Abstract
With the growth of qualitative health research in low- and middle-income countries, local health professionals are increasingly involved in facilitating interviews with their fellow health workers. Understanding the methodological implications of such situations is required to ensure high-quality study findings and to build capacity and skills for interviewers with clinical backgrounds working with limited resources. This article reports a qualitative process evaluation of a study that assessed barriers and enablers of implementing bubble continuous positive airway pressure in Malawi. Findings were summarized through an iterative process of reflection on what worked, what did not work, areas for improvement, structural challenges, negotiating dual roles as nurses and researchers and the professional hierarchy within the health care system. Comprehensive practical training was critical to conducting qualitative research in a health setting. Interviewers were health workers themselves and required skills in reflexivity to effectively probe and navigate interviewing other health professionals, including senior staff. The main challenge in conducting interviews in a resource-limited healthcare setting was time constraints, which were compounded by staffing shortages. Lessons from this qualitative evaluation highlight the importance of training in reflexivity, engaging interviewers as collaborators and reserving adequate time to accommodate healthcare workers’ multiple roles and responsibilities.
Keywords
Introduction
Qualitative research is becoming increasingly important in the health sector for contextualizing social phenomena and enriching quantitative findings. (Al-Busaidi, 2008; Mack et al., 2005; Ruark & Fielding-Miller, 2016). Qualitative methods are used to uncover meanings, perspectives, attitudes, and experiences (Chafe, 2017; Creswell & Clark, 2006; Mack et al., 2005) and can provide insights into intervention implementation during feasibility studies, process evaluation, and program evaluation (Holtrop et al., 2018).
Interviews are widely used to access insights into participants’ perceptions and lived experiences in line with study objectives globally (Alsaawi, 2018; Barrett & Twycross, 2018). For qualitative research in a clinical setting, using health professionals and social scientists to facilitate interviews is common (Hunt et al., 2011). In such situations, the interviewer and participant share professional characteristics (Quinney et al., 2016). The relationship between interviewer and participant can influence study findings due to power dynamics embedded within medical hierarchies (Coar et al., 2006). In addition, interviews in a hospital setting may affect the behavior and responses of interviewers and interviewees (Quinney et al., 2016).
When health professionals are used as interviewers for other health care workers in clinical settings, they must understand the methodological implications. Studies conducted in high-income countries (HICs) have explored the roles of research nurses, the challenges they encounter and their impact on clinical research (Hernon et al., 2019; Larkin et al., 2017; Tinkler and Lisa, 2020). However, a gap exists in the methodological implications of utilizing health workers as interviewers for qualitative health research in low-and middle-income countries (LMICs) and the process of strengthening the capacity of local researchers to conduct interviews.
Workforce shortages, limited resources and health infrastructure gaps are serious concerns for delivering health services in LMICs, and we seek to explore their implications for qualitative health research (Bangdiwala et al., 2010; O’Brien & Gostin, 2011). The purpose of this methods paper is to document the qualitative process evaluation of a recently completed qualitative healthcare research study in Malawi. Local capacity for qualitative research in our country is relatively limited. Systematically documenting methodological issues encountered during qualitative studies can help inform and strengthen future research projects.
Methods
Qualitative Evaluation of a Data Collection Process
This is a qualitative process evaluation of the “Integrating a neonatal healthcare package for Malawi” primary study, which is part of the Innovating for Maternal and Child Health in Africa (IMCHA) initiative. Qualitative evaluation is a method of assessment that aims to gain an in-depth understanding of a program or process, focusing on the why and how questions (Patton, 1999). This qualitative evaluation is based on the experiences and observations of interviewers with a health background who participated in data collection with health worker professionals in a clinical setting in Malawi. We formalize this qualitative evaluation by analyzing reflections and observations on our experiences preparing for and conducting interviews with health professionals in a health care setting and providing lessons learned to inform and strengthen future research in similar settings.
