to take part in the survey. The scripts and consent forms were shared through email and WhatsApp. People who wanted to participate filled out a quick screening form. Participants were free to pick interview times, and video conferencing was used to keep the interviews secure.
Data Collection
The data was collected by asking open questions in semi-structured interviews that lasted around 45–60 minutes. Before each session, informed consent was taken from the participants. Every interview was recorded, and the text was prepared by a professional transcription company. After checking for accuracy, participants were allowed to review their transcripts to confirm they were correct.
Data Saturation
Data saturation took place after interviewing 15 individuals because no additional themes or ideas appeared. The same patterns appeared in various positions and areas of work. The process of collecting data came to an end when the team identified that the answers were repeating and all the research questions had been answered.
Data Analysis
Data Preparation
All of the interview transcripts were imported into NVivo 12 for analysis. All transcripts were confirmed to be accurate and were made anonymous. While reviewing transcriptions, preliminary notes were taken and used for guidance while coding the data.
Coding and Theme Development
A thematic analysis method was applied to understand the findings in the study (Morgan, 2022). All transcripts were coded at first, and the data were then organized into wider themes. In the beginning, in vivo and descriptive coding were applied to save the way participants used language. Codes were repeatedly reviewed, put into groups, and refined. Codes were arranged, frequency was kept track of, and theme relationships were shown using NVivo. The themes found connected smoothly to my research questions and were confirmed by selected quotes.
Summary
This chapter covered the procedures used to study how healthcare professionals view machine learning with regard to EHR confidentiality. A qualitative case study method was used, guided by a constructivist approach. It included explanations on how to sample, form instruments, gather and process data, and ensure trustworthiness. The next chapter explains the study’s findings, divided into themes based on the participant interviews.
References Ahmed, S. K. (2024). The pillars of trustworthiness in qualitative research. Journal of Medicine, Surgery, and Public Health, 2, 100051. https://doi.org/10.1016/j.glmedi.2024.100051 Allsop, D. B., Chelladurai, J. M., Kimball, E. R., Marks, L. D., & Hendricks, J. J. (2022). Qualitative methods with NVivo software: A practical guide for analyzing qualitative data. Psych, 4(2), 142-159. https://doi.org/10.3390/psych4020013 Fajeriadi, H., & Irhasyuarna, Y. (2021). The practicality of natural science learning devices on the concept of environmental pollution with problem-solving learning models. Journal of Physics: Conference Series, 012025. https://iopscience.iop.org/article/10.1088/1742-6596/2104/1/012025/meta Hensen, B., Mackworth-Young, C. R., Simwinga, M., Abdelmagid, N., Banda, J., Mavodza, C., & Weiss, H. A. (2021). Remote data collection for public health research in a COVID-19 era: ethical implications, challenges and opportunities. Health policy and planning, 36(3), 360-368. https://doi.org/10.1093/heapol/czaa158 Kimberley, A. (2021). Reflexivity as a vital skill for future researchers and professionals. Electronic Journal of Business Research Methods, 19(1), 14-26. https://doi.org/10.34190/ejbrm.19.1.2124 Morgan, H. (2022). Understanding thematic analysis and the debates involving its use. The qualitative report, 27(10), 2079-2090. https://doi.org/10.46743/2160-3715/2022.5912 Ruslin, R., Mashuri, S., Rasak, M. S., Alhabsyi, F., & Syam, H. (2022). Semi-structured Interview: A methodological reflection on the development of a qualitative research instrument in educational studies. IOSR Journal of Research & Method in Education (IOSR-JRME), 12(1), 22-29. DOI: 10.9790/7388-1201052229 Seh, A. H., Al-Amri, J. F., Subahi, A. F., Agrawal, A., Pathak, N., Kumar, R., & Khan, R. A. (2022). An analysis of integrating machine learning in healthcare for ensuring confidentiality of the electronic records. Computer Modeling in Engineering & Sciences, 130(3), 1387-1422. DOI: 10.32604/cmes.2022.018163 She, A. H., Al-Amri, J. F., Subahi, A. F., Agrawal, A., Pathak, N., Kumar, R., & Khan, R. A. (2022). An analysis of integrating machine learning in healthcare to ensure the confidentiality of electronic records. Computer Modelling in Engineering & Sciences, 130(3), 1387–1422. https://edepot.wur.nl/579611 Vu, T. T. (2021). Understanding validity and reliability from qualitative and quantitative research traditions. VNU Journal of Foreign Studies, 37(3), 88. https://doi.org/10.25073/2525-2445/vnufs.4672