Description of the Primary “Integrating a Neonatal Healthcare Package for Malawi” Study
The primary descriptive qualitative study assessed the barriers and enablers of implementing bubble continuous positive airway pressure (CPAP) from the perspective of health professionals in Malawi. Bubble CPAP supports breathing for newborns with respiratory distress syndrome (Nyondo-Mipando et al., 2020). Affordable and robust bubble CPAP systems have been designed for low-resource health settings (Brown et al., 2013). Between June to August 2018, 46 in-depth interviews with health professionals (nurses, clinical officers, registrars, consulting pediatricians, district health officers, district medical officers, and district nursing officers) were conducted for the primary study at one large tertiary hospital and three district-level hospitals. Research Ethics approvals for the “Integrating a neonatal healthcare package for Malawi” study were obtained from the University of Malawi College of Medicine (P.08/15/1783) and the University of British Columbia (H15-01463-A003). The qualitative process evaluation reported in the current paper was part of the main study to improve the data collection process.
Process of Preparing for and Conducting Evaluation Interviews
Characteristics of Data Collectors.
While our team had previous research experience, it was primarily quantitative research supporting randomized controlled trials. Three interviewers had inadequate knowledge and skills to moderate qualitative interviews, differentiating open and closed-ended questions, probing, conducting member checking after an interview to validate responses, taking field notes that provide context to qualitative interviews and employing reflexivity (Tolley et al., 2016). The gap in qualitative research knowledge and skills posed a challenge of risking collecting invalid information that is not rich enough to reflect the actual situation because the interviewer potentially affects the study’s outcomes. Consequently, as one way of mitigating the researcher as instrument effect (Tolley et al., 2016), there were three training sessions for data collectors. The data collectors already had training in good clinical practice. The current training included background on approaches to qualitative research, methods used in qualitative data collection, confidentiality, good clinical practices, data management and practice interview sessions. An experienced qualitative researcher led the formal training. At the same time, the final workshop was peer-led by one of the data collectors, who had previous experience in qualitative research and conducting interviews. Data collectors reviewed the interview guide and discussed how to rephrase leading questions during interview practice.
During training, interviewers focused on the importance of learning the difference between open- and closed-ended questions, and they practiced using “how” and “what” questions to obtain rich responses. For example, asking health workers what influenced their decision to commence a baby on bubble CPAP encouraged reflection on their own experiences. This helped to elicit examples or stories rather than simply responding “yes” or “no”. A key point described during training was not to make assumptions about familiarity with clinical knowledge based on the respondent’s qualifications. For example, when a nurse participant was interviewed, interviewers were advised to probe for explicit description instead of presuming that the participant’s method of checking vital signs was the same as what the interviewer practiced.
Data collectors used a semi-structured topic guide developed through a scoping literature review and consultation with Malawian health professionals. After each topic section, the interviewer summarized what was discussed to check the content and provide an opportunity for the participant to add.
The interviews ranged between 30 and 60 minutes and were audio-recorded with the permission of the study participants. Participants could conduct the interviews in English or the Malawi’s local language (Chichewa), depending on their preference. However, health workers largely responded in English with a few local language terms interspersed as English is the language of clinical training in Malawi. All the interviewers were bilingual, which worked well in accommodating participants’ preferences. Audio files were transcribed, and local language terms were translated into English for thematic analysis. Further details on the methods and findings of the primary study on barriers and facilitators to implementing bubble CPAP in hospitals are reported elsewhere (Nyondo-Mipando et al., 2020).
Our Method: A Qualitative Evaluation of the Data Collection Process
The current qualitative process evaluation of the data collection methods was guided by Patton’s (2003) evaluation checklist. It included all five data collection team members who conducted interviews for the primary study (O’Brien, Harris, Beckman et al., 2014). We held debriefing meetings among the five primary study interviewers, and two co-investigators moderated the debriefing interviews. The debriefing interviews continued during the entire data collection period between June to October 2018 as an ongoing process evaluation to improve data collection tools and address any ethical and research issues for subsequent interviews for the primary study. We used an inductive approach to the evaluation to understand the interviewers’ experiences conducting interviews with fellow health workers in a clinical setting (Patton, 1999). The process mainly informed the qualitative methods evaluation of data collection reflections, lessons learned and overall study’s planned debriefing sessions amongst the data collectors.
Data collectors also discussed the interview field notes, including comments on the setting, behaviors, thoughts, questions and concerns. For example, data collectors noted concerns about time for the interviews and reflected on contrasting responses between participants. These debriefing sessions followed the ‘SWOT’ framework (strength, weakness, opportunity and threats) to encourage reflection and asked interviewers to share examples from their experiences of strengths (what worked), weaknesses (what did not work well), opportunities (recommendations for improvement) and threats in the primary study (systemic barriers to conducting the research) (Helms, 2010; Nam et al., 2019; Wazir et al., 2013).
SWOT became the study’s conceptual framework, and the strengths, weaknesses, opportunities, and threats became deductive codes. Debriefing meeting notes, discussions and reflections constituted narratives of the data collection process and the unit of analysis. They were documented in Word (Microsoft Office) at the time of collection by one researcher and verified by two researchers who also attended the debriefing meetings (Mack et al., 2005). One researcher hand-coded the narratives in collaboration with the project research team. Debriefing meetings were not audio recorded. Credibility was also supported by several debriefing meetings conducted, which provided prolonged engagement with the data collectors during the data collection period from June to August 2018 and the evaluation of the process from September to October 2018 to understand their experiences in conducting the qualitative interviews (Patton, 1999). The complete documented narratives were shared with the data collectors to verify the interpretation as a form of member checking. De-identified research narratives are provided from the perspective of the interviewers.
Findings
Four major themes emerged from the narratives of the data collection processes: 1) the importance of practical component in training, 2) inadequate time for interviews, 3) positionality of the interviewer, and 4) professional hierarchies in healthcare. The themes provide a platform to understand the interview process in a resource-constrained health setting; the findings could be used to strengthen qualitative health research.
Importance of Practical Component in Training
All data collectors found the practical component of qualitative interview training essential. The interactive training included role-playing and a thorough review of the interview guide. Data collectors practiced rephrasing questions to best elicit responses from interviewees. The practice interviews gave insight into the process, helped familiarize data collectors with the guide and developed facilitation skills. ‘I received three training days that included theoretical background to qualitative research and interview practice. The practice itself was important because the theory is useless without practice.
Inadequate Time for Interviews
Interviewers shared that time was a challenge in conducting qualitative interviews with health care workers. The interviewers indicated that some health professionals did not have enough time for interviews, while others appeared rushed. Interviewers realized the tight schedule for health workers and recommended a semi-structured interview guide to focus on critical areas of interest and expertise, especially for senior clinicians. A semi-structured interview guide allowed flexibility to modify the phrasing of questions and skip questions that participants may have already addressed. ‘When I walked into her office, the doctor I was about to interview half-jokingly mentioned that I was 3 minutes late. With her position and the hospital’s setup, I appreciated that she took the time to do an interview…She appeared rushed throughout the interview, but I knew she had other responsibilities as interviews were conducted within the hospital setting...’ Clinicians faced challenges in setting aside time for interviews without compromising clinical duties and data collectors reflected on the numerous interruptions during the interviews. Data collectors’ flexibility to accommodate time adjustments and interruptions made the interviews possible. ‘I was given a 9 a.m. appointment, but when I arrived, the officer was already busy. I had to wait 45 minutes for her to finish...There were a few interruptions during the interview...
Positionality of the Interviewer
In particular, interviewers with a nursing background stressed their complex positionality in the interview process. Data collectors who also worked as nurses highlighted that they did not put on their nursing uniforms while in their role as researchers. By focusing on their researcher role, interviewers reported feeling freer to facilitate interviews without being recognized as nurses. Interview participants may be more comfortable sharing their experiences without feeling judged and would explain issues in detail without assuming that the interviewer already knew. ‘While conducting an interview, a fellow data collector in the study who is not a nurse asked me, “Why didn’t you put on a uniform?” I am a nurse by training. My response was that I was gathering data as a researcher, not as a nurse. I stated that I did not want my nurse position to influence the interview’s outcome. This was done strategically to create a safe environment for sharing rather than one in which one feels their skills are being evaluated and judged, encouraging explaining issues rather than assuming that I already know as a nurse…
The strong health background of data collectors led to a deep understanding of the issues raised by participants. Nurse-interviewers expressed how their clinical experience allowed them to ask relevant probing questions and clarify points regarding the health contexts where responses were unclear. ‘I had previously used bubble CPAP while working as a nurse on the wards, and that background knowledge assisted me in probing the participants to delve deeper into issues during the interviews...’
Professional Hierarchies
It was a challenge for nurse-interviewers to navigate the medical professional hierarchy. Some mentioned feelings of trepidation when interviewing senior medical personnel. Due to existing authority structures in the medical field, some interviewers were hesitant to probe deeper into questions for fear that senior clinicians would challenge their level of knowledge. Data collectors reported it was important to reflect on their roles as researchers rather than nurses within the study’s context and conduct interviews external to their usual workplace, particularly when interviewing senior staff. ‘I was nervous about interviewing a pediatric consultant because I considered myself to be of a lower medical caliber than the person I was questioning. I was nervous because I expected to be questioned about my understanding of our conversation as an interviewer. Despite my negative thoughts, I quickly reminded myself that, in the context of our interviews, I was not a nurse but rather an interviewer. I overcame my reservations by approaching the interviewee as a layperson…Because I was hired as an interviewer rather than a nurse, I knew I could question them about any ambiguity and that they would explain things in medical terms even if I already knew what they meant...’
Discussion
This qualitative methods evaluation paper, embedded within a qualitative health research project in Malawi, demonstrates how a strong background aided interviewers’ ability to conduct successful interviews in health. On the other hand, researchers should consider how their knowledge and skills may influence study findings.
We found that comprehensive training for qualitative interviews was required before data collection. Most importantly, the practical component of the training was highly beneficial in collecting high-quality data. Due to high workload and staff shortages in resource-limited hospital settings, scheduling interviews with health professionals was difficult. It was beneficial to schedule interviews for health workers outside of working hours and to use skilled interviewers to keep the interview focused.
Lessons Learned on Building Capacity for Qualitative Research
Human resource constraints limit the ability to conduct qualitative research in health contexts worldwide (Liu et al., 2016). This is especially true in resource-constrained health settings in LMICs. However, the pressure of interviewing fellow health care workers regarding reflexivity and hierarchy is standard in both HIC and LMICs. The following are four lessons learned on building capacity for qualitative health research in Malawi. Although the lessons were learned in Malawi, the following sections discuss the implications for high-income countries and other low- and middle-income countries.
The Importance of Comprehensive Interview Training and Skill Practice
Extensive training with a focus on practice is essential for developing interviewing skills. Qualitative research in HIC and LMIC settings necessitates using skilled and knowledgeable data collectors to elicit and engage study participants in opening up and sharing their lived experiences. Interviewing techniques can impact the data collection process, and several sources recommend that novice researchers be trained before conducting interviews to improve their skills. (Choo et al., 2016; Ranney et al., 2015). Practical sessions with constructive feedback were important in the training process to enhance the data collectors’ skills. Data collectors needed the training to build interview skills, including how to rephrase questions and follow up on issues the participant spontaneously brings up.
In addition, training on good listening skills was valuable in capturing detailed participant narratives, helping interviewers to summarize participants’ thoughts and probe with questions that furthered understanding of issues and events that participants had raised. This flexibility was possible due to a detailed briefing on the research objectives of the study, comprehensive training on qualitative interviewing methods and engaging our interviewers as research partners empowered to adapt the questionnaire in real-time.
Considering Health Workers’ Time in Resource-Constrained Settings
Finding time to conduct interviews with health professionals within a health facility setting is challenging. Providing detailed narratives took time because health professionals needed to reflect and share their experiences. At the same time, the health professionals were expected to provide care to clients. Data collectors need to arrange for interviews outside working hours to create time for the interviews. However, recruiting health care worker participants outside working hours may be challenging, particularly in resource-constrained health settings due to staffing shortages and often being called back during their off working hours to cover the wards (Mangham, 2007). For instance, one of the district hospitals in a rural area where interviews were conducted had a pool of five nurses who worked in the newborn unit. With two nurses scheduled per shift, the small pool of nursing staff meant limited time off-duty to fulfil their personal and family commitments and frequently being called back to cover holidays, sick days, and when nurses are pulled into other wards. Although our study never conducted interviews immediately after the working hours of the study participants, conducting interviews immediately after working hours is another option for researchers to consider. However, there is a risk of having interviews rushed through because participants may look forward to knocking off after a long day with a high workload.
Flexibility and patience in rescheduling interview appointments and working around the health professionals’ schedules can support qualitative research in resource-limited health facility settings. While data collectors’ understanding of time constraints was beneficial, it is important to note that it can have a negative influence if interviewers are not skillful to adjust in real-time. In a busy and understaffed hospital, which is common in LMICs, interviews may be interrupted or rushed.
Recognizing that their Position as Nurses could Affect Data Collection
The findings of this study highlighted reflexivity as a vital component of training for our interviewers. Reflexivity is examining oneself as an interviewer and noting personal values that could affect data collection and interpretation (Palaganas et al., 2017). Interviewers with strong health worker backgrounds and reflexivity training were essential to promote an understanding how their knowledge, skills and position could affect data collection. There was a required role shift from clinician to researcher to minimize bias and preconceived notions health professionals may hold. If reflexivity was not practiced, research nurses in both HIC and LMIC settings might neglect to probe more deeply due to an assumption of shared knowledge. Berger (2015) reflected on how a researcher’s positionality can affect reflexivity in qualitative research. As an immigrant, Berger revealed how her insider position helped participants feel more open and willing to participate in her research on immigrant women. However, she was challenged by the participants’ assumption that she would already know the information. These dual roles require continuous reflection and negotiation and warrant discussion among team members, as was done in our study.
Similarly, Chew-Graham et al., (2002), in qualitative research conducted in a HIC setting, indicated respondents’ expressions of shared understanding about a research subject with the interviewer when they knew that the interviewer was a fellow professional. Being aware of both positives and negatives of the professional identity of interviewers and finding ways to mitigate the negatives, like employing reflexivity, is important regardless of the research setting. Chew-Gram (2002) suggest that studies should also report and discuss when reporting findings of such particular work.
Our nurse researchers employed reflexivity to balance their outsider-insider position as nurses to ensure equal power-sharing with the respondents. Researchers can practice reflexivity in several ways, such as holding debriefing sessions with the research team, internally reflecting on the research process, member checking and keeping a journal of the research process.
Internal reflection could be achieved both by prospective and retrospective reflexivity. Our interviewers prospectively reflected that putting on their nursing uniforms could influence data collection, and they avoided it. Participants in the primary study did not know that most interviewers were fellow professionals. Our interviewers for the primary study purposively blinded their profession because they reflected that respondents could assume that they were testing their knowledge. In agreement with Chew-Gram (2002), in lessons drawn from interviewing fellow professionals, they reported that some respondents thought they were trying their knowledge and regarded interviewers as experts.
Interestingly, some of our interviewers shared in their reflection about feeling anxious when interviewing senior professional participants thinking they would bring medical terms or discussions they would not understand. Therefore, interviewers must reflect on their attitudes and the participants as professionals in the same field and how they could potentially affect the collected data. In retrospective debriefing sessions, the data collection team reflected on how their health backgrounds may have influenced how they approached the interview process, which was essential to add to the context of the interviews when analyzing the data. Attia and Edge (2017) use prospective and retrospective reflexivity to guide how she could collect data among her fellow teachers without compromising data quality. Attia reflected on her professional background to inform her study design and retrospectively reflected on actions she was involved in during data collection (Attia & Edge, 2017).
The Use of Healthcare Workers as Interviewers
The ability of health workers to understand medical terms, contexts, and social interactions within hospital settings is one of the benefits of using them in data collection (Faulkner-Gurstein et al., 2019). We demonstrated that strong background knowledge in the research area should be considered when planning interviews in a health setting. On the other hand, interviewers with health backgrounds must balance their roles as health professionals and researchers both in HICs and LMICs. According to Kaikelame et al. (2018, 2019), the data collector and the research subject contribute significantly to the research process; thus, a positive relationship between data collectors and the research is essential. According to Larkin et al. (2017), research nurses must maintain good working relationships with fellow health workers and the nurse-patient therapeutic relationship while carrying out their research roles.
Health professionals working as interviewers in both HIC and LMIC settings face challenges if the study occurs where they work because they are more familiar with patients and coworkers (Yanos & Ziedonis, 2006). Research nurses may be tempted to support clinical care in resource-limited health settings, mainly when there is a shortage of staff and a heavy workload at the time of interviews. In other cases, where the healthcare worker respondent and patients know that the interviewer is a nurse, they may also expect them to provide clinical support or advice. A systematic review to develop a typology of clinician-researcher dual-role experiences in health research with the patient –participants reported situations where patients asked the researcher treatment-related questions, which made researchers uncomfortable, fearing influencing data quality (Hay-Smith et al., 2016). Hill (2018) describes a similar experience in which clinical research nurses assisted with clinical care while clinical nurses were conducting interviews. Because nurse researchers’ roles can blur, they may jeopardize participants’ privacy and cast doubt on the confidentiality of their findings. Our interviewers concealed their identity as health care workers and moderated interviews in a different health facility. However, the interviewers reflected on the urge to provide support with care for the infant on bubble CPAP when there were several interruptions during the interviews.
Strengths and Limitations
Our interview methods study was not a planned component of the primary study but rather emerged as an important topic of reflection to improve the data collection process with a focus on debriefing meetings with data collectors, which represents a limitation of the study as the analysis was post hoc. These reflections are also based on the experiences of five data collectors; thus, further research is needed to elucidate and confirm our exploratory study’s findings in other contexts. The strength of our study is the continued engagement with data collectors to support reflection on the interviewing process and confirm interpretations, which supports the credibility of the findings. In addition, we conducted the qualitative process evaluation in Malawi an understudied LMICs.
Conclusion
The health care setting in Malawi is challenged by staff shortages that render the available workers to have multiple roles. However, we still need to research health facility settings to improve several aspects of health delivery. Qualitative research requires adequate time to capture detailed and in-depth narratives. The availability of sufficient time is challenged by the human resource gap in health care settings in low-income countries. Therefore, researchers must plan and think through ways to overcome those challenges, as reflected in our study. Healthcare workers are uniquely positioned to conduct qualitative health research. Still, they need to be aware of their position or influence and how it may impact the interview dynamics and the validity of the results obtained. Successful implementation of evidence-based health interventions requires understanding context, for which qualitative research is a valuable tool. This article provides a good basis for planning and conducting qualitative research that accommodates context in LMICs.
Footnotes
Acknowledgments
This manuscript is part of the “Integrating a neonatal healthcare package for Malawi” project within the IMCHA initiative. The authors would like to express their gratitude to the IMCHA team especially the data collectors for their support and all of the health workers who participated in the study. We are thankful for the institutional support from the Hospitals for allowing us to conduct the study in their facilities. The last author is supported by Malawi HIV Implementation Research Scientist Training program.
Declaration of Conflicting Interest
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: This work was supported by the authors disclosed receipt of the following financial support for the research: Integrating a neonatal healthcare package for Malawi is funded by the Canadian International Development Research Centre (IDRC) in partnership with Global Affairs Canada (GAC) and the Canadian Institutes for Health Research (CIHR).
